DETAILED ACTION
Applicant's amendment of April 10, 2026 overcomes the following:
Objections to claims 2-9 and 13
Specification objections
Rejection of claims 17-18 based on 35 USC 101
Rejection of claims 1-18 under 35 U.S.C. 112(b), pre-AIA 35 U.S.C. 112, second paragraph
Applicant has amended claims 1 -18. Claims 19-20 are new. Claims 1-20 are pending.
Response to Arguments
Applicant’s arguments filed on April 10, 2026 with respect to pending claims have been considered but are moot in view of the new ground(s) of rejection. The amended claims resulted in changes to the scope and contents; therefore, the grounds of rejection are modified accordingly.
Regarding interpretation of claims under 35 U.S.C. 112(f), pre-AIA 35 U.S.C. 112, sixth paragraph, indicated in the Non-Final Office action (OA), Applicant asserts that “because the term "means for” is not recited with respect to any of these limitations, a rebuttable presumption arises that 35 U.S.C. 112(f) does not apply… the terms "machine learning module" take their name from the functions the machine learning module performs. Thus, Applicant respectfully submits that the Office Action has not successfully rebutted the presumption discussed above, i.e., that the terms provide sufficient structure for performing the corresponding functions, when considered in light of the specification and commonly accepted meanings in the technological art” (Remarks, Pg. 11-12).
Examiner respectfully disagrees.
As previously indicated in Pg. 7 of the Non-Final OA, this “application includes one or more claim limitations that do not use the word "means," but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier”.
For example, claims 1-4 and 10-13 recite generic placeholders “module”, “modules”, “selector”, “system”, “forward-projector”, “comparator”, “determiner”, “back-projector”, “updater”, and “reconstructor”, respectively, that act as substitutes for the word “means”, and are coupled with corresponding recited functional language “module for facilitating an iterative reconstruction operation”, “module configured to: receive… predict… provide”, “module acts”, “modules, configured for… reduction”, “selector for receiving”, “system for iterative reconstruction”, “forward-projector to map”, “comparator configured to establish”, “determiner… to provide”, “back-projector configured to back-project”, “updater configured to apply”, “system configured to train”, “system for generating”, “reconstructor configured to:… to process”, respectively, without reciting sufficient structure to perform each recited function and each generic placeholder is not preceded by a structural modifier, as previously indicated in the Non-Final OA.
As indicated in MPEP § 2173 et seq. and Supplementary Examination Guidelines for Determining Compliance With 35 U.S.C. 112 and for Treatment of Related Issues in Patent Applications, 76 FR 7162, 7167 (Feb. 9, 2011), “a claim limitation that does not use the phrase ‘‘means for’’ or ‘‘step for’’ will trigger the rebuttable presumption that § 112, ¶6 does not apply.67 This presumption is a strong one that is not readily overcome.68 This strong presumption may be overcome if the claim limitation is shown to use a non-structural term that is ‘‘a nonce word or a verbal construct that is not recognized as the name of structure’’ but is merely a substitute for the term ‘‘means for,’’ associated with functional language.69… When the claim limitation does not use the phrase ‘‘means for’’ or ‘‘step for,’’ examiners should determine whether the claim limitation uses a non-structural term (a term that is simply a substitute for the term ‘‘means for’’). Examiners will apply §112, ¶6 to a claim limitation that uses a non-structural term associated with functional language, unless the non-structural term is (1) preceded by a structural modifier, defined in the specification as a particular structure or known by one skilled in the art, that denotes the type of structural device (e.g., ‘‘filters’’), or (2) modified by sufficient structure or material for achieving the claimed function. The following is a list of non-structural terms that may invoke § 112, ¶6: ‘‘mechanism for,’’ ‘‘module for,’’ ‘‘device for,’’ ‘‘unit for,’’ ‘‘component for,’’ ‘‘element for,’’ ‘‘member for,’’ ‘‘apparatus for,’’ ‘‘machine for,’’ or ‘‘system for.’’72 This list is not exhaustive, and other non-structural terms may invoke §112, ¶6”.
Based on above rationale, the use of the non-structural terms “module for”, “module configured to”, “modules, configured for”, “selector for”, “system for”, “forward-projector to”, “comparator configured to”, “determiner… to”, “back-projector configured to”, “updater configured to”, “system configured to”, “reconstructor configured to”, respectively, which are substitutes for the term “means for” performing (i.e. to perform) the claimed functions recited in the claims, as indicated above, do not suggest sufficient definite structure to one of ordinary skill in the art to preclude application of 35 U.S.C. 112(f), pre-AIA 35 U.S.C. 112, sixth paragraph.
Therefore, based on above, Applicant’s remarks regarding interpretation of claims under 35 U.S.C. 112(f), pre-AIA 35 U.S.C. 112, sixth paragraph, previously indicated in the last OA have been respectfully found unconvincing because the claim limitation(s) recited above use generic placeholders that are coupled with functional language without reciting sufficient structure to perform the recited functions and the generic placeholders are not preceded by a structural modifier, as indicated above, and previous claim interpretation indicated in the Non-Final OA is believed to be proper and is hereby maintained.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-6 and 8-20 are rejected under 35 U.S.C. 103 as being unpatentable over CAO et al. (US PG Publication No. 2021/0104023 A1), hereafter referred to as CAO, in view of RA et al. (US PG Publication No. 2015/0243070), hereafter referred to as RA.
Regarding claim 1, CAO discloses a system (Par. [0002]: systems and methods for iterative image reconstruction) comprising trained at least one machine learning module (Par. [0008]: optimizing model may include a machine learning model; Par. [0102]: the optimizing model may be pre-trained and stored in a storage device… the optimizing model may include a machine learning model, for example, a neural network model) for facilitating an iterative reconstruction operation (Par. [0002-3]: systems and methods for iterative image reconstruction… CT image can be iteratively reconstructed; Par. [0115]: processing device 120 (e.g., the reconstruction module 420) (e.g., the processing circuits of the processor 210) may receive an image to be processed in the current iterative reconstruction… the image to be processed is an optimized image determined in a previously adjacent iterative reconstruction), wherein, in a plurality of steps, imagery in image domain is reconstructable from measured projection data in projection domain (Par. [0057]: a reconstructed image is generated by performing a plurality of iteration steps each of which includes a first optimization operation (which is used to preliminarily optimize the image to be processed) and at least one second optimization operation (which is used to further optimize an updated image generated in the first operation) implemented via a machine learning model, that is, a preliminary optimization and a further optimization via a machine learning model are used in combination, which can improve the image quality of the reconstructed image and reduce the noise of the reconstructed image; Par. [0092-96]: processing device 120 (e.g., the reconstruction module 420) (e.g., the processing circuits of the processor 210) may determine an updated image… processing device 120 may determine forward projection data by performing a forward projection transformation on the image to be processed in the iteration step. According to the forward projection transformation… transform data (e.g., the image to be processed) in an image domain to data (e.g., the forward projection data) in a projection domain… determine the updated image based at least in part on back projection data of a weighted error between the forward projection data and originally acquired projection data associated with the initial image. As used herein, the originally acquired projection data associated with the initial image may refer to original projection data acquired by the scanning device 110, for example, original projection data upon which the initial image is determined… determine a difference between the forward projection data and originally acquired projection data as an error between the forward projection data and originally acquired projection data… determine the back projection data of the weighted error by performing a backward projection transformation on the weighted error. According to the back projection transformation… transform data (e.g., the weighted error) in a projection domain to data (e.g., the back projection data) in an image domain; wherein, in a plurality of steps, imagery in image domain is reconstructable from measured projection data in projection domain (e.g. systems and methods for iterative image reconstruction, in which reconstructed images (i.e. imagery) are generated by performing a plurality of iteration steps (i.e. wherein, in a plurality of steps, imagery is reconstructable), for example, include transforming data in a projection domain, including originally acquired projection data (i.e. measured projection data in projection domain), to data (i.e. imagery) in an image domain (i.e. imagery in image domain is reconstructable from measured projection data), as indicated above), for example), the at least one machine learning module configured to (Par. [0030-31]: system may further include a training module. The training module may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model… a non-transitory computer readable medium including executable instructions. When the executable instructions are executed by at least one processor, the executable instructions may direct the at least one processor to perform a method. The method may include obtaining an initial image to be processed and generating a reconstructed image by performing a plurality of iteration steps on the initial image. Each of the plurality of iteration steps may include a first optimization operation and at least one second optimization operation; Par. [0063-80]: storage device 150 may store data, instructions, and/or any other information… the storage device 150 may store data obtained from the scanning device 110, the terminal device 130, and/or the processing device 120… the storage device 150 may store data and/or instructions that the processing device 120 may execute or use to perform exemplary methods described in the present disclosure… training module 430 may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples… the reconstruction module 420 may be divided into a first reconstruction unit configured to perform the first optimization operation and a second reconstruction unit configured to perform the at least one second optimization operation):
receive input correction data generated in the iterative reconstruction operation;
predict, based on the input correction data, output correction data (Par. [0031-32]: obtaining an initial image to be processed and generating a reconstructed image by performing a plurality of iteration steps on the initial image. Each of the plurality of iteration steps may include a first optimization operation and at least one second optimization operation… The first optimization operation may include receiving an image to be processed in the iteration step and determining an updated image by preliminarily optimizing the image to be processed. The at least one second optimization operation may include determining an optimized image by reducing interference information of the updated image and designating the optimized image as a next image to be processed in a next iteration step. The interference information may include noise information and/or artifact information… obtaining projection data to be processed and generating a reconstructed image by processing the projection data based on an image reconstruction model. The image reconstruction model may include a plurality of sequentially connected sub-models. Each of the plurality of sequentially connected sub-models include a processing layer and a reconstruction layer… configured to receive an image to be processed in the sub-model and determine a regularization result by regularizing the image to be processed based on the projection data. The reconstruction layer may be configured to determine a sub-reconstructed image based on the regularization result and a previous sub-reconstructed image determined by a previously adjacent sub-model and designate the sub-reconstructed image as a next image to be processed in a next sub-model; Par. [0056]: obtain an initial image (e.g., a CT image) to be processed and generate a reconstructed image by performing a plurality of iteration steps on the initial image. At least one of the plurality of iteration steps may include a first optimization operation and at least one second optimization operation… In the first optimization operation in an iteration step, the systems may receive an image to be processed (e.g., the initial image for the first iteration step) in the iteration step and determine an updated image by preliminarily optimizing the image to be processed. In the at least one second optimization operation in the iteration step… determine an optimized image by reducing interference information (e.g., noise information, artifact information) of the updated image and designate the optimized image as a next image to be processed in a next iteration step. The systems may determine the optimized image by reducing the interference information of the updated image based on a machine learning model (e.g., a deep learning model); Par. [0078]: generate a reconstructed image by performing a plurality of iteration steps on the initial image… each of the plurality of iteration steps may include a first optimization operation and at least one second optimization operation… the first optimization operation and the at least one second optimization operation may be executed sequentially. Take a specific iteration step as an example, the first optimization operation may include receiving an image to be processed in the iteration step (e.g., for the first iteration step, the image to be processed is the initial image) and determining an updated image by preliminarily optimizing the image to be processed. The at least one second optimization operation may include determining an optimized image by reducing interference information (e.g., noise information, artifact information) of the updated image and designating the optimized image as a next image to be processed in a next iteration step… for the at least one second optimization operation, the reconstruction module 420 may determine the optimized image by reducing the interference information based on an optimizing model (e.g., a machine learning model); Par. [0095-106]: processing device 120 may determine a first quality weight associated with the originally acquired projection data. For example, the processing device 120 may determine the first quality weight based on interference information (e.g., noise information, artifact information) in the originally acquired projection data… the interference information may include noise information and/or artifact information… the updated image may include different types of interference information (e.g., the noise information, the artifact information). In order to eliminate the different types of interference information as much as possible… determine an optimizing model corresponding to each type of interference information. For example, the processing device 120 may determine an artifact optimizing model for reducing or eliminating the artifact information of the updated image… determine a noise optimizing model for reducing or eliminating the noise information of the updated image; Par. [0142-149]: processing device 120 may input the projection data to be processed and/or an initial image) into the image reconstruction model… the image reconstruction model may include a plurality of sequentially connected sub-models. Each of the plurality of sequentially connected sub-models may include a processing layer and a reconstruction layer. The processing layer may be configured to receive an image to be processed (for the first sub-model, the image to be processed is the initial image) in the sub-model and determine a regularization result by regularizing the image to be processed based on the projection data. The reconstruction layer may be configured to determine a sub-reconstructed image based on the regularization result and a previous sub-reconstructed image determined by a previously adjacent sub-model and designate the sub-reconstructed image as a next image to be processed in a next sub-model… the image reconstruction model 1000 may include a plurality of sequentially connected sub-models, such as a first sub-model 1010 and one or more second sub-models 1020 (e.g., 1021 and 1022)… the first sub-model 1010 may be configured to receive the input (e.g., the originally acquired projection data and/or the initial image) of the image reconstruction model 1000 and generate an output. Each of the second sub-model 1020 may be configured to receive an output of a previously adjacent sub-model connected to the second sub-model 1020 and generate an output. The output of the image reconstruction model 1000 may be the output of the last second sub-model… each of the plurality of sub-models may include a processing layer… and a reconstruction layer… The reconstruction layer may be configured to determine a sub-reconstructed image based on the regularization result and a previous sub-reconstructed image determined by a previously adjacent sub-model… for the first sub-model 1010, the input of the reconstruction layer (e.g., the reconstruction layer 1) may include an output (i.e., the regularization result of the initial image) of the processing layer (e.g., the processing layer 1) of the first sub-model 1010 and the input (e.g., the projection data) of the image reconstruction model 1000; for the second sub-model 1020 (e.g., 1021), the input of the reconstruction layer (e.g., the reconstruction layer 2) may include an output (i.e., the regularization result of the sub-reconstructed image) of a processing layer (e.g., the processing layer 2) of the second sub-model 1020 (e.g., 1021) and a sub-reconstructed image output by a reconstruction layer (e.g., the reconstruction layer 1) in a previously adjacent sub-model (e.g., the first sub-model 1010). The reconstruction layer may be also configured to designate the sub-reconstructed image as a next image to be processed in a next sub-model. As illustrated, the output of the reconstruction layer of each of the plurality of sub-models may include a sub-reconstructed image corresponding to the sub-model. Accordingly, the output of the last reconstruction layer (e.g., the reconstruction layer n) of the last second sub-model (e.g., 1022) may be the reconstructed image; receive input correction data generated in the iterative reconstruction operation; predict, based on the input correction data, output correction data (e.g. systems and methods for iterative image reconstruction, in which reconstructed images (i.e. imagery) are generated by performing a plurality of iteration steps (i.e. iterative reconstruction operation) on originally acquired projection data (i.e. measured projection data in projection domain) and/or an initial image (i.e. imagery in image domain), for example, include receiving an image to be processed in an iteration step as input (i.e. receive input data in the iterative reconstruction operation), such as the initial image, for example, and determining (i.e. predicting) an updated image as output generated in the iterative reconstruction operation, such as a sub-reconstructed image, for example, by preliminarily optimizing the image to be processed in order to reduce (i.e. correct) interference (i.e. noise, error, deviation, etc.) information (i.e. receive input correction data generated in the iterative reconstruction operation), including noise information and artifact information of the updated image, for example, by reducing (i.e. correcting) interference information of the updated image (i.e. output correction data) in order to determine an optimized image as output (i.e. predict, based on the input correction data, output correction data) and designate the optimized image as a next image to be processed in a next iteration step, as indicated above), for example);
provide the output correction data for facilitating correcting a current image into a new image, wherein the new image is used for a next iteration step of the plurality of steps or output as a final image (Par. [0031-32]: obtaining an initial image to be processed and generating a reconstructed image by performing a plurality of iteration steps on the initial image. Each of the plurality of iteration steps may include a first optimization operation and at least one second optimization operation… first optimization operation may include receiving an image to be processed in the iteration step and determining an updated image by preliminarily optimizing the image to be processed. The at least one second optimization operation may include determining an optimized image by reducing interference information of the updated image and designating the optimized image as a next image to be processed in a next iteration step. The interference information may include noise information and/or artifact information… obtaining projection data to be processed and generating a reconstructed image by processing the projection data based on an image reconstruction model. The image reconstruction model may include a plurality of sequentially connected sub-models. Each of the plurality of sequentially connected sub-models include a processing layer and a reconstruction layer. The processing layer may be configured to receive an image to be processed in the sub-model and determine a regularization result by regularizing the image to be processed based on the projection data. The reconstruction layer may be configured to determine a sub-reconstructed image based on the regularization result and a previous sub-reconstructed image determined by a previously adjacent sub-model and designate the sub-reconstructed image as a next image to be processed in a next sub-model; Par. [0056]: obtain an initial image… to be processed and generate a reconstructed image by performing a plurality of iteration steps on the initial image. At least one of the plurality of iteration steps may include a first optimization operation and at least one second optimization operation. The first optimization operation and the at least one second optimization operation may be executed sequentially. In the first optimization operation in an iteration step, the systems may receive an image to be processed (e.g., the initial image for the first iteration step) in the iteration step and determine an updated image by preliminarily optimizing the image to be processed. In the at least one second optimization operation in the iteration step… determine an optimized image by reducing interference information (e.g., noise information, artifact information) of the updated image and designate the optimized image as a next image to be processed in a next iteration step… determine the optimized image by reducing the interference information of the updated image based on a machine learning model (e.g., a deep learning model); Par. [0078]: generate a reconstructed image by performing a plurality of iteration steps on the initial image… each of the plurality of iteration steps may include a first optimization operation and at least one second optimization operation… Take a specific iteration step as an example, the first optimization operation may include receiving an image to be processed in the iteration step (e.g., for the first iteration step, the image to be processed is the initial image) and determining an updated image by preliminarily optimizing the image to be processed. The at least one second optimization operation may include determining an optimized image by reducing interference information (e.g., noise information, artifact information) of the updated image and designating the optimized image as a next image to be processed in a next iteration step… for the at least one second optimization operation, the reconstruction module 420 may determine the optimized image by reducing the interference information based on an optimizing model (e.g., a machine learning model); Par. [0095-106]: processing device 120 may determine a first quality weight associated with the originally acquired projection data. For example, the processing device 120 may determine the first quality weight based on interference information (e.g., noise information, artifact information) in the originally acquired projection data… the interference information may include noise information and/or artifact information… the updated image may include different types of interference information (e.g., the noise information, the artifact information). In order to eliminate the different types of interference information as much as possible… determine an optimizing model corresponding to each type of interference information. For example, the processing device 120 may determine an artifact optimizing model for reducing or eliminating the artifact information of the updated image. As another example, the processing device 120 may determine a noise optimizing model for reducing or eliminating the noise information of the updated image; Par. [0142-149]: processing device 120 may input the projection data to be processed and/or an initial image) into the image reconstruction model… the image reconstruction model may include a plurality of sequentially connected sub-models. Each of the plurality of sequentially connected sub-models may include a processing layer and a reconstruction layer. The processing layer may be configured to receive an image to be processed (for the first sub-model, the image to be processed is the initial image) in the sub-model and determine a regularization result by regularizing the image to be processed based on the projection data. The reconstruction layer may be configured to determine a sub-reconstructed image based on the regularization result and a previous sub-reconstructed image determined by a previously adjacent sub-model and designate the sub-reconstructed image as a next image to be processed in a next sub-model… the image reconstruction model 1000 may include a plurality of sequentially connected sub-models, such as a first sub-model 1010 and one or more second sub-models 1020 (e.g., 1021 and 1022)… the first sub-model 1010 may be configured to receive the input (e.g., the originally acquired projection data and/or the initial image) of the image reconstruction model 1000 and generate an output. Each of the second sub-model 1020 may be configured to receive an output of a previously adjacent sub-model connected to the second sub-model 1020 and generate an output. The output of the image reconstruction model 1000 may be the output of the last second sub-model… each of the plurality of sub-models may include a processing layer… and a reconstruction layer… The reconstruction layer may be configured to determine a sub-reconstructed image based on the regularization result and a previous sub-reconstructed image determined by a previously adjacent sub-model… for the first sub-model 1010, the input of the reconstruction layer (e.g., the reconstruction layer 1) may include an output (i.e., the regularization result of the initial image) of the processing layer (e.g., the processing layer 1) of the first sub-model 1010 and the input (e.g., the projection data) of the image reconstruction model 1000; for the second sub-model 1020 (e.g., 1021), the input of the reconstruction layer (e.g., the reconstruction layer 2) may include an output (i.e., the regularization result of the sub-reconstructed image) of a processing layer (e.g., the processing layer 2) of the second sub-model 1020 (e.g., 1021) and a sub-reconstructed image output by a reconstruction layer (e.g., the reconstruction layer 1) in a previously adjacent sub-model (e.g., the first sub-model 1010). The reconstruction layer may be also configured to designate the sub-reconstructed image as a next image to be processed in a next sub-model. As illustrated, the output of the reconstruction layer of each of the plurality of sub-models may include a sub-reconstructed image corresponding to the sub-model. Accordingly, the output of the last reconstruction layer (e.g., the reconstruction layer n) of the last second sub-model (e.g., 1022) may be the reconstructed image; provide the output correction data for facilitating correcting a current image into a new image, wherein the new image is used for a next iteration step of the plurality of steps or output as a final image (e.g. systems and methods for iterative image reconstruction, in which reconstructed images (i.e. imagery) are generated by performing a plurality of iteration steps (i.e. iterative reconstruction operation) on originally acquired projection data (i.e. measured projection data in projection domain) and/or an initial image (i.e. imagery in image domain), for example, include receiving an image to be processed in an iteration step as input (i.e. receive input data in the iterative reconstruction operation), such as the initial image, for example, and determining (i.e. predicting) an updated image (i.e. a current image) as output generated in the iterative reconstruction operation, such as a sub-reconstructed image, for example, by preliminarily optimizing the image to be processed in order to reduce (i.e. correct) interference (i.e. noise, error, deviation, etc.) information, including noise information and artifact information of the updated image, for example, by reducing (i.e. correcting) interference information of the updated image (i.e. output correction data) in order to determine an optimized image as output (i.e. provide the output correction data for facilitating correcting a current image into a new image) and designate the optimized image as a next image to be processed in a next iteration step (i.e. wherein the new image is used for a next iteration step), as indicated above), for example),
wherein the input correction data is in the image domain, based on back-projected error projection data (Par. [0096]: determine the back projection data of the weighted error by performing a backward projection transformation on the weighted error… transform data (e.g., the weighted error) in a projection domain to data (e.g., the back projection data) in an image domain… determine the updated image based on the back projection data of the weighted error), the back-projected error projection data representing a deviation (i.e. difference) between the measured (i.e. originally acquired) projection data and estimated projection data (i.e. forward projection data), and the estimated projection data being obtained by forward-projection into the projection domain of the current image (Par. [0094-95]: determine the updated image based at least in part on back projection data of a weighted error between the forward projection data and originally acquired projection data associated with the initial image… the originally acquired projection data associated with the initial image may refer to original projection data acquired by the scanning device… determine a first quality weight associated with the originally acquired projection data… determine the first quality weight based on interference information (e.g., noise information, artifact information) in the originally acquired projection data… determine a difference between the forward projection data and originally acquired projection data as an error between the forward projection data and originally acquired projection data), or
wherein the input correction data is in the projection domain (Par. [0093]: determine forward projection data by performing a forward projection transformation on the image to be processed in the iteration step. According to the forward projection transformation, the processing device 120 may transform data (e.g., the image to be processed) in an image domain to data (e.g., the forward projection data) in a projection domain), based on the deviation (i.e. difference) between the measured projection data (i.e. originally acquired) and the estimated projection (i.e. forward projection data) data (Par. [0094-95]: determine the updated image based at least in part on back projection data of a weighted error between the forward projection data and originally acquired projection data associated with the initial image… the originally acquired projection data associated with the initial image may refer to original projection data acquired by the scanning device… determine a first quality weight associated with the originally acquired projection data… determine the first quality weight based on interference information (e.g., noise information, artifact information) in the originally acquired projection data… determine a difference between the forward projection data and originally acquired projection data as an error between the forward projection data and originally acquired projection data), and the estimated projection data being obtained by the forward-projection into the projection domain of the current image (Par. [0093-94]: determine forward projection data by performing a forward projection transformation on the image to be processed in the iteration step. According to the forward projection transformation, the processing device 120 may transform data (e.g., the image to be processed) in an image domain to data (e.g., the forward projection data) in a projection domain… determine the updated image based at least in part on back projection data of a weighted error between the forward projection data and originally acquired projection data associated with the initial image).
CAO discloses systems and methods for iterative image reconstruction, in which reconstructed images (i.e. imagery) are generated by performing a plurality of iteration steps (i.e. iterative reconstruction operation) on originally acquired projection data (i.e. measured projection data in projection domain) and/or an initial image (i.e. imagery in image domain), for example, include receiving an image to be processed in an iteration step as input (i.e. receive input data in the iterative reconstruction operation), such as the initial image, for example, and determining (i.e. predicting) an updated image as output generated in the iterative reconstruction operation by preliminarily optimizing the image to be processed in order to reduce (i.e. correct) interference (i.e. noise, error, deviation, etc.) information, including noise information and artifact information of the updated image, for example, by reducing (i.e. correcting) interference information of the updated image (i.e. output correction data) in order to determine an optimized image as output (i.e. predict, based on the input correction data, output correction data) and designate the optimized image as a next image to be processed in a next iteration step, as indicated above, for example, but does not expressly recite “correction” and “correcting”.
However, RA teaches “correction” and “correcting” (Par. [0027]: image reconstructor may be further configured to correct the first information based on predicted data acquired by forward projecting the predicted third image with respect to the measured data; Par. [0156] correct the first information… correction of the first information by the image reconstructor 620; Par. [0266]: mage reconstructor 620 may acquire second information by correcting the first information; Par. [0279-288]: correction of the first information by the image reconstructor 620… first information is corrected based on a difference between the predicted sinogram and the measured sinogram, corrected first information 1620 may be obtained by correcting first information… image reconstructor 620 may compare a measured image obtained by back-projecting measured data acquired… with a predicted image, and correct the first information such that a difference between the measured image and the predicted image decreases).
CAO and RA are considered to be analogous art because they pertain to medical image processing applications. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to modify the systems and methods for iterative image reconstruction (as disclosed by CAO) with “correction” and “correcting” (as taught by RA, Abstract, Par. [0156, 266, 279-288]) to reduce an occurrence of motion artifacts within a reconstructed tomography by comparing predicted data with measured data to correct first information such that a difference between the predicted data and the measured data decreases (RA, Abstract, Par. [00019-23, 25-30]).
Regarding claim 2, claim 1 is incorporated and the combination of CAO and RA, as a whole, teaches the system (CAO, Par. [0002]), wherein the at least one machine learning module improves data quality so that the output correction data has a higher quality than the input correction data (CAO, Par. [0003]: systems and methods for image reconstruction with improved image quality; [0009-10]: determining a quality feature of the updated image… the quality feature may include a noise feature; Par. [0057]: a reconstructed image is generated by performing a plurality of iteration steps each of which includes a first optimization operation (which is used to preliminarily optimize the image to be processed) and at least one second optimization operation (which is used to further optimize an updated image generated in the first operation) implemented via a machine learning model, that is, a preliminary optimization and a further optimization via a machine learning model are used in combination, which can improve the image quality of the reconstructed image and reduce the noise of the reconstructed image; Par. [0085-86]: optimization operation may include determining an optimized image by reducing interference information (e.g., noise information, artifact information) of the updated image and designating the optimized image as a next image to be processed in a next iteration step… the optimized image by reducing the interference information based on an optimizing model (e.g., a machine learning model)… determine whether a termination condition is satisfied in the current iteration step. Exemplary termination conditions may include… the optimized image in the current iteration step has reached a desired image quality (e.g., a noise rate is less than a threshold)).
Regarding claim 3, claim 2 is incorporated and the combination of CAO and RA, as a whole, teaches the system (CAO, Par. [0002]), wherein the data quality comprises noise (CAO, Par. [0003]: systems and methods for image reconstruction with improved image quality; Par. [0009-10]: determining a quality feature of the updated image… the quality feature may include a noise feature; Par. [0103]: determine a quality feature of the updated image… input the updated image and the quality feature into the optimizing model and determine the optimized image based on an output of the optimizing model. The quality feature may include a noise feature), and wherein the at least one machine learning module comprises a plurality of machine learning modules (CAO, Par. [0102-108] optimizing model may be trained based on a plurality of training samples… plurality of interference information reduction operations may be executed based on a plurality of optimizing algorithms or optimizing models… the plurality of optimizing models may be models of different types… the plurality of optimizing models may be models of the same type but the same structure or different structures. For example, the plurality of optimizing models may be deep neural network models with different numbers of neural networks or/and different numbers of neurons. As another example, the plurality of optimizing models may be deep neural network models with different activation modes and/or different model structures; ), configured (i.e. trained) for a plurality of different noise reduction level levels, respectively (CAO, Par. [0126-128]: processing device 120 (e.g., the training module 430) (e.g., the processing circuits of the processor 210) may obtain a plurality of training samples… each of the plurality of training samples may include a sample image and a sample quality feature of the sample image… the sample quality feature may include a sample noise feature… The sample noise feature may include a sample noise distribution, a sample noise intensity, a sample global noise intensity, a sample noise rate… the plurality of training samples may correspond to various quality levels (e.g., various noise intensities, various artifact intensities)).
Regarding claim 4, claim 3 is incorporated and the combination of CAO and RA, as a whole, teaches the system (CAO, Par. [0002]), further comprising a selector for receiving a selector signal (CAO, Par. [0066-80]: the processing device 120 and/or the terminal device 130 may be implemented on the computing device 200 … computing device 200 may include a processor 210, a storage 220, an input/output (I/O) 230, and a communication port 240… I/O 230 may input and/or output signals… input information received through the input device may be transmitted to another component (e.g., the processing device 120) via, for example, a bus, for further processing… communication port 240 may establish connections between the processing device 120 and one or more components (e.g., the scanning device 110, the storage device 150, and/or the terminal device 130) of the medical system 100… processing device 120 may also include a transmission module (not shown) configured to transmit signals (e.g., electrical signals, electromagnetic signals) to one or more components (e.g., the scanning device 110, the terminal device 130, the storage device 150) of the medical system 100) to select one of the plurality of machine learning modules (CAO, Par. [0103-107]: determine a quality feature of the updated image… The quality feature may include a noise feature… the updated image may include different types of interference information (e.g., the noise information, the artifact information). In order to eliminate the different types of interference information as much as possible, the processing device 120 may determine an optimizing model corresponding to each type of interference information… determine a noise optimizing model for reducing or eliminating the noise information of the updated image), based on a desired noise reduction level (CAO, Par. [0086-103]: the optimized image in the current iteration step has reached a desired image quality (e.g., a noise rate is less than a threshold)… determine a quality feature of the updated image… The quality feature may include a noise feature… determine a noise optimizing model for reducing or eliminating the noise information of the updated image).
Regarding claim 5, claim 1 is incorporated and the combination of CAO and RA, as a whole, teaches the system (CAO, Par. [0002]), wherein the at least one machine learning module includes an artificial neural network model (CAO, Par. [0102-108]: optimizing model may include a machine learning model, for example, a neural network model… the plurality of optimizing models may be deep neural network models with different numbers of neural networks or/and different numbers of neurons. As another example, the plurality of optimizing models may be deep neural network models with different activation modes and/or different model structures).
Regarding claim 6, claim 5 is incorporated and the combination of CAO and RA, as a whole, teaches the system (CAO, Par. [0002]), wherein the artificial neural network model is a convolutional neural network (CAO, Par. [0102-108]: the optimizing model may include a machine learning model, for example, a neural network model. The neural network model may include… a convolutional neural network (CNN) model)
Regarding claim 8, claim 1 is incorporated and the combination of CAO and RA, as a whole, teaches the system (CAO, Par. [0002]), wherein the measured projection data is measured by an imaging apparatus (CAO, Par. [0053]: systems and methods for non-invasive biomedical imaging/treatment, such as for disease diagnostic, disease therapy, or research purposes… the systems may include an imaging system; Par. [0094]: originally acquired projection data associated with the initial image may refer to original projection data acquired by the scanning device 110… obtain the originally acquired projection data from the scanning device 110 or a storage device), the imaging apparatus comprising one of: an X-ray imaging apparatus (CAO, Par. [0053]: system may include, for example… an X-ray imaging system), a positron emission tomography (PET) (Par. [0053-59]: a positron emission tomography (PET) system… scanning device 110 may include… a positron emission tomography (PET) device)/single photon emission computed tomography (SPECT) imaging apparatus (Par. [0059]: a scanning device 110 may include… a single-photon emission computed tomography (SPECT) device)/, or a magnetic resonance imaging (MRI) imaging apparatus Par. [0053-59]: system may include, for example… a magnetic resonance imaging (MRI) system… a scanning device 110 may include… a magnetic resonance imaging (MRI) device).
Regarding claim 9, claim 2 is incorporated and the combination of CAO and RA, as a whole, teaches the system (CAO, Par. [0002]), wherein the quality pertains to at least one of: noise (CAO, Par. [0095-103]: determine the first quality weight based on interference information (e.g., noise information… quality feature may include a noise feature), resolution (Par. [0103]: quality feature may include… a resolution) or image artifact (Par. [0095-103]: determine the first quality weight based on interference information (e.g., noise information, artifact information)… quality feature may include… an artifact feature).
Regarding claim 10, claim 1 is incorporated and the combination of CAO and RA, as a whole, teaches the system (CAO, Par. [0002]) including a system for iterative reconstruction of imagery (CAO, Par. [0002-3]: systems and methods for iterative image reconstruction… CT image can be iteratively reconstructed; Par. [0115]: processing device 120 (e.g., the reconstruction module 420) (e.g., the processing circuits of the processor 210) may receive an image to be processed in the current iterative reconstruction… the image to be processed is an optimized image determined in a previously adjacent iterative reconstruction) in the image domain from the measured projection data in the projection domain (CAO, Par. [0092-96]: processing device 120 (e.g., the reconstruction module 420) (e.g., the processing circuits of the processor 210) may determine an updated image… processing device 120 may determine forward projection data by performing a forward projection transformation on the image to be processed in the iteration step. According to the forward projection transformation… transform data (e.g., the image to be processed) in an image domain to data (e.g., the forward projection data) in a projection domain), the system comprising:
memory on which is stored the current image for the given iteration step of the plurality of steps (CAO, Par. [0073-74]: FIG. 3 is a schematic diagram illustrating exemplary hardware and/or software components of an exemplary mobile device according to some embodiments of the present disclosure… one or more components (e.g., the terminal device 130, the processing device 120) of the medical system 100 may be implemented on one or more components of the mobile device 300… the mobile device 300 may include a communication platform 310, a display 320, a graphics processing unit (GPU) 330, a central processing unit (CPU) 340, an 1/O 350, a memory 360, and a storage 390);
a forward-projector to map the current image into the projection domain to obtain the estimated projection data (CAO, Par. [0093]: determine forward projection data by performing a forward projection transformation on the image to be processed in the iteration step. According to the forward projection transformation, the processing device 120 may transform data (e.g., the image to be processed) in an image domain to data (e.g., the forward projection data) in a projection domain);
a comparator configured to establish the back-projected error projection data as the deviation, if any, between the measured projection data and the estimated projection data (RA, Par. [0028]: image reconstructor may be further configured to correct the first information based on predicted data acquired by forward projecting the predicted third image with respect to the measured data… image reconstructor may be further configured to compare the predicted data with the measured data and to correct the first information such that a difference between the predicted data and the measured data decreases);
a back-projector to back-project the back-projected error projection data into the image domain to obtain the input correction data when there is the deviation (CAO, Par. [0094]: determine the updated image based at least in part on back projection data of a weighted error between the forward projection data and originally acquired projection data associated with the initial image);
a correction data determiner, implemented as the trained at least one machine learning module of claim 1 (CAO, Par. [0101-102]: determine the optimized image by reducing interference information of the updated image based on an optimizing model or an optimizing algorithm… optimizing model may be pre-trained and stored in a storage device… the optimizing model may be trained based on a plurality of training samples), to provide the output correction image data (CAO, Par. [0099-101]: determine an optimized image by reducing (or eliminating) interference information of the updated image… determine the optimized image by reducing interference information of the updated image based on an optimizing model or an optimizing algorithm); and
an updater configured to apply the output correction data to the current image to update the current image into the new image (CAO, Par. [0099-109]: determine an optimized image by reducing (or eliminating) interference information of the updated image… determine the optimized image by reducing interference information of the updated image based on an optimizing model or an optimizing algorithm… the updated image may include different types of interference information (e.g., the noise information… the processing device 120 may determine a noise optimizing model for reducing or eliminating the noise information of the updated image… processing device 120 may execute the plurality of interference information reduction operations on the updated image sequentially. For example, the processing device 120 may designate a result image obtained in the current interference information reduction operation as a next updated image in a next interference information reduction operation. Further, the processing device 120 may designate a result image obtained in the last interference information reduction operation as the optimized image).
The same motivation to combine above-mentioned teachings applies, as previously indicated in claim 1.
Regarding claim 11, claim 1 is incorporated and the combination of CAO and RA, as a whole, teaches the system (CAO, Par. [0002]) including a system for iterative reconstruction of imagery (CAO, Par. [0002-3]: systems and methods for iterative image reconstruction… CT image can be iteratively reconstructed; Par. [0115]: processing device 120 (e.g., the reconstruction module 420) (e.g., the processing circuits of the processor 210) may receive an image to be processed in the current iterative reconstruction… the image to be processed is an optimized image determined in a previously adjacent iterative reconstruction) in the image domain from the measured projection data in the projection domain (CAO, Par. [0092-96]: processing device 120 (e.g., the reconstruction module 420) (e.g., the processing circuits of the processor 210) may determine an updated image… processing device 120 may determine forward projection data by performing a forward projection transformation on the image to be processed in the iteration step. According to the forward projection transformation… transform data (e.g., the image to be processed) in an image domain to data (e.g., the forward projection data) in a projection domain), comprising:
memory on which is stored the current image for the given iteration step of the plurality of steps (CAO, Par. [0073-74]: FIG. 3 is a schematic diagram illustrating exemplary hardware and/or software components of an exemplary mobile device according to some embodiments of the present disclosure… one or more components (e.g., the terminal device 130, the processing device 120) of the medical system 100 may be implemented on one or more components of the mobile device 300… the mobile device 300 may include a communication platform 310, a display 320, a graphics processing unit (GPU) 330, a central processing unit (CPU) 340, an 1/O 350, a memory 360, and a storage 390);
a forward-projector to map the current image into the projection domain to obtain the estimated projection data (CAO, Par. [0093]: determine forward projection data by performing a forward projection transformation on the image to be processed in the iteration step. According to the forward projection transformation, the processing device 120 may transform data (e.g., the image to be processed) in an image domain to data (e.g., the forward projection data) in a projection domain);
a comparator configured to establish the input correction data as the deviation, if any, between the measured projection data and the estimated projection data (RA, Par. [0028]: image reconstructor may be further configured to correct the first information based on predicted data acquired by forward projecting the predicted third image with respect to the measured data… image reconstructor may be further configured to compare the predicted data with the measured data and to correct the first information such that a difference between the predicted data and the measured data decreases);
a correction data determiner, implemented as the trained at least one machine learning module of claim 1 (CAO, Par. [0101-102]: determine the optimized image by reducing interference information of the updated image based on an optimizing model or an optimizing algorithm… optimizing model may be pre-trained and stored in a storage device… the optimizing model may be trained based on a plurality of training samples), to provide the output correction data when there is the deviation;
a back-projector configured to back-project the output correction data into the image domain (CAO, Par. [0094-97]: determine the updated image based at least in part on back projection data of a weighted error between the forward projection data and originally acquired projection data associated with the initial image… determine the back projection data of the weighted error by performing a backward projection transformation on the weighted error. According to the back projection transformation, the processing device 120 may transform data (e.g., the weighted error) in a projection domain to data (e.g., the back projection data) in an image domain… determine the updated image based at least in part on the back projection data of the weighted error between the forward projection data and originally acquired projection data associated with the initial image; and
an updater configured to apply the correction data to the current image to update the current image into the new image (CAO, Par. [0099-109]: determine an optimized image by reducing (or eliminating) interference information of the updated image… determine the optimized image by reducing interference information of the updated image based on an optimizing model or an optimizing algorithm… the updated image may include different types of interference information (e.g., the noise information… the processing device 120 may determine a noise optimizing model for reducing or eliminating the noise information of the updated image… processing device 120 may execute the plurality of interference information reduction operations on the updated image sequentially. For example, the processing device 120 may designate a result image obtained in the current interference information reduction operation as a next updated image in a next interference information reduction operation. Further, the processing device 120 may designate a result image obtained in the last interference information reduction operation as the optimized image).
The same motivation to combine above-mentioned teachings applies, as previously indicated in claim 1.
Regarding claim 12, claim 1 is incorporated and the combination of CAO and RA, as a whole, teaches the system (CAO, Par. [0002]) including a training system configured to train, based on training data, the at least one machine learning module of claim 1 (CAO, Par. [0030]: system may further include a training module. The training module may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples. Each of the plurality of training samples may include a sample image and a sample quality feature of the sample image. A loss function of the optimizing model may be positively related to a second quality weight. The second quality weight may be determined based on the sample quality feature; Par. [0079]: training module 430 may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples; Par. [0130]: the processing device 120 (e.g., the training module 430) (e.g., the processing circuits of the processor 210) may obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples).
Regarding claim 13, claim 12 is incorporated and the combination of CAO and RA, as a whole, teaches the system (CAO, Par. [0002]) including a a system for generating the training data for use in the training system of claim 12 (CAO, Par. [0030]: system may further include a training module. The training module may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples. Each of the plurality of training samples may include a sample image and a sample quality feature of the sample image. A loss function of the optimizing model may be positively related to a second quality weight. The second quality weight may be determined based on the sample quality feature; Par. [0079]: training module 430 may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples; Par. [0130]: the processing device 120 (e.g., the training module 430) (e.g., the processing circuits of the processor 210) may obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples), the system comprising an iterative reconstructor (CAO, Par. [0002]: systems and methods for iterative image reconstruction) configured to:
i) process projection data to obtain correction image data at a first quality and a second quality, the second quality being higher than the first quality, the correction image data at the first and second qualities being provided as the training data for input and training target in the training system (CAO, Par. [0030]: the system may further include a training module. The training module may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples. Each of the plurality of training samples may include a sample image and a sample quality feature of the sample image. A loss function of the optimizing model may be positively related to a second quality weight. The second quality weight may be determined based on the sample quality feature), wherein the processing includes performing iterative reconstructions using different amounts (i.e. weights) of the projection data to obtain the correction image data at the first and second qualities (CAO, Par. [0095]: processing device 120 may determine a first quality weight associated with the originally acquired projection data. For example, the processing device 120 may determine the first quality weight based on interference information (e.g., noise information, artifact information) in the originally acquired projection data… determine the weighted error between the forward projection data and originally acquired projection data based on the first quality weight, the forward projection data, and the originally acquired projection data. For example, the processing device 120 may determine a difference between the forward projection data and originally acquired projection data as an error between the forward projection data and originally acquired projection data. Further, the processing device 120 may determine the weighted error based on the error and the first quality weight… the processing device 120 may determine a quality feature of the image to be processed in the iteration step and determine a second quality weight based on the quality feature… Further, the processing device 120 may determine the weighted error based on the error and the second quality weight), wherein the correction image data is based on respective back-projected error projection data (CAO, Par. [0096]: determine the back projection data of the weighted error by performing a backward projection transformation on the weighted error… transform data (e.g., the weighted error) in a projection domain to data (e.g., the back projection data) in an image domain… determine the updated image based on the back projection data of the weighted error), the respective back-projected error projection data representing a respective deviation (i.e. difference) between respective measured projection data (i.e. originally acquired) and respective estimated projection data (i.e. forward projection data), and the respective estimated projection data being obtained by respective forward-projection into projection domain of a respective current image (Par. [0094-95]: determine the updated image based at least in part on back projection data of a weighted error between the forward projection data and originally acquired projection data associated with the initial image… the originally acquired projection data associated with the initial image may refer to original projection data acquired by the scanning device… determine a first quality weight associated with the originally acquired projection data… determine the first quality weight based on interference information (e.g., noise information, artifact information) in the originally acquired projection data… determine a difference between the forward projection data and originally acquired projection data as an error between the forward projection data and originally acquired projection data); or
ii) process projection data to obtain correction projection data at a first quality and a second quality (CAO, Par. [0095]: processing device 120 may determine a first quality weight associated with the originally acquired projection data. For example, the processing device 120 may determine the first quality weight based on interference information (e.g., noise information, artifact information) in the originally acquired projection data… determine the weighted error between the forward projection data and originally acquired projection data based on the first quality weight, the forward projection data, and the originally acquired projection data. For example, the processing device 120 may determine a difference between the forward projection data and originally acquired projection data as an error between the forward projection data and originally acquired projection data. Further, the processing device 120 may determine the weighted error based on the error and the first quality weight… the processing device 120 may determine a quality feature of the image to be processed in the iteration step and determine a second quality weight based on the quality feature… Further, the processing device 120 may determine the weighted error based on the error and the second quality weight), the correction projection data at the first and a second qualities being provided as the training data for the input and training target in the training system (CAO, Par. [0030]: the system may further include a training module. The training module may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples. Each of the plurality of training samples may include a sample image and a sample quality feature of the sample image. A loss function of the optimizing model may be positively related to a second quality weight. The second quality weight may be determined based on the sample quality feature), wherein the processing includes performing iterative reconstructions using different amounts (i.e. weights) of the projection data to obtain the correction projection data, based on the respective deviation between the respective measured projection data and the respective estimated projection data (CAO, Par. [0095]: processing device 120 may determine a first quality weight associated with the originally acquired projection data. For example, the processing device 120 may determine the first quality weight based on interference information (e.g., noise information, artifact information) in the originally acquired projection data… determine the weighted error between the forward projection data and originally acquired projection data based on the first quality weight, the forward projection data, and the originally acquired projection data. For example, the processing device 120 may determine a difference between the forward projection data and originally acquired projection data as an error between the forward projection data and originally acquired projection data. Further, the processing device 120 may determine the weighted error based on the error and the first quality weight… the processing device 120 may determine a quality feature of the image to be processed in the iteration step and determine a second quality weight based on the quality feature… Further, the processing device 120 may determine the weighted error based on the error and the second quality weight), the respective estimated projection data being obtained by respective forward-projection into the projection domain of a respective current image reconstructed at a given respective iteration step (CAO, Par. [0094-95]: determine the updated image based at least in part on back projection data of a weighted error between the forward projection data and originally acquired projection data associated with the initial image… the originally acquired projection data associated with the initial image may refer to original projection data acquired by the scanning device… determine a first quality weight associated with the originally acquired projection data… determine the first quality weight based on interference information (e.g., noise information, artifact information) in the originally acquired projection data… determine a difference between the forward projection data and originally acquired projection data as an error between the forward projection data and originally acquired projection data).
Regarding claim 14, CAO discloses a method for facilitating an iterative reconstruction operation (Par. [0002]: systems and methods for iterative image reconstruction). The steps further recited in method claim 14 correspond to claim 1 when executed and are rejected as applied to apparatus claim 1 above.
Regarding claim 15, claim 1 is incorporated and the combination of CAO and RA, as a whole, teaches the system (CAO, Par. [0002]) including a method of training the at least one machine learning module as claimed in of claim 1 based on training data (CAO, Par. [0030]: system may further include a training module. The training module may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples. Each of the plurality of training samples may include a sample image and a sample quality feature of the sample image. A loss function of the optimizing model may be positively related to a second quality weight. The second quality weight may be determined based on the sample quality feature; Par. [0079]: training module 430 may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples; Par. [0130]: the processing device 120 (e.g., the training module 430) (e.g., the processing circuits of the processor 210) may obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples).
Regarding claim 16, CAO discloses a method (Par. [0002]: systems and methods for iterative image reconstruction) of generating training data for use in a training a machine learning model (CAO, Par. [0030]: system may further include a training module. The training module may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples. Each of the plurality of training samples may include a sample image and a sample quality feature of the sample image. A loss function of the optimizing model may be positively related to a second quality weight. The second quality weight may be determined based on the sample quality feature; Par. [0079]: training module 430 may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples; Par. [0130]: the processing device 120 (e.g., the training module 430) (e.g., the processing circuits of the processor 210) may obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples), the method comprising:
i) processing projection data to obtain correction image data at a first quality and a second quality, the second quality being higher than the first quality (CAO, Par. [0030]: the system may further include a training module. The training module may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples. Each of the plurality of training samples may include a sample image and a sample quality feature of the sample image. A loss function of the optimizing model may be positively related to a second quality weight. The second quality weight may be determined based on the sample quality feature), wherein the processing includes performing iterative reconstructions using different amounts (i.e. weights) of the projection data to obtain the correction image data at the first and second qualities (CAO, Par. [0095]: processing device 120 may determine a first quality weight associated with the originally acquired projection data. For example, the processing device 120 may determine the first quality weight based on interference information (e.g., noise information, artifact information) in the originally acquired projection data… determine the weighted error between the forward projection data and originally acquired projection data based on the first quality weight, the forward projection data, and the originally acquired projection data. For example, the processing device 120 may determine a difference between the forward projection data and originally acquired projection data as an error between the forward projection data and originally acquired projection data. Further, the processing device 120 may determine the weighted error based on the error and the first quality weight… the processing device 120 may determine a quality feature of the image to be processed in the iteration step and determine a second quality weight based on the quality feature… Further, the processing device 120 may determine the weighted error based on the error and the second quality weight), wherein the correction image data is based on respective back-projected error projection data (CAO, Par. [0096]: determine the back projection data of the weighted error by performing a backward projection transformation on the weighted error… transform data (e.g., the weighted error) in a projection domain to data (e.g., the back projection data) in an image domain… determine the updated image based on the back projection data of the weighted error), t the respective back-projected error projection data representing a respective deviation (i.e. difference) between respective measured projection data (i.e. originally acquired) and respective estimated projection data (i.e. forward projection data), and the respective estimated projection data being obtained by respective forward-projection into projection domain of a respective current image (Par. [0094-95]: determine the updated image based at least in part on back projection data of a weighted error between the forward projection data and originally acquired projection data associated with the initial image… the originally acquired projection data associated with the initial image may refer to original projection data acquired by the scanning device… determine a first quality weight associated with the originally acquired projection data… determine the first quality weight based on interference information (e.g., noise information, artifact information) in the originally acquired projection data… determine a difference between the forward projection data and originally acquired projection data as an error between the forward projection data and originally acquired projection data); and
providing the correction image data at the first and second qualities as training data for input and training target in a training system (CAO, Par. [0030]: the system may further include a training module. The training module may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples. Each of the plurality of training samples may include a sample image and a sample quality feature of the sample image. A loss function of the optimizing model may be positively related to a second quality weight. The second quality weight may be determined based on the sample quality feature), or
ii) processing the projection data to obtain correction projection data at a first quality and a second quality (CAO, Par. [0095]: processing device 120 may determine a first quality weight associated with the originally acquired projection data. For example, the processing device 120 may determine the first quality weight based on interference information (e.g., noise information, artifact information) in the originally acquired projection data… determine the weighted error between the forward projection data and originally acquired projection data based on the first quality weight, the forward projection data, and the originally acquired projection data. For example, the processing device 120 may determine a difference between the forward projection data and originally acquired projection data as an error between the forward projection data and originally acquired projection data. Further, the processing device 120 may determine the weighted error based on the error and the first quality weight… the processing device 120 may determine a quality feature of the image to be processed in the iteration step and determine a second quality weight based on the quality feature… Further, the processing device 120 may determine the weighted error based on the error and the second quality weight), wherein the processing includes performing iterative reconstructions using different amounts (i.e. weights) of the projection data to obtain the correction projection data, based on the respective deviation between the respective measured projection data and the respective estimated projection data (CAO, Par. [0095]: processing device 120 may determine a first quality weight associated with the originally acquired projection data. For example, the processing device 120 may determine the first quality weight based on interference information (e.g., noise information, artifact information) in the originally acquired projection data… determine the weighted error between the forward projection data and originally acquired projection data based on the first quality weight, the forward projection data, and the originally acquired projection data. For example, the processing device 120 may determine a difference between the forward projection data and originally acquired projection data as an error between the forward projection data and originally acquired projection data. Further, the processing device 120 may determine the weighted error based on the error and the first quality weight… the processing device 120 may determine a quality feature of the image to be processed in the iteration step and determine a second quality weight based on the quality feature… Further, the processing device 120 may determine the weighted error based on the error and the second quality weight), the respective estimated projection data being obtained by the respective forward-projection into the projection domain of a respective current image reconstructed at a given respective iteration step (CAO, Par. [0094-95]: determine the updated image based at least in part on back projection data of a weighted error between the forward projection data and originally acquired projection data associated with the initial image… the originally acquired projection data associated with the initial image may refer to original projection data acquired by the scanning device… determine a first quality weight associated with the originally acquired projection data… determine the first quality weight based on interference information (e.g., noise information, artifact information) in the originally acquired projection data… determine a difference between the forward projection data and originally acquired projection data as an error between the forward projection data and originally acquired projection data); and
providing the correction projection data at the first and second qualities as training data for the input and training target in the training system (CAO, Par. [0030]: the system may further include a training module. The training module may be configured to obtain a plurality of training samples and obtain the optimizing model by training a preliminary optimizing model based on the plurality of training samples. Each of the plurality of training samples may include a sample image and a sample quality feature of the sample image. A loss function of the optimizing model may be positively related to a second quality weight. The second quality weight may be determined based on the sample quality feature).
Regarding claim 17, claim 14 is incorporated and the combination of CAO and RA, as a whole, teaches the system (CAO, Par. [0002]) including at least one non-transitory computer readable medium storing instructions, which, when executed by at least one processing unit, cause the at least one processing unit to perform the method as claimed in claim 14 (CAO, Par. [0031]: a non-transitory computer readable medium including executable instructions. When the executable instructions are executed by at least one processor, the executable instructions may direct the at least one processor to perform a method. The method may include obtaining an initial image to be processed and generating a reconstructed image by performing a plurality of iteration steps on the initial image. Each of the plurality of iteration steps may include a first optimization operation and at least one second optimization operation… first optimization operation may include receiving an image to be processed in the iteration step and determining an updated image by preliminarily optimizing the image to be processed… second optimization operation may include determining an optimized image by reducing interference information of the updated image and designating the optimized image as a next image to be processed in a next iteration step. The interference information may include noise information and/or artifact information).
Regarding claim 18, claim 17 is incorporated and the combination of CAO and RA, as a whole, teaches the computer readable medium (CAO, Par. [0031]) further storing the training trained machine learning module model (CAO, Par. [0102]: optimizing model may be pre-trained and stored in a storage device… the optimizing model may be trained based on a plurality of training samples; Par. [0126]: the processing device 120 (e.g., the training module 430) (e.g., the processing circuits of the processor 210) may obtain a plurality of training samples).
Regarding claim 19, claim 14 is incorporated and the combination of CAO and RA, as a whole, teaches the method (CAO, Par. [0002]) wherein the machine learning model improves data quality so that the output correction data has a higher quality than the input correction data (CAO, Par. [0003]: systems and methods for image reconstruction with improved image quality; [0009-10]: determining a quality feature of the updated image… the quality feature may include a noise feature; Par. [0057]: a reconstructed image is generated by performing a plurality of iteration steps each of which includes a first optimization operation (which is used to preliminarily optimize the image to be processed) and at least one second optimization operation (which is used to further optimize an updated image generated in the first operation) implemented via a machine learning model, that is, a preliminary optimization and a further optimization via a machine learning model are used in combination, which can improve the image quality of the reconstructed image and reduce the noise of the reconstructed image; Par. [0085-86]: optimization operation may include determining an optimized image by reducing interference information (e.g., noise information, artifact information) of the updated image and designating the optimized image as a next image to be processed in a next iteration step… the optimized image by reducing the interference information based on an optimizing model (e.g., a machine learning model)… determine whether a termination condition is satisfied in the current iteration step. Exemplary termination conditions may include… the optimized image in the current iteration step has reached a desired image quality (e.g., a noise rate is less than a threshold)).
Regarding claim 20, claim 16 is incorporated and the combination of CAO and RA, as a whole, teaches the method (CAO, Par. [0002]) including at least one non-transitory computer readable medium storing instructions, that, when executed by at least one processing unit, cause the at least one processing unit to perform the method of claim 16 (CAO, Par. [0031]: a non-transitory computer readable medium including executable instructions. When the executable instructions are executed by at least one processor, the executable instructions may direct the at least one processor to perform a method. The method may include obtaining an initial image to be processed and generating a reconstructed image by performing a plurality of iteration steps on the initial image. Each of the plurality of iteration steps may include a first optimization operation and at least one second optimization operation… first optimization operation may include receiving an image to be processed in the iteration step and determining an updated image by preliminarily optimizing the image to be processed… second optimization operation may include determining an optimized image by reducing interference information of the updated image and designating the optimized image as a next image to be processed in a next iteration step. The interference information may include noise information and/or artifact information).
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over CAO, in view of RA, as applied to claim 1 above, in further view of LU et al. (US PG Publication No. 2021/0007695 A1), hereafter referred to as LU.
Regarding claim 7, claim 6 is incorporated and the combination of CAO and RA, as a whole, teaches the system (CAO, Par. [0002]), wherein the convolutional neural network has an effective receptive field larger than a neighborhood (Par. [0035]: a filter using a total-variation (TV) minimization regularization term can be applied if imaged region supports an assumption of uniformity over large areas; Par. [0073]: a non-limiting example in which the DL network 170 is a convolutional neural network (CNN)… For example, CNNs can be used for image-processing optimization by using multiple layers of small neuron collections which process portions of the input image, called receptive fields. The outputs of these collections can then tiled so that they overlap, to obtain a better representation of the original image) of a voxel in the image domain or of a pixel in the projection domain (Par. [0035]: a filter using a total-variation (TV) minimization regularization term can be applied if imaged region supports an assumption of uniformity over large areas; Par. [0048-54]: disparate physical models can be integrated into the data fidelity term of the objective function in both the image domain… and the projection domain… For example, the initial reconstructed image… can be used to estimate the scatter contribution to each of the pixel values in the projection data custom-character… For example, the initial reconstructed image… can be segmented into material components (e.g., using a material decomposition or by mapping the Hounsfield Units in the respective voxels to respective material components… material decomposition can be used to separate the attenuation in the respective voxels into material components, which are then used for the forward projection to account for beam-hardening corrections… the physical models can include an advanced footprint method for precise forward projection… to account for the system geometry… The forward projection model… can be performed by mapping pixel boundaries and detector boundaries to a common axis, and then applying a kernel operation to map data from one set of boundaries to another… Based on these boundaries, the length of overlap is calculated between each image pixel and each detector cell, and length of overlap is used to normalize the weight used in projection and backprojection. This corresponds to applying the distance-driven kernel operation to the mapped boundary locations, which is achieved by performing a loop over all boundaries; Par. [0073]: a non-limiting example in which the DL network 170 is a convolutional neural network (CNN)… For example, CNNs can be used for image-processing optimization by using multiple layers of small neuron collections which process portions of the input image, called receptive fields. The outputs of these collections can then tiled so that they overlap, to obtain a better representation of the original image).
CAO, RA and LU are considered to be analogous art because they pertain to because they pertain to medical image processing applications. Therefore, the combined teachings of CAO, RA and LU, as a whole, would have rendered obvious the invention recited in claim 7 with a reasonable expectation of success in order to modify the systems and methods for iterative image reconstruction of a voxel in the image domain or of a pixel in the projection domain (as disclosed by CAO) with wherein the convolutional neural network has an effective receptive field larger than a neighborhood (as taught by LU, Abstract, Par. [0035, 48-54, 73, ) to improve the image quality based on the physical model (LU, Abstract, Par. [00 2-7, 16-19, 31, 35, 39 48-54, 73]).
Conclusion
Applicant’s amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any extension fee pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Contact
Any inquiry concerning this communication or earlier communications from the examiner should be directed to GUILLERMO RIVERA-MARTINEZ whose telephone number is 571-272-4979. The examiner can normally be reached on Monday-Friday (8am - 5pm Eastern Time). If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Andrew Bee can be reached on 571-270-5183. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/GUILLERMO M RIVERA-MARTINEZ/ Primary Examiner, Art Unit 2677
Prosecution Insights