DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendments
The amendments to claims 1-10 and 12 are accepted and entered.
The amendment to FIG. 7 of the Drawings is accepted and entered.
The amendments to the specification are accepted and entered.
Claims 1-12 are pending regarding this application.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 05/10/2024 and 06/12/2024 are considered and attached.
Claim Objections
Claims 3 and 6 are objected to because of the following informalities:
Claim 2 recites “the m-th first patch” in line 11. However, no prior reference to the variable “m” is recited prior to the above recitation. As such, please amend to recite “a m-th first patch”.
Claim 6 recites “a number of the parts” in line 2, “a number of the parts” in line 3. Please amend line 3 to recite “the number of the parts”.
Claim 6 recites “a ratio of the first number of patches to the second number of patches” in lines 8-9 and “a ratio of the first number to the second number” in line 9. Each instance “ratio” clearly refers to two separate and distinct ratios. As such, please amend to recite “a first ratio of the first number of patches to the second number of patches” in lines 8-9 and “a second ratio of the first number to the second number” in line 9.
As a result, the claims will be analyzed below assuming these changes were made. Appropriate correction is required.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-12 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 1 recites “at least one first patch” in line 5, “the first patch” in line 6, and “the first patches” in line 7. Claims 2 and 4 recite “the m-th first patch”, in line 11 and lines 3-4, respectively. As such, it is unclear how “at least one patch” can be both “the first patch” and “the first patches” throughout claim 1, and claims 2-10, which depend upon claim 1. For example, how is it possible that the first patch can be stored by a storage unit, as recited in line 6 of claim 1, but then any of the first patches from the storage unit are read out as claimed in lines 7-9 of claim 1. Here, although only one patch is stored in the storage unit, multiple first patches are then referenced as residing in the storage unit, which presents an unclear recitation of whether one or more first patched are stored in the storage unit. Applicant discusses the first patch(es) throughout the specification. However, nowhere in the specification does the applicant clarify whether the references to “the first patch”, “the first patches”, “at least one patch”, or “the m-th first patch” (see claims 2 and 4) are equivalent or distinct. Similarly, instances of “the first patch”, “the first patches”, and “the m-th first patch” as recited in claims 1, 2, 5, 6, 7, 8, and 9 should be amended to create consistency between the first recitation of “at least one first patch” and each subsequent reference to the “at least one first patch”. If applicant wishes to claim a plurality of first patches, applicant must amend to introduce an initial instance of a plurality of first patches. Therefore, claims 1, 2, 5, 6, 7, 8, and 9 fail to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor, regards as the invention.
Similar analysis can be applied to corresponding independent claims 11 and 12. Claims 2-10 are rejected due to their dependency upon rejected claim 1.
Claim 2 recites “a parameter for controlling randomness of a patch” in line 4. However, at least one first patch and at least one second patch are already introduced in claim 1. Therefore, it is unclear whether the patch as recited in line 4 of claim 2 is equivalent to or distinct from the at least one first patch and/or the at least second patch as claimed in claim 1. Applicant discusses the parameter for controlling randomness in para. [0053]. However, this section does not discuss the parameter in the context of any specific patch. Therefore, it is unclear what patch is being referenced in line 4 of claim 2.
Furthermore, claim 2 recites “when the three-dimensional data or a subset of the three-dimensional data is denoted by X’” in lines 7-8. However, claim 1 recites “at least one first patch being a subset of the three-dimensional data”. As such, it is unclear whether the subset of the three-dimensional data as claimed in claim 2 is equivalent to or distinct from the subset of three-dimensional data as claimed in claim 1. Furthermore, it is unclear whether X’ is equivalent to or distinct from the at least one first patch. Applicant’s specification discusses X’ in para. [0056]-[0065]. Applicant’s specification discusses the subset of three-dimensional data in para. [0028] and [0056]. However, none of these sections clarify whether the subset of three-dimensional data as claimed in both claim 1 and claim 2 are equivalent or distinct. As such, claim 2 fails to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor, regards as the invention.
Claims 3 and 4 are additionally rejected due to their dependency upon rejected claim 2.
Claim 4 recites “a parameter of the second patch” in line 4. Claim 2, upon which claim 4 depends, recites “a parameter for controlling randomness of a patch” in line 4. Here, it is unclear whether the parameter introduced in claim 2 is equivalent to or distinct from the parameter claimed in line 4 of claim 4. Applicant’s specification discusses a parameter in para. [0053]-[0054] and para. [0056]-[0065]. However, none of these sections clarify a distinction between the parameter as described in claim 2 and the parameter as described in claim 4. Para. [0053]-[0054] discusses a parameter for controlling randomness of a patch and para. [0056]-[0065] discusses a parameter of the second patch. However, it is not clear from these sections whether the many different parameters as discussed in both sections may be equivalent or distinct. As such, claim 4 fails to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor, regards as the invention.
Claim 5 recites “the three-dimensional data include labels set by a plurality of parts in the object” in line 3. However, para. [0025] of applicant’s specification recites “Further, the three-dimensional data may include labels set by a plurality of parts in the object. For example, when the three-dimensional data are point cloud data, a label is set for each of a plurality of points composing the point cloud data”. Here, it is unclear how the “plurality of parts in the object” can set labels included in the three-dimensional data. Furthermore, it seems that the recitation of “three-dimensional data include labels set by a plurality of parts in the object” in claim 5 should instead recite “the three-dimensional data include labels set for a plurality of parts in the object” (emphasis added), since the parts (or points) themselves are not performing the act of setting labels. However, it remains unclear what applicant is attempting to claim in the aforementioned subject matter. As such, claim 5 fails to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor, regards as the invention. In the prior art rejection below, claim 5 is being interpreted as though the following amendment has been made: “the three-dimensional data include labels set for a plurality of parts in the object” (emphasis amended).
Claim 6 is additionally rejected due to its dependency upon rejected claim 5.
Claim 6 recites “a first number of the parts having the first label” in line 2 and “a second number being a number of the parts having the second label” in lines 2-3. Both “the first label” and “the second label” lack antecedent basis. Although the term “labels” is introduced in claim 5, upon which claim 6 depends, it remains unclear whether the first/second label in claim 6 is equivalent to/distinct from/a subset of the “labels” as claimed in claim 5. As such, there is insufficient antecedent basis for these limitations in the claim.
Additionally, claim 6 recites “a first number being a number of the parts having the first label”, “a second number being a number of the parts having the second label”, “a number of the first patches”, “a first number of patches”, “a number of the first patches”, and “a second number of patches” in lines 2-6. There are multiple instances of “a number”, “a first number”, and “a second number” that results in a lack of clarity regarding the scope of the term “number” in each instance. A suggestion that may resolve this specific issue for this section (please note that 112b issue regarding the first/second label which is not addressed in this suggested amendment) may be to amend lines 1-9 of claim 6 to recite: “wherein a first number of the parts having the first label is less than a second number of the parts having the second label, and when a third number of the first patches including the first label is defined and a fourth number of the first patches including the second label is defined, the at least one processor is configured to execute the instructions to set a ratio of the third number to the fourth number higher than the ratio of the first number to the second number.”.
Claim 9 recites ”a plurality of subsets” in line 5. However, two instances of “a subset” are introduced in claim 1 (see line 5 and line 8). As such, it is unclear whether the plurality of subsets as defined in claim 9 is equivalent to or distinct from multiple instances of “a subset” defined in claim 1. Applicant discusses the above subject matter in para. [0043] in the specification. It seems as though the applicant may intend to imply that the plurality of subsets refers to a plurality of second patches. However, this is never explicitly stated in either the claim or the specification. As such, although the claims will be examined as though the plurality of subsets is equivalent to a plurality of second patches, it remains unclear what applicant wishes to claim by the recitation of “a plurality of subsets”, and if these subsets are equivalent or distinct from either of the previously claimed subsets in claim 1.
Claim 9 further recites “in at least one of distance and shape” in lines 5-6. However, multiple instances of a “shape” are already introduced in claim 1, upon which claim 9 depends. As such, it is unclear whether the “shape” as claimed in claim 9 is equivalent or distinct from the “three-dimensional data indicating a shape of an object” as claimed in lines 4-5 of claim 1 and “evaluating a three-dimensional shape” as claimed in line 11 of claim 1. Applicant’s specification discusses the “shape” in para. [0043]. However, applicant does not clarify whether the “shape” is equivalent or distinct from the “shape” as recited in the aforementioned sections of claim 1. As such, claim 9 fails to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor, regards as the invention. In the prior art rejection below, the “shape” as claimed in claim 9 is interpreted as separate and distinct from the “shape” as recited in the aforementioned sections of claim 1.
Regarding claim 10, line 5 of claim 10 recites “a location” and “an object”. However, “an object” is already introduced in claim 1, upon which claim 10 depends, and “a location” is previously introduced in line 4 of claim 10. As such, it is unclear whether the object as defined in claim 1 is equivalent or distinct from the object claimed in claim 10. Likewise analysis can be applied to the multiple references of “a location” in claim 10. Applicant’s specification discusses a location of abnormality in an object in para. [0047]. That section is the only section which discusses a location of abnormality, and the section provides no further clarity regarding the distinction between the two recitations of the location of abnormality in the object. As such, claim 10 is rejected under 112(b).
Due to the 112(b) rejections of claims 2-4 and 6 a proper art rejection of these claims was unable to be made, as the specification fails to provide proper contextual information to ascertain the scope of claims 2-4 and 6.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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.
Claim(s) 1, 7-9, 11, and 12 are rejected under 35 U.S.C. 103 as being unpatentable over Liu et al. (U.S. Publication No. 2018/0089530 A1), hereinafter Liu, in view of Suleyman et al. (U.S. Publication 2014/0270488 A1), hereinafter Suleyman.
Regarding claim 1, Liu teaches a model training apparatus comprising:
at least one memory configured to store instructions (Liu teaches “the computer program instructions may be stored in a storage device 512 (e.g., magnetic disk) and loaded into memory 510 when execution of the computer program instructions is desired” in para. [0027]); and
at least one processor configured to execute the instructions (Liu teaches “Computer 502 contains a processor 504, which controls the overall operation of the computer 502 by executing computer program instructions which define such operation” in para. [0027]) to:
generate, by using three-dimensional data indicating a shape of an object, at least one first patch being a subset of the three-dimensional data, and (Liu teaches generating a plurality of training images patches (set of voxels) to detect landmarks in 3D medical images in para. [0022] and FIG. 1, wherein the plurality of training image patches of the 3d medical images are interpreted as equivalent to the subset of the three-dimensional data. See also that “the method of FIG. 1 can be performed to implement a first stage of detecting a pose parameters (position, orientation, and scale) for the anatomical object in the medical image” in para. [0026]);
read out any of the first patches from the storage unit, and generates at least one second patch being a subset of the first patch (Liu teaches “the deep neural network is trained directly on voxels sampled from training image patches based on the predetermined sampling pattern, such as the sampling pattern of FIG. 2 or the sampling pattern of FIG. 3” in para. [0022], wherein “the sampling pattern 300 samples pixels within a respective circular (log-polar) image patch centered at each pixel of a medical image” as shown in para. [0019] and FIG. 3. Here, the segmented regions (patches) of the image patch (first patch) in FIG. 3 and described in para. [0019] are interpreted as the claimed second patch(es));
train, with the second patch as training data, a model for evaluating a three-dimensional shape (Liu teaches “at step 104, a subset of voxels in each of a plurality of image patches of the medical image are input to a trained deep neural network classifier based on a sampling pattern” in para. [0016]. This subset may be the second patches as defined in the above citation in para. [0019] and FIG. 3. Para. [0023] further recites that “the output of the hidden layers can be treated as high-level image features and used to train a discriminative classifier for detecting the anatomical object”. See also “the method of FIG. 1 can be performed to implement a first stage of detecting a pose parameters (position, orientation, and scale) for the anatomical object in the medical image” in para. [0026]).
Liu fails to teach causing a storage unit to store the first patch; and, repeat processes of generating the second patch and training the model until a criterion is satisfied.
However, Suleyman teaches causing a storage unit to store the first patch (Suleyman teaches a process of extracting a subsection from an image based on the shape of an object in para. [0080] and FIG. 3, wherein the subsection is interpreted as the claimed first patch. See also para. [0081]-[0089]. Para. [0033]-[0043] discusses a process of storing a characterization of the input images in a database, wherein the characterization includes the subsection (first patch). See also para. [0047]); and,
repeat processes of generating the second patch and training the model until a criterion is satisfied (Suleyman teaches a process of repeatedly extracting a sub-patch (second patch) from a subsection (first patch) in order to train a network until a “requisite number of patches have been added to the overall record” in para. [0079], [0090]-[0091], and FIG. 5. See also FIG. 3 regarding the generation of the subsection (first patch)).
Liu and Suleyman are both considered to be analogous to the claimed invention because they are in the same field of analyzing patches of an image in order to perform a characterization. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Liu to incorporate the teachings of Suleyman and “causing a storage unit to store the first patch; and, repeat processes of generating the second patch and training the model until a criterion is satisfied”. The motivation for doing so would have been that, “by using a plurality of patches from within the same subsection, the training set is increased in size dramatically”, and that “the advantages of creating such patches include the creation of a larger training set which is more focused on the pattern. Accordingly, the training network will be able to learn the characteristics of this pattern from a smaller training set than if the whole images were used”, as suggested by Suleyman in para. [0027] and para. [0079], respectively. Therefore, it would have been obvious to one of ordinary skill at the time the invention was filed to combine Liu with Suleyman to obtain the invention specified in claim 1.
Regarding claim 7, Liu and Suleyman teach the model training apparatus according to claim 1,
wherein the at least one processor is configured to execute the instructions to superpose a predetermined shape on the first patch, and generate the second patch by using a result of the superposition (Liu teaches “the deep neural network is trained directly on voxels sampled from training image patches based on the predetermined sampling pattern” in para. [0022]. See FIGs. 2 and regarding examples of a superimposed sampling pattern which is used to generate the subset of voxels (second patches) from the image patches (first patches) in para. [0022]).
Regarding claim 8, Liu and Suleyman teach the model training apparatus according to claim 1, wherein the at least one processor is configured to execute the instructions to
generate the second patch by selecting a reference point in the first patch, and selecting another part from the first patch according to a predetermined rule from the reference point (Liu teaches “a log-polar sampling pattern 300 having a higher density near the patch center and a lower density when farther away from the patch center is used to select the subset of voxels to input to the deep neural network”, wherein “the sampling pattern 300 samples voxels a respective spherical image patch centered at each voxel of the medical image” as shown in para. [0019]. Here, the subsets of voxels (second patches) are determined based on the sampling pattern involving “a respective spherical image patch centered at each voxel of the medical image” as shown in para. [0019]. As such, each voxel at which an image patch is centered acts as the reference point for the subsequent generation of the second patch).
Regarding claim 9, Liu and Suleyman teach the model training apparatus according to claim 1, wherein the at least one processor is configured to execute the instructions to
generate the second patch by dividing the first patch into a plurality of subsets that are similar to each other in at least one of distance and shape (Liu teaches a method of generating subsets of voxels involving a sampling pattern in which “the grid sampling pattern can be implemented by skipping every n voxels in each direction between each voxel input to the deep neural network” as shown in para. [0018]. As a result of this method, each sampled voxel (subset of the first patch) is similar to each other in distance).
Regarding claim 11, Liu teaches a model training method comprising, by a computer:
generating, by using three-dimensional data indicating a shape of an object, at least one first patch being a subset of the three-dimensional data, and (Liu teaches generating a plurality of training images patches (set of voxels) to detect landmarks in 3D medical images in para. [0022] and FIG. 1, wherein the plurality of training image patches of the 3d medical images are interpreted as equivalent to the subset of the three-dimensional data. See also that “the method of FIG. 1 can be performed to implement a first stage of detecting a pose parameters (position, orientation, and scale) for the anatomical object in the medical image” in para. [0026]);
reading out any of the first patches from the storage unit, and generates at least one second patch being a subset of the first patch (Liu teaches “the deep neural network is trained directly on voxels sampled from training image patches based on the predetermined sampling pattern, such as the sampling pattern of FIG. 2 or the sampling pattern of FIG. 3” in para. [0022], wherein “the sampling pattern 300 samples pixels within a respective circular (log-polar) image patch centered at each pixel of a medical image” as shown in para. [0019] and FIG. 3. Here, the segmented regions (patches) of the image patch (first patch) in FIG. 3 and described in para. [0019] are interpreted as the claimed second patch(es));
training, with the second patch as training data, a model for evaluating a three-dimensional shape (Liu teaches “at step 104, a subset of voxels in each of a plurality of image patches of the medical image are input to a trained deep neural network classifier based on a sampling pattern” in para. [0016]. This subset may be the second patches as defined in the above citation in para. [0019] and FIG. 3. Para. [0023] further recites that “the output of the hidden layers can be treated as high-level image features and used to train a discriminative classifier for detecting the anatomical object”. See also “the method of FIG. 1 can be performed to implement a first stage of detecting a pose parameters (position, orientation, and scale) for the anatomical object in the medical image” in para. [0026]).
Liu fails to teach causing a storage unit to store the first patch; and, wherein the generating the second patch and the training the model are repeated until a criterion is satisfied.
However, Suleyman teaches causing a storage unit to store the first patch (Suleyman teaches a process of extracting a subsection from an image based on the shape of an object in para. [0080] and FIG. 3, wherein the subsection is interpreted as the claimed first patch. See also para. [0081]-[0089]. Para. [0033]-[0043] discusses a process of storing a characterization of the input images in a database, wherein the characterization includes the subsection (first patch). See also para. [0047]); and,
wherein the generating the second patch and the training the model are repeated until a criterion is satisfied (Suleyman teaches a process of repeatedly extracting a sub-patch (second patch) from a subsection (first patch) in order to train a network until a “requisite number of patches have been added to the overall record” in para. [0079], [0090]-[0091], and FIG. 5. See also FIG. 3 regarding the generation of the subsection (first patch)).
Liu and Suleyman are both considered to be analogous to the claimed invention because they are in the same field of analyzing patches of an image in order to perform a characterization. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Liu to incorporate the teachings of Suleyman and “causing a storage unit to store the first patch; and, wherein the generating the second patch and the training the model are repeated until a criterion is satisfied”. The motivation for doing so would have been that, “by using a plurality of patches from within the same subsection, the training set is increased in size dramatically”, and that “the advantages of creating such patches include the creation of a larger training set which is more focused on the pattern. Accordingly, the training network will be able to learn the characteristics of this pattern from a smaller training set than if the whole images were used”, as suggested by Suleyman in para. [0027] and para. [0079], respectively. Therefore, it would have been obvious to one of ordinary skill at the time the invention was filed to combine Liu with Suleyman to obtain the invention specified in claim 11.
Regarding claim 12, Liu teaches a non-transitory computer-readable storage medium that stores a program (Liu teaches that “computer program instructions may be stored in a storage device 512 (e.g., magnetic disk) and loaded into memory 510 when execution of the computer program instructions is desired” in para. [0027]; here, the example of a magnetic disk is an example of a non-transitory computer-readable medium) causing a computer to execute:
generating, by using three-dimensional data indicating a shape of an object, at least one first patch being a subset of the three-dimensional data, and (Liu teaches generating a plurality of training images patches (set of voxels) to detect landmarks in 3D medical images in para. [0022] and FIG. 1, wherein the plurality of training image patches of the 3d medical images are interpreted as equivalent to the subset of the three-dimensional data. See also that “the method of FIG. 1 can be performed to implement a first stage of detecting a pose parameters (position, orientation, and scale) for the anatomical object in the medical image” in para. [0026]);
reading out any of the first patches from the storage unit, and generates at least one second patch being a subset of the first patch (Liu teaches “the deep neural network is trained directly on voxels sampled from training image patches based on the predetermined sampling pattern, such as the sampling pattern of FIG. 2 or the sampling pattern of FIG. 3” in para. [0022], wherein “the sampling pattern 300 samples pixels within a respective circular (log-polar) image patch centered at each pixel of a medical image” as shown in para. [0019] and FIG. 3. Here, the segmented regions (patches) of the image patch (first patch) in FIG. 3 and described in para. [0019] are interpreted as the claimed second patch(es));
training, with the second patch as training data, a model for evaluating a three-dimensional shape (Liu teaches “at step 104, a subset of voxels in each of a plurality of image patches of the medical image are input to a trained deep neural network classifier based on a sampling pattern” in para. [0016]. This subset may be the second patches as defined in the above citation in para. [0019] and FIG. 3. Para. [0023] further recites that “the output of the hidden layers can be treated as high-level image features and used to train a discriminative classifier for detecting the anatomical object”. See also “the method of FIG. 1 can be performed to implement a first stage of detecting a pose parameters (position, orientation, and scale) for the anatomical object in the medical image” in para. [0026]).
Liu fails to teach causing a storage unit to store the first patch; and, repeating processes of generating the second patch and training the model until a criterion is satisfied.
However, Suleyman teaches causing a storage unit to store the first patch (Suleyman teaches a process of extracting a subsection from an image based on the shape of an object in para. [0080] and FIG. 3, wherein the subsection is interpreted as the claimed first patch. See also para. [0081]-[0089]. Para. [0033]-[0043] discusses a process of storing a characterization of the input images in a database, wherein the characterization includes the subsection (first patch). See also para. [0047]); and,
repeating processes of generating the second patch and training the model until a criterion is satisfied (Suleyman teaches a process of repeatedly extracting a sub-patch (second patch) from a subsection (first patch) in order to train a network until a “requisite number of patches have been added to the overall record” in para. [0079], [0090]-[0091], and FIG. 5. See also FIG. 3 regarding the generation of the subsection (first patch)).
Liu and Suleyman are both considered to be analogous to the claimed invention because they are in the same field of analyzing patches of an image in order to perform a characterization. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Liu to incorporate the teachings of Suleyman and “causing a storage unit to store the first patch; and, repeating processes of generating the second patch and training the model until a criterion is satisfied”. The motivation for doing so would have been that, “by using a plurality of patches from within the same subsection, the training set is increased in size dramatically”, and that “the advantages of creating such patches include the creation of a larger training set which is more focused on the pattern. Accordingly, the training network will be able to learn the characteristics of this pattern from a smaller training set than if the whole images were used”, as suggested by Suleyman in para. [0027] and para. [0079], respectively. Therefore, it would have been obvious to one of ordinary skill at the time the invention was filed to combine Liu with Suleyman to obtain the invention specified in claim 12.
Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Liu et al. (U.S. Publication No. 2018/0089530 A1), hereinafter Liu, in view of Suleyman et al. (U.S. Publication 2014/0270488 A1), hereinafter Suleyman and Bengtsson et al. (U.S. Publication No. 2021/0401392 A1), hereinafter Bengtsson.
Regarding claim 5, Liu and Suleyman teach the model training apparatus according to claim 1.
While Liu teaches annotated images (see para. [0021]), Liu and Suleyman fail to teach wherein the three-dimensional data include labels set by a plurality of parts in the object, and the at least one processor is configured to execute the instructions to generate a plurality of the first patches by using the labels.
However, Bengtsson teaches wherein the three-dimensional data include labels set by a plurality of parts in the object, and the at least one processor is configured to execute the instructions to generate a plurality of the first patches by using the labels (Bengtsson teaches “us[ing] the labels provided for the multiple anatomical regions as markers to select segments within the two-dimensional segmentation mask 220”, wherein “patches may be defined by the border of the segment (i.e., the pixels or voxels classified as being part of the segment such as tumor tissue), a bounding box with coordinates generated to encompass the segment, or the border or the bounding box plus a buffer zone of predetermined number of pixels or voxels to ensure the entire segment is included within the patch of image data” as shown in para. [0087]).
Liu, Suleyman, and Bengtsson are all considered to be analogous to the claimed invention because they are in the same field of analyzing patches of an image in order to perform a characterization. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Liu (as modified by Suleyman) to incorporate the teachings of Bengtsson and include “wherein the three-dimensional data include labels set by a plurality of parts in the object, and the at least one processor is configured to execute the instructions to generate a plurality of the first patches by using the labels”. The motivation for doing so would have been that “splitting and labeling narrows the imaging space and allows for select CNN models to be trained on different anatomical regions, which improves overall learning of the CNN models”, as suggested by Bengtsson in para. [0102]. Therefore, it would have been obvious to one of ordinary skill at the time the invention was filed to combine Liu and Suleyman with Bengtsson to obtain the invention specified in claim 5.
Claim 10 is rejected under 35 U.S.C. 103 as being unpatentable over Liu et al. (U.S. Publication No. 2018/0089530 A1), hereinafter Liu, in view of Suleyman et al. (U.S. Publication 2014/0270488 A1), hereinafter Suleyman and Lyman et al. (U.S. Publication No. 2020/0161005 A1), hereinafter Lyman.
Regarding claim 10, Liu and Suleyman teach the model training apparatus according to claim 1.
While Liu teaches annotating training data indicating a location of the target object (see para. [0021]), Liu and Suleyman fail to teach wherein the training data includes information indicating whether the second patch includes a location of abnormality generated in the object, and the model is a model for detecting a location of abnormality in an object.
However, Lyman teaches wherein the training data includes information indicating whether the second patch includes a location of abnormality generated in the object (Lyman teaches “smaller sub-regions can be selected from previous sub-regions of the image data for each subsequently applied inference function, based on localization data indicated in each subsequently generated abnormality data that further localizes a detected abnormality” in para. [0420]. Additionally, para. [0516]-[0517] states that inference functions which generate the set of abnormality data can be applied to the plurality of sub-regions within the training sets, wherein “the abnormality detection data indicates region of interest data indicating a location of the first one of the plurality of abnormality types in the image data” as shown in para. [0529]-[0530]), and
the model is a model for detecting a location of abnormality in an object (Lyman teaches fine-tuned models geared towards indicating a location of a type of abnormality in para. [0529]-[0530]).
Liu, Suleyman, and Lyman are all considered to be analogous to the claimed invention because they are in the same field of analyzing patches of an image in order to perform a characterization. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Liu (as modified by Suleyman) to incorporate the teachings of Lyman and include “wherein the training data includes information indicating whether the second patch includes a location of abnormality generated in the object, and the model is a model for detecting a location of abnormality in an object”. The motivation for doing so would have been that “the pixel data corresponding to the three-dimensional subregions is utilized input to the forward propagation algorithm when the training step 1352 is employed to populate the weight vector and/or other model parameter data 1355” in order to “detect and/or classify abnormalities, or otherwise generate the inference data 1370”, as suggested by Lyman in para. [0147]. Therefore, it would have been obvious to one of ordinary skill at the time the invention was filed to combine Liu and Suleyman with Lyman to obtain the invention specified in claim 10.
Allowable Subject Matter
Claims 2-4 and 6 would be allowable if rewritten to overcome the rejection(s) under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), 2nd paragraph, set forth in this Office action and to include all of the limitations of the base claim and any intervening claims.
The following is a statement of reasons for the indication of allowable subject matter.
The best prior art of record is Liu, Suleyman, Lyman, and Bengtsson. Prior art applied alone or in combination with fails to anticipate or render obvious claims 2-4 and 6.
Claim 2
Regarding Claim 2, dependent upon claim 1, Liu in view of Suleyman teaches the limitation recited in claim 1.
Suleyman further teaches wherein the at least one processor is configured to execute the instructions to: generate N pieces of the first patches by using a parameter for controlling randomness of a patch; generate the second patch by using the parameter.
However, neither Liu, nor Suleyman, nor Lyman, nor Bengtsson, nor the combination, teaches when the three-dimensional data or a subset of the three-dimensional data is denoted by X', a set of the parameters satisfying a condition that the second patch is available as training data and includes at least a part of the X' is denoted by PX', and a range that the parameter may take for the m-th first patch when a condition that the second patch generated from the first patch can be generated also directly from the three-dimensional data is satisfied is denoted by Param (Pm), ensure that, upon computing a union of the Param (Pm) from m=1 to N, the PX' is included in the union.
Claims 3 and 4 include allowable subject matter by virtue of being dependent upon claim 2.
Claim 6
Regarding Claim 6, dependent upon claim 5, Liu in view of Suleyman and Bengtsson teaches the limitation recited in claim 1.
Bengtsson further teaches wherein a first number being a number of the parts having the first label is less than a second number being a number of the parts having the second label, and when a number of the first patches including the first label is denoted by a first number of patches and a number of the first patches including the second label is denoted by a second number of patches.
However, neither Liu, nor Suleyman, nor Lyman, nor Bengsston, nor the combination, teaches wherein the at least one processor is configured to execute the instructions to set a ratio of the first number of patches to the second number of patches higher than a ratio of the first number to the second number.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner
should be directed to KYLA G ALLEN whose telephone number is (703)756-5315. The examiner can
normally be reached M-F 7:30am - 4:30pm EST.
Examiner interviews are available via telephone, in-person, and video conferencing using a
USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use
the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor,
John Villecco can be reached on (571) 272-7319. The fax phone number for the organization where this
application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from
Patent Center. Unpublished application information in Patent Center is available to registered users. To
file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit
https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and
https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional
questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like
assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or
571-272-1000.
/Kyla Guan-Ping Tiao Allen/
Examiner, Art Unit 2661
/JOHN VILLECCO/Supervisory Patent Examiner, Art Unit 2661
Prosecution Insights