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 .
Claim Objections
Claim 8 is objected to because of the following informalities:
In Claim 8, “displaying the simulated image data output from the machine learning model on thea user interface” should read as “displaying the simulated image data output from the machine learning model on the user interface.”
Appropriate correction is required.
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claims 1, 3-10, 12, and 14-22 are rejected under 35 U.S.C. 102 (a)(1) as being anticipated by Petkov et. al (US10339695B2).
Regarding Claim 1, Petkov teaches a computer implemented method for simulating a preview of medical imaging scan data based on first image data, comprising:
Paragraph [0006]: “By way of introduction, the preferred embodiments described below include methods, systems, instructions, and computer readable media for machine training an artificial intelligence and use of that artificial intelligence to provide rendering settings.”
Paragraph [0015]: “Content-based photorealistic rendering of medical images is based on machine learning.”
receiving from a medical imaging apparatus first image data of an anatomy of a patient obtained in a first scan, wherein the first image data is 3D image data;
Paragraph [0007]: “A medical dataset representing a three-dimensional region of a patient is loaded from memory.”
Paragraph [0062]: “Scan data representing a 3D volume is loaded as a medical dataset.”
obtaining an indication of at least a first set of one or more candidate imaging scan parameters for a subsequent imaging scan to be performed on the anatomy;
Paragraph [0007]: “The machine-learned model is trained with deep learning to extract features from the medical dataset and trained to output values for two or more physically-based rendering parameters based on input of the medical dataset.”
Paragraph [0018]: “The output of the trained artificial intelligence system is the optimal set of settings for the rendering parameters.”
Explanation: Rendering parameters correspond to candidate scan parameters that define how the subsequent image would appear.
retrieving a machine learning model, wherein the machine learning model is trained to perform image-to-image translation, and is adapted to receive at least a subset of a 3D image dataset of an anatomy, and is adapted to output simulated image data of the anatomy, wherein the simulated image data simulates an expected appearance of the image data which would be acquired from the medical imaging apparatus if run with the at least first set of candidate scan parameters;
Paragraph [0007]: “A machine applies the medical dataset to a machine-learnt non-linear model. The machine-learned model is trained with deep learning to extract features from the medical dataset and trained to output values for two or more physically-based rendering parameters based on input of the medical dataset.”
Paragraph [0009]: “The rendering parameters include windowing, transfer function, and lighting, and the settings are learned to provide a first image from the data similar to one or more second images for a same diagnostic context.”
Paragraph [0037]: “The machine learns to output renderer settings to model the resulting rendered image for a given scan dataset on a desired image.”
Paragraph [0070]: “Where the training was for consistency, the application of the machine-learnt model is more likely to output values in act 38 resulting in the photorealistic image corresponding to a standard image despite differences in the medical dataset.”
Explanation: These disclosures teach an ML model, an input 3D dataset, and a transformed/simulated image appearance output, which corresponds to image-to-image translation. The rendered image is predicted/simulated, based on parameter settings, and represents expected output appearance.
processing the first image data using the machine learning model, to obtain simulated image data;
Paragraph [0007]: “A physically-based renderer renders a photorealistic image of the three-dimensional region of the patient using the output values resulting from the applying. The photorealistic image is transmitted.”
Paragraph [0065]: “In act 34, the medical dataset is applied to the machine-learnt model.”
Paragraph [0067]: “The machine inputs the feature vector, resulting in the machine-learnt model outputting the rendering settings based on the learned knowledge.”
and acquiring second image data of the anatomy using the medical imaging apparatus with a second set of imaging scan parameters in a second scan, wherein the second set of parameters are the same or different to the first set of parameters.
Paragraph [0002]: ‘Due to the many different scan settings and patient variability, renderings for different patients or at different times appear different.”
Paragraph [0009]: “A medical scanner is configured to scan a patient.”
Paragraph [0015]: “Due to the variability between scan parameters, data contrast, noise, and/or the large number of rendering parameters…”
Paragraph [0086]: “For example, scan data is acquired and used for diagnosis or surgical planning, such as identifying a lesion or treatment location.”
Explanation: Performing scans, using different parameters, and producing different image data are disclosed.
Regarding Claim 3, Petkov teaches the method of claim 1, further comprising controlling a user interface to display a visual representation of the simulated image data.
Paragraph [0078]: “For example, the image is loaded into a buffer and output from the buffer to a display for viewing by a physician to aid diagnosis or pre-operative planning.”
Paragraph [0080]: “In another embodiment, the transmission is to a display as an initial image for interactive viewing. The machine-learnt model is used to initialize interactive viewing applications in the clinical workflow.”
Paragraph [0095]: “The display 54 is a monitor, LCD, projector, plasma display, CRT, printer, or other now known or later developed device for displaying the photorealistic image or images.”
Regarding Claim 4, Petkov teaches the method of claim 1, further comprising:
accessing a datastore storing a plurality of versions of the machine learning model, each respective version trained to output simulated image data which simulates an expected appearance of the input image data if acquired from the medical imaging apparatus run with a different respective set of candidate scan parameters;
Paragraph [0032]: “Alternatively, different models are trained for different situations. The different situations may include different scan modalities (e.g., different model for computed tomography, magnetic resonance, ultrasound, positron emission tomography, and single photon emission computed tomography). The different situations may include different types of tissue of interest (e.g., liver versus kidney), different diagnostic purpose or workflow (e.g., cancerous lesion versus bone calcification), and/or different users (e.g., different operators may have different preferences for visualization).”
Paragraph [0037]: “The machine learns to output renderer settings to model the resulting rendered image for a given scan dataset on a desired image.”
Paragraph [0055]: “The trained model is stored in a memory. Any memory may be used. The memory used for the training data may be used. For application, the memory may be in other devices. For example, the trained model is stored in a memory of a server. As another example, multiple copies of the trained model are provided to different physicians, medical scanners, and/or workstations for use by different physicians.”
Explanation: Different models correspond to different parameter sets and different expected appearances.
retrieving from the datastore one of the versions of the machine learning model, in dependence upon the obtained at least first set of candidate scan parameters;
Paragraph [0063]: “The diagnostic information may be used to select the machine-learnt model to use and/or be part of the input feature vector.”
Paragraph [0065]: “The sensor provides a measure of the luminosity. This measure may be used as an input in the input feature vector and/or to select a particular machine-learnt model to use.”
Explanation: Model selection = retrieval based on conditions (parameters)
and applying the retrieved version to the received first image data.
Paragraph [0065]: “In act 34, the medical dataset is applied to the machine-learnt model.”
Regarding Claim 5, Petkov teaches the method of claim 4, further comprising:
a plurality of sets of simulated image data for each of a plurality of sets of candidate imaging scan parameters, based on, for each candidate set of scan parameters: retrieving from the datastore one of the machine learning models;
Paragraph [0028]: “In one embodiment, additional samples are created by perturbing the path tracing rendering parameters of an input sample. The perturbing creates a collection of sets of the path tracing rendering parameters for each scan data set. Rendering is then performed using the sets of rendering parameters in the pool to produce a pool of corresponding images.”
Paragraph [0032]: “Alternatively, different models are trained for different situations. The different situations may include different scan modalities (e.g., different model for computed tomography, magnetic resonance, ultrasound, positron emission tomography, and single photon emission computed tomography). The different situations may include different types of tissue of interest (e.g., liver versus kidney), different diagnostic purpose or workflow (e.g., cancerous lesion versus bone calcification), and/or different users (e.g., different operators may have different preferences for visualization).”
Paragraph [0088]: “For learning, the machine 50 is configured by one or more machine learning algorithms.”
Explanation: Directly teaches multiple parameter sets and multiple simulated images. It also supports retrieval per parameter scenario.
and applying the machine learning model to the candidate set of scan parameters.
Paragraph [0028]: “The artificial intelligence system is applied on the input scan dataset to generate rendering parameters, which are then perturbed to generate a pool of rendering parameters.”
Paragraph [0050]: “The model learns to output rendering parameter settings based, in part, on settings for one or more parameters under the control of the user.”
Regarding Claim 6, Petkov teaches the method of claim 5, further comprising:
controlling a user interface to display a visual representation of each set of simulated image data, and an indication of the candidate scan parameters corresponding to each set of simulated image data;
Paragraph [0080]: “In another embodiment, the transmission is to a display as an initial image for interactive viewing.”
Paragraph [0095]: “The display 54 is a monitor, LCD, projector, plasma display, CRT, printer, or other now known or later developed device for displaying the photorealistic image or images.”
Paragraph [0061]: “While the training may use samples from many patients to learn features and/or learn to classify input medical data to provide values for rendering, the learnt model is applied to the medical data for a patient to output values for rendering parameters for that patient.”
Explanation: This teaches display of simulated images and associated rendering parameters, which corresponds to scan parameters.
and receiving from the user-interface a user-indicated selection of one of the candidate sets of scan parameters;
Paragraph [0044]: “In other or additional reinforcement learning, the training is modeled after steps taken by an expert or other user to achieve the desired visualization for the training data. The user's sequence of adjustments to provide the desired rendering from the scan dataset is monitored (e.g., select transfer function T1 , then window function W1 , then select a different transfer function T2 , then select a material reflectance MR1 , . . . )…The training learns to provide final rendering settings based on the monitored sequence.”
Paragraph [0058]: “In yet another example, acts for manual adjustment and/or initial setting of one or more rendering parameters are provided, such as for an interactive rendering workflow.”
Paragraph [0083]: “Other methods or acts may be implemented, such as providing a user input (e.g., mouse, trackball, touch pad, and/or keyboard) and user interface for interactive rendering.”
wherein the second set of scan parameters used for acquiring the second image data is defined by the user selection.
Paragraph [0075]: “The output values of the machine-learnt model are used as settings by the physically-based renderer.”
Explanation: Selected rendering parameters [Wingdings font/0xE0] applied for subsequent imaging/rendering
Regarding Claim 7, Petkov teaches the method of claim 6, comprising controlling the user interface such that the user-indicated selection of one of the candidate sets of scan parameters is achieved through a user-selection of one of the displayed sets of simulated image data.
Paragraph [0078]: “For example, the image is loaded into a buffer and output from the buffer to a display for viewing by a physician to aid diagnosis or pre-operative planning.”
Paragraph [0080]: “In another embodiment, the transmission is to a display as an initial image for interactive viewing. The machine-learnt model is used to initialize interactive viewing applications in the clinical workflow.”
Explanation: Selection via displayed images is shown through interactive image-based workflows.
Regarding Claim 8, Petkov teaches the method of claim 1, further comprising:
providing on a user interface a user control permitting a user to define a custom set of candidate imaging scan parameters;
Paragraph [0058]: “In yet another example, acts for manual adjustment and/or initial setting of one or more rendering parameters are provided, such as for an interactive rendering workflow.”
Paragraph [0083]: “Other methods or acts may be implemented, such as providing a user input (e.g., mouse, trackball, touch pad, and/or keyboard) and user interface for interactive rendering.”
receiving from the user interface a user-defined custom set of candidate scan parameters;
Paragraph [0044]: “In other or additional reinforcement learning, the training is modeled after steps taken by an expert or other user to achieve the desired visualization for the training data. The user's sequence of adjustments to provide the desired rendering from the scan dataset is monitored (e.g., select transfer function T1 , then window function W1 , then select a different transfer function T2 , then select a material reflectance MR1 , . . . )…The training learns to provide final rendering settings based on the monitored sequence.”
wherein retrieving the machine learning model from the datastore comprises retrieving a machine learning model trained to generate simulated image data which simulates the appearance of image data acquired with the user-defined custom set of candidate scan parameters;
Paragraph [0032]: “Alternatively, different models are trained for different situations. The different situations may include different scan modalities (e.g., different model for computed tomography, magnetic resonance, ultrasound, positron emission tomography, and single photon emission computed tomography). The different situations may include different types of tissue of interest (e.g., liver versus kidney), different diagnostic purpose or workflow (e.g., cancerous lesion versus bone calcification), and/or different users (e.g., different operators may have different preferences for visualization).”
Paragraph [0065]: “The sensor provides a measure of the luminosity. This measure may be used as an input in the input feature vector and/or to select a particular machine-learnt model to use.”
Explanation: Model selection based on conditions (parameters).
and displaying the simulated image data output from the machine learning model on the user interface.
Paragraph [0095]: “The display 54 is a monitor, LCD, projector, plasma display, CRT, printer, or other now known or later developed device for displaying the photorealistic image or images.”
Regarding Claim 9, Petkov teaches the method of claim 8, wherein a change in the custom scan parameters by the user via the user control automatically triggers generation of new simulated image data in accordance with the changed custom scan parameters, by retrieving a machine learning model trained to simulate the appearance of image data acquired with the custom scan parameters.
Paragraph [0028]: “In one embodiment, additional samples are created by perturbing the path tracing rendering parameters of an input sample. The perturbing creates a collection of sets of the path tracing rendering parameters for each scan data set. The artificial intelligence system is applied on the input scan dataset to generate rendering parameters, which are then perturbed to generate a pool of rendering parameters. Rendering is then performed using the sets of rendering parameters in the pool to produce a pool of corresponding images.”
Explanation: Changing parameters [Wingdings font/0xE0] new outputs generated.
Regarding Claim 10, Petkov teaches the method of claim 1, further comprising a quality assessment to the simulated image data to derive a quality indicator for the simulated image data.
Paragraph [0039]: “In one embodiment, the consistency is learned using a measured or calculated metric… The metric is used as an indicator of strength of importance of a given training sample.”
Paragraph [0040]: “The output image rendered from a given set of rendering settings for a sample of scan data and the associated quality metrics (i.e., similarity) are fed-back into the learning system for additional refinement of the learned parameters (e.g., in deep supervised and/or reinforcement learning). A metric measuring the similarity of one rendered image to one or a group of rendered images with the desired quality is calculated, and used as the “reward” to train the artificial intelligent agent using the deep reinforcement learning technique.”
Paragraph [0041]: “The quality metric in the reinforcement learning is used to learn to provide the optimal rendering parameters for increasing the conspicuity of the pathology of interest.”
Regarding Claim 12, Petkov teaches the method of claim 1, further comprising:
controlling a user interface to display a visual representation of the simulated image data;
Paragraph [0095]: “The display 54 is a monitor, LCD, projector, plasma display, CRT, printer, or other now known or later developed device for displaying the photorealistic image or images.”
receiving from the user interface a user-indicated adjusted set of candidate imaging scan parameters, subsequent to display of the simulated image data;
Paragraph [0044]: “In other or additional reinforcement learning, the training is modeled after steps taken by an expert or other user to achieve the desired visualization for the training data. The user's sequence of adjustments to provide the desired rendering from the scan dataset is monitored (e.g., select transfer function T1 , then window function W1 , then select a different transfer function T2 , then select a material reflectance MR1 , . . . )…The training learns to provide final rendering settings based on the monitored sequence.”
Paragraph [0058]: “In yet another example, acts for manual adjustment and/or initial setting of one or more rendering parameters are provided, such as for an interactive rendering workflow.”
Paragraph [0080]: “In another embodiment, the transmission is to a display as an initial image for interactive viewing. The machine-learnt model is used to initialize interactive viewing applications in the clinical workflow.”
Paragraph [0083]: “Other methods or acts may be implemented, such as providing a user input (e.g., mouse, trackball, touch pad, and/or keyboard) and user interface for interactive rendering.”
Explanation: User modifies parameters after viewing images.
wherein the received adjusted set of imaging scan parameters are used as the second set of scan parameters.
Paragraph [0037]: “The machine learns to output renderer settings to model the resulting rendered image for a given scan dataset on a desired image.”
Explanation: Selected/adjusted parameters are used for subsequent imaging/rendering.
Regarding Claim 14, Petkov teaches the method of claim 1, further comprising deriving the second scan parameters by adjusting the first candidate scan parameters, to thereby to derive second scan parameters different to the first scan parameters, and wherein adjustment of the first scan parameters to derive the second scan parameters is performed in dependence upon a user input received from a user interface, and/or adjustment of the first scan parameters to derive the second scan parameters is performed at least in part by an automated adjustment operation, in dependence upon processing applied to the derived simulated image data.
Paragraph [0007]: “A machine applies the medical dataset to a machine-learnt non-linear model. The machine-learned model is trained with deep learning to extract features from the medical dataset and trained to output values for two or more physically-based rendering parameters based on input of the medical dataset.”
Paragraph [0074]: “The user may, however, make any number of adjustments.”
Paragraph [0050]: “The model learns to output rendering parameter settings based, in part, on settings for one or more parameters under the control of the user.”
Paragraph [0075]: “In act 42, a physically-based renderer renders the photorealistic image of the 3D region of the patient using the values output from the application.”
Explanation: Output parameters derived from the input dataset corresponds to adjusted parameters, which satisfies deriving second parameters from first candidate parameters. Rendering corresponds to processing applied to simulated image data which drives parameter adjustment.
Regarding Claim 15, Petkov teaches the method of claim 1, wherein the method comprises:
controlling a user interface to display a visual representation of the simulated image data;
Paragraph [0078]: “For example, the image is loaded into a buffer and output from the buffer to a display for viewing by a physician to aid diagnosis or pre-operative planning.”
Paragraph [0095]: “The display 54 is a monitor, LCD, projector, plasma display, CRT, printer, or other now known or later developed device for displaying the photorealistic image or images.”
and wherein the second set of imaging scan parameters are determined based on a user input;
Paragraph [0044]: “In other or additional reinforcement learning, the training is modeled after steps taken by an expert or other user to achieve the desired visualization for the training data. The user's sequence of adjustments to provide the desired rendering from the scan dataset is monitored (e.g., select transfer function T1 , then window function W1 , then select a different transfer function T2 , then select a material reflectance MR1 , . . . )…The training learns to provide final rendering settings based on the monitored sequence.”
Paragraph [0058]: “In yet another example, acts for manual adjustment and/or initial setting of one or more rendering parameters are provided, such as for an interactive rendering workflow.”
Paragraph [0083]: “Other methods or acts may be implemented, such as providing a user input (e.g., mouse, trackball, touch pad, and/or keyboard) and user interface for interactive rendering.”
and/or wherein the method comprises automated determination of the second imaging scan parameters based at least in part on the simulated image data.
Paragraph [0018]: “Using many examples, the machine training learns to provide rendering settings for photorealistic rendering based on input data for a specific patient.”
Paragraph [0041]: “The quality metric in the reinforcement learning is used to learn to provide the optimal rendering parameters for increasing the conspicuity of the pathology of interest.”
Regarding Claim 16, Petkov teaches the method of claim 1, wherein the method further comprises:
processing source 3D image data of the anatomy of the patient to generate: a first 2D image representing a view of the anatomy across a first view plane, the first view plane representing a view from a first angular position relative to the anatomy, and to generate a second 2D image representing a view of the anatomy across a second view plane, the second view plane representing a view from a second angular position relative to the anatomy;
Paragraph [0007]: “A physically-based renderer renders a photorealistic image of the three-dimensional region of the patient using the output values resulting from the applying. The photorealistic image is transmitted.”
Paragraph [0047]: “Viewing design parameters include type of camera, position of the camera, orientation of the camera, intrinsic parameters for viewing, or others.”
Paragraph [0062]: “Scan data representing a 3D volume is loaded as a medical dataset.”
Explanation: Rendering produces 2D projections from 3D data, and different camera orientations correspond to different view planes.
obtaining an indication of one or more volumetric sub-regions within the source image data which each contains a respective anatomy of interest;
Paragraph [0032]: “The different situations may include different types of tissue of interest (e.g., liver versus kidney), different diagnostic purpose or workflow (e.g., cancerous lesion versus bone calcification), and/or different users (e.g., different operators may have different preferences for visualization).”
Paragraph [0041]: “Different rendering settings increase the conspicuity of different pathologies. The quality metric in the reinforcement learning is used to learn to provide the optimal rendering parameters for increasing the conspicuity of the pathology of interest.”
Explanation: Region/anatomy selection is explicitly tied to diagnostic interest.
for each of the one or more volumetric sub-regions, extracting from the source image data at least a first and second 2D slice through the respective sub-region, each 2D slice being orthogonal to both the first and second view planes;
Paragraph [0062]: “The scan data may be from multiple two-dimensional scans or may be formatted from a 3D scan.”
Explanation: 3D scan supports slice extraction in multiple orientations.
controlling a user interface to display a representation of: the first 2D image, the second 2D image, a boundary outline of each volumetric sub-region superposed on the first 2D image and second 2D image, and each of the generated 2D slices.
Paragraph [0078]: “For example, the image is loaded into a buffer and output from the buffer to a display for viewing by a physician to aid diagnosis or pre-operative planning.”
Regarding Claim 17, Petkov teaches the method of claim 16, wherein the method further comprises receiving from the user interface a user-indicated adjustment of the volumetric sub-region, and wherein the adjustment is a change in a position of the volumetric sub-region relative to the patient anatomy, or a change in the volume defined by the sub-region.
Paragraph [0047]: “Viewing design parameters include type of camera, position of the camera, orientation of the camera, intrinsic parameters for viewing, or others.”
Paragraph [0074]: “The user may, however, make any number of adjustments.”
Explanation: Position/extent adjustments correspond to sub-region changes.
Regarding Claim 18, Petkov teaches the method of claim 17, wherein the method comprises controlling the user interface so that, responsive to the user indicating the adjustment to the sub-region, the displayed boundary of the sub-region is automatically adjusted in accordance with the user- indicated adjustment.
Paragraph [0096]: “For interactive rendering, new images may be generated as settings for one or more rendering parameters are changed by a user.”
Explanation: UI automatically updates displayed results based on adjustments.
Regarding Claim 19, Petkov teaches the method of claim 17, wherein, responsive to receipt of the user-indicated adjustment of a given sub-region, the method further comprises extracting a new first and second 2D slice through each volumetric sub-region, and displaying the new 2D slices in place of the 2D slices previously displayed on the user interface.
Paragraph [0077]: “A sequence of images due to alteration of values for one or more rendering parameters may be output.”
Paragraph [0096]: “For interactive rendering, new images may be generated as settings for one or more rendering parameters are changed by a user.”
Explanation: Adjustment [Wingdings font/0xE0] new images generated and displayed.
Regarding Claim 20, Petkov teaches the method of claim 16, wherein obtaining the indication of the sub-region of interest comprises applying an image analysis operation to the source image data to detect an anatomical object of interest, and subsequently determining a volumetric-subregion in dependence upon a boundary of the identified anatomical object of interest within the image data.
Paragraph [0007]: ‘A physically-based renderer renders a photorealistic image of the three-dimensional region of the patient using the output values resulting from the applying.”
Paragraph [0016]: “Image features, and optionally, available non-image data are mapped to sets of rendering parameters that produce optimal images, where the relationship between the two is highly non-linear in the general case.”
Paragraph [0025]: “Parametric image information derived from medical image analysis algorithms (e.g. cardiac strain map or elasticity)”
Paragraph [0032]: “The different situations may include different types of tissue of interest (e.g., liver versus kidney), different diagnostic purpose or workflow (e.g., cancerous lesion versus bone calcification), and/or different users (e.g., different operators may have different preferences for visualization).”
Paragraph [0041]: “Different rendering settings increase the conspicuity of different pathologies. The quality metric in the reinforcement learning is used to learn to provide the optimal rendering parameters for increasing the conspicuity of the pathology of interest.”
Paragraph [0062]: “Scan data representing a 3D volume is loaded as a medical dataset.”
Paragraph [0069]: “By applying the deep-learnt model in act 36, features are extracted from the medical dataset.”
Explanation: These excerpts disclose image analysis operations applied to source image data for identifying meaningful structures (anatomy/features), as well as explicit identification of anatomy/pathology of interest. Once anatomy is identified within a 3D volume, a volumetric region/sub-region is defined.
Regarding Claim 21, Petkov teaches the method of claim 1, wherein the method further comprises:
identifying a slab of interest within 3D source image data, wherein the 3D source image data is the first image data or the simulated image data, and wherein a slab is a volumetric region consisting of a consecutive stack of image slices in the source image data;
Paragraph [0024]: “A two-dimensional (2D), 3D, 2D+time sequence, 3D+time sequence, and/or other image or scan data may be used.”
Paragraph [0062]: “Scan data representing a 3D volume is loaded as a medical dataset.”
Explanation: Explicit support for 3D data and stack of slices.
extracting a stack of image slices from the source image data corresponding to the slab of interest;
Paragraph [0062]: “The scan data may be from multiple two-dimensional scans or may be formatted from a 3D scan.”
Explanation: 3D dataset enables extraction of slice stacks.
generating a volume rendering of the slab of interest;
Paragraph [0007]: ‘A physically-based renderer renders a photorealistic image of the three-dimensional region of the patient using the output values resulting from the applying.”
Explanation: Rendering of a selected portion of a volume = slab rendering
displaying the volume rendering on a user interface.
Paragraph [0078]: “For example, the image is loaded into a buffer and output from the buffer to a display for viewing by a physician to aid diagnosis or pre-operative planning.”
Regarding Claim 22, Petkov teaches the method of claim 21, wherein the method further comprises:
applying an image analysis operation to identify an anatomical region of interest in the source image data;
Paragraph [0025]: “Parametric image information derived from medical image analysis algorithms (e.g. cardiac strain map or elasticity)”
Paragraph [0069]: “By applying the deep-learnt model in act 36, features are extracted from the medical dataset.”
identifying the slab of interest within the source image data in dependence upon the identified anatomical region of interest by setting a position and thickness of the slab within the image data so as to overlap with at least a portion of the anatomical region of interest.
Paragraph [0016]: “Image features, and optionally, available non-image data are mapped to sets of rendering parameters that produce optimal images, where the relationship between the two is highly non-linear in the general case.”
Paragraph [0018]: “Using many examples, the machine training learns to provide rendering settings for photorealistic rendering based on input data for a specific patient.”
Paragraph [0032]: “The different situations may include different types of tissue of interest (e.g., liver versus kidney), different diagnostic purpose or workflow (e.g., cancerous lesion versus bone calcification), and/or different users (e.g., different operators may have different preferences for visualization).”
Paragraph [0047]: “Viewing design parameters include type of camera, position of the camera, orientation of the camera, intrinsic parameters for viewing, or others.”
Explanation: Region of interest drives selection of relevant portion of volume. There is also parameterized control of spatial positioning and viewing/extent (which is equivalent to thickness).
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.
Claim 2 is rejected under 35 U.S.C. 103 as being unpatentable over Petkov et. al in view of Moghari et. al (“Estimation of full-dose 4D CT perfusion images from low-dose images using conditional generative adversarial networks”).
Regarding Claim 2, Petkov teaches the method of claim 1, but fails to teach that the first image data has lower spatial resolution than the second image data; and/or wherein the first scan administers a lower radiation dose than the second scan.
However, Moghari teaches that a first scan administers a lower radiation dose than the second scan, stating that “we simulated low-dose CTP images corresponding to tube currents of 100 mAs and 45 mAs…” (Abstract), and that “in our institution, a standard dose CTP scan is performed at 200 mAs…” (Section III, Results and Discussion, pg. 2). This teaches that a first image has a lower dose (45-100 mAs) while a second image has a higher dose (200 mAs). Additionally, Moghari teaches that the first image data has lower spatial resolution than the second image data, stating that “a reduction in dose is accompanied by an increase in noise…” (Abstract), which establishes that a lower dose leads to higher noise and lower effective image quality/resolution.
Thus, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify Petkov’s system to include a low-dose first scan and a higher-dose second scan as taught by Moghari. One of ordinary skill in the art would have been motivated to apply the low-dose and standard-dose imaging approach to the system of Petkov in order to reduce radiation exposure to the patient while maintaining sufficient image quality, since the reference explicitly teaches that “the ability to reduce this dose without compromising the accuracy of image-based stroke modelling is highly desirable… there is a trade-off between radiation dose and image quality” (Introduction, pg. 1). Such modification would have yielded predictable results because adjusting imaging dose parameters to balance image quality and radiation exposure is a well-understood and routine optimization in medical imaging systems.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Song (CN107146266A) teaches an image reconstruction method comprising the steps of according to first scanning data which correspond with a to-be-reconstructed target image and a preset reconstruction parameter, generating a preview image, wherein the reconstruction parameter comprises an interlayer interval parameter and a layer thickness parameter; based on an operation behavior to the preview image, adjusting a reconstruction range parameter, wherein the reconstruction range parameter comprises a central position parameter and a reconstruction visual field parameter of the reconstructed image; and performing image reconstruction according to the reconstruction range parameter and the corresponding second scanning data, and obtaining a target image.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to WILLIAM ADU-JAMFI whose telephone number is (571) 272-9298. The examiner can normally be reached M-T 8:00-6:00.
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, Andrew Bee can be reached at (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 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.
/WILLIAM ADU-JAMFI/Examiner, Art Unit 2677
/ANDREW W BEE/Supervisory Patent Examiner, Art Unit 2677
Prosecution Insights