# US 12,223,648 B2 · Optimizing rendering condition for medical image data based on detection certainty from machine learning model > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,223,648 B2 - Document ID: US12223648B2 - Status: Granted - Kind: B2 - Application number: 17/815,644 - Filed: 2022-07-28 - Published: 2024-02-01 - Granted: 2025-02-11 - Earliest claimed: 2022-07-28 - Assignees: Canon Medical Systems Corporation - Applicants: Canon Medical Systems Corporation - Inventors: Paul Thomson - Web page: https://patentgrep.com/patent/US12223648B2 ## CPC - G06T 7/0012: {Biomedical image inspection} - G06T 15/08: Volume rendering - G06T 15/40: Hidden part removal - G06T 15/506: {Illumination models} - G06T 2207/20081: Training; Learning - G06T 2207/20084: Artificial neural networks [ANN] - G06T 2207/30008: Bone - G06T 2207/30096: Tumor; Lesion - G06T 2210/41: Medical - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL ## Abstract A medical image processing apparatus comprising processing circuitry configured to: receive medical image data including a region of interest; generate first rendering data by rendering the medical image data based on a first rendering condition; receive a characteristic value relating to detection certainty with respect to the region of interest by inputting the first rendering data to a machine learning model; determine a property of the first rendering condition based on the characteristic value. ## Claims ### Claim 1 A medical image processing apparatus comprising processing circuitry configured to: receive medical image data including a region of interest; generate first rendering data by rendering the medical image data based on a first rendering condition; receive a characteristic value relating to detection certainty with respect to the region of interest by inputting the first rendering data to a machine learning model; determine a property of the first rendering condition based on the characteristic value. ### Claim 2 (depends on 1) The apparatus according to claim 1, wherein the processing circuitry is further configured to: generate second rendering data by rendering the medical image data based on a second rendering condition which is different from the first rendering condition when the characteristic value is at least one of a), b), c) and d): a) lower than a predetermined threshold; b) higher than a predetermined threshold; c) in a predetermined range; d) out of a predetermined range. ### Claim 3 (depends on 2) The apparatus according to claim 2, wherein the processing circuitry is further configured to: generate further rendering data by rendering the medical image data based on at least one further rendering condition which is different from the first and second rendering conditions in an iterative optimization process until a stopping criterion is reached. ### Claim 4 (depends on 3) An apparatus according to claim 3, wherein the stopping criterion is at least one of: a determination that a local minimum or maximum of the optimization process has been reached; a predetermined characteristic value or values; a predetermined number of further rendering conditions; and a change in the characteristic value is less than a predetermined percentage over a predetermined number of further rendering conditions. ### Claim 5 (depends on 3) The apparatus according to claim 3, wherein the processing circuitry is further configured to at least one of: determine a property of the further rendering condition at which the stopping criterion was reached; and display the further rendering data based on the further rendering condition at which the stopping criterion was reached. ### Claim 6 (depends on 1) The apparatus according to claim 1, wherein the processing circuitry is further configured to render the medical image data based on a global illumination method. ### Claim 7 (depends on 6) An apparatus according to claim 6, wherein the property of the rendering conditions is a visualization parameter value or particular configuration of visualization parameter values of the global illumination method. ### Claim 8 (depends on 1) The apparatus according to claim 1, wherein the processing circuitry is further configured to: render the medical image data using at least one of: a differentiable renderer and a gradient based method. ### Claim 9 (depends on 1) An apparatus according to claim 1, wherein the characteristic value relates to detection certainty of at least one of: a feature in the region of interest, bone fracture, calcification, anatomical abnormalities, tumours, and soft tissue injuries. ### Claim 10 (depends on 1) The apparatus according to claim 1, wherein the first rendering condition is at least one of a), b) or c): a) pre-set; b) manually defined; c) specifically selected with respect to the region of interest. ### Claim 11 (depends on 1) The apparatus according to claim 1, wherein the processing circuitry is further configured to: modify the rendering conditions by limiting to be within a predetermined distance in parameter space from the first rendering condition. ### Claim 12 (depends on 1) The apparatus according to claim 1, wherein the processing circuitry is further configured to: determine the characteristic value using at least one of: an ensemble and dropout approach. ### Claim 13 (depends on 3) The apparatus according to claim 3, wherein the iterative optimization process includes a term to discourage rendering conditions that result in an unrealistic image. ### Claim 14 (depends on 13) The apparatus according to claim 13, wherein the term is derived from an adversarially trained classifier that distinguishes between images obtained with given rendering conditions, and images that have been chosen as good images. ### Claim 15 (depends on 1) The apparatus according to claim 1, wherein the processing circuitry is further configured to at least one of: select the region of interest of the medical image data and/or of the rendering data automatically or based on input from a user, and use a hidden-surface determination algorithm to prevent occlusion of the region of interest. ### Claim 16 (depends on 1) The apparatus according to claim 1, wherein the characteristic value is a single scalar output. ### Claim 17 A medical image processing method, comprising: receiving medical image data including a region of interest; generating first rendering data by rendering the medical image data based on a first rendering condition; receiving a characteristic value relating to detection certainty with respect to the region of interest by inputting the first rendering data to a machine learning model; determining a property of the first rendering condition based on the characteristic value. ### Claim 18 A medical image processing apparatus comprising processing circuitry configured to: train a machine learning model to perform a detection task on first rendering data with respect to a region of interest of medical image data, wherein the training is performed on a training set of rendering data that are annotated. ### Claim 19 A training method comprising: training a machine learning model to perform a detection task on first rendering data with respect to a region of interest of medical image data, wherein the training is performed on a training set of rendering data that are annotated. ### Claim 20 (depends on 19) The training method of claim 19, wherein the training is performed on the training set of rendering data that are annotated to indicate whether at least one of a) a feature, b) a bone fracture, c) calcification, d) an anatomical abnormality, e) a tumour and f) a soft tissue injury is present in the region of interest.