# US 12,217,852 B2 · Image processing device, image processing method, image processing program, learning device, learning method, learning program, and derivation model > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,217,852 B2 - Document ID: US12217852B2 - Status: Granted - Kind: B2 - Application number: 17/699,190 - Filed: 2022-03-21 - Published: 2022-07-07 - Granted: 2025-02-04 - Earliest claimed: 2019-09-27 - Assignees: FUJIFILM Corporation - Applicants: FUJIFILM Corporation - Inventors: Atsushi Tachibana - Web page: https://patentgrep.com/patent/US12217852B2 ## CPC - G06T 12/30: Image post-processing, e.g. metal artefact correction - G16H 30/40: for processing medical images, e.g. editing - A61B 5/055: involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging - G06F 17/12: Simultaneous equations {, e.g. systems of linear equations} - G06N 3/0464: Convolutional networks [CNN, ConvNet] - G06N 3/09: Supervised learning - G06T 7/0012: {Biomedical image inspection} - G06T 11/10: Texturing; Colouring; Generation of textures or colours (retouching, inpainting or scratch removal G06T5/77) - G06N 3/045: Combinations of networks - G06T 2207/10081: Computed x-ray tomography [CT] - G06T 2207/30088: Skin; Dermal - G06T 2211/441: AI-based methods, deep learning or artificial neural networks - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL - G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA - A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D) - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract An image acquisition unit acquires at least one target medical image having a representation format different from an MRI image. A tissue eigenvalue derivation unit has a derivation model constructed by machine learning using a plurality of teacher data to, in a case in which at least one medical image having the representation format different from the MRI image is input, output a tissue eigenvalue of MRI for the input medical image, and inputs the target medical image to the derivation model to derive the tissue eigenvalue for the target medical image. ## Claims ### Claim 1 An image processing device comprising at least one processor, wherein the processor is configured to: acquire at least one target medical image having a representation format different from an MRI image; and input the target medical image to a derivation model to derive the tissue eigenvalue for the target medical image, wherein the derivation model being constructed by machine learning using a plurality of teacher data to, in a case in which at least one medical image having the representation format different from the MRI image is input, output a plurality of tissue eigenvalues of MRI for the input medical image, wherein the plurality of tissue eigenvalues represents different representation formats of MRI; and generate a derived MRI image based on the tissue eigenvalue of MRI derived by the derivation model through the target medical image having the representation format different from the MRI image. ### Claim 2 (depends on 1) The image processing device according to claim 1, wherein the target medical image and the medical image are CT images. ### Claim 3 (depends on 1) The image processing device according to claim 1, wherein the target medical image and the medical image are a plurality of CT images for the same subject, which are acquired by using radiation having different energy distributions, respectively. ### Claim 4 (depends on 1) The image processing device according to claim 1, wherein the derivation model is constructed by performing machine learning using the plurality of teacher data including a medical image for learning including a specific structure of a subject, and a tissue eigenvalue for learning for the same subject as the subject. ### Claim 5 (depends on 1) The image processing device according to claim 1, wherein the tissue eigenvalue is a T1 value, a T2 value, and a proton density (PD) value. ### Claim 6 (depends on 1) The image processing device according to claim 1, wherein the processor is further configured to derive a predetermined type of MRI image by using the tissue eigenvalue. ### Claim 7 (depends on 6) The image processing device according to claim 6, wherein the processor is further configured to display the MRI image on a display. ### Claim 8 (depends on 7) The image processing device according to claim 7, wherein the processor is configured to display the target medical image on the display. ### Claim 9 A learning device comprising: at least one processor, wherein the processor is configured to construct a derivation model by performing machine learning using a plurality of teacher data to, in a case in which at least one medical image having a representation format different from an MRI image is input, output a plurality of tissue eigenvalues of MRI for the medical image, wherein the plurality of tissue eigenvalues represent different representation formats of MRI. ### Claim 10 (depends on 9) The learning device according to claim 9, wherein the teacher data includes a medical image for learning including a specific structure of a subject, and a tissue eigenvalue for learning for the same subject as the subject, and the processor is configured to perform the machine learning by inputting the medical image for learning to the derivation model, outputting the tissue eigenvalue from the derivation model, and using a difference between the output tissue eigenvalue and the tissue eigenvalue for learning as a loss. ### Claim 11 An image processing method comprising: acquiring at least one target medical image having a representation format different from an MRI image; and inputting the target medical image to a derivation model to derive the tissue eigenvalue for the target medical image, wherein the derivation model being constructed by machine learning using a plurality of teacher data to, in a case in which at least one medical image having the representation format different from the MRI image is input, output a plurality of tissue eigenvalues of MRI for the input medical image, wherein the plurality of tissue eigenvalues represents different representation formats of MRI; and generating a derived MRI image based on the tissue eigenvalue of MRI derived by the derivation model through the target medical image having the representation format different from the MRI image. ### Claim 12 A learning method comprising: constructing a derivation model by performing machine learning using a plurality of teacher data to, in a case in which at least one medical image having a representation format different from an MRI image is input, output a plurality of tissue eigenvalues of MRI for the medical image, wherein the plurality of tissue eigenvalues represent different representation formats of MRI. ### Claim 13 A non-transitory computer-readable storage medium that stores an image processing program causing a computer to execute: a procedure of acquiring at least one target medical image having a representation format different from an MRI image; and a procedure of inputting the target medical image to a derivation model to derive the tissue eigenvalue for the target medical image, wherein the derivation model being constructed by machine learning using a plurality of teacher data to, in a case in which at least one medical image having the representation format different from the MRI image is input, output plurality of tissue eigenvalues of MRI for the input medical image, wherein the plurality of tissue eigenvalues represents different representation formats of MRI; and a procedure of generating a derived MRI image based on the tissue eigenvalue of MRI derived by the derivation model through the target medical image having the representation format different from the MRI image. ### Claim 14 A non-transitory computer-readable storage medium that stores a learning program causing a computer to execute: a procedure of constructing a derivation model by performing machine learning using a plurality of teacher data to, in a case in which at least one medical image having a representation format different from an MRI image is input, output a plurality of tissue eigenvalues of MRI for the medical image, wherein the plurality of tissue eigenvalues represent different representation formats of MRI.