Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
GENERATING SYNTHETIC MEDICAL IMAGES, FEATURE DATA FOR TRAINING IMAGE SEGMENTATION, AND INPAINTED MEDICAL IMAGES USING GENERATIVE MODELS
Document Type and Number:
WIPO Patent Application WO/2024/059693
Kind Code:
A3
Abstract:
Generative models (e.g., a denoising diffusion probabilistic model ("DDPM") or other suitable generative model) are used to create synthetic medical images (e.g., synthetic digital radiographic images), feature data useful as a training data set for training an image segmentation model, inpainted medical images that depict a predicted postoperative outcome for a patient, and/or deidentified medical images in which radiographic markers have been removed.

Inventors:
WYLES CODY C (US)
ROUZROKH POURIA (US)
KHOSRAVI BARDIA (US)
TAUNTON MICHAEL J (US)
ERICKSON BRADLEY J (US)
Application Number:
PCT/US2023/074166
Publication Date:
April 25, 2024
Filing Date:
September 14, 2023
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
MAYO FOUND MEDICAL EDUCATION & RES (US)
International Classes:
G06T5/00; G16H50/50
Foreign References:
US20220233242A12022-07-28
Other References:
DU TIANMING ET AL: "Morphology Reconstruction of Obstructed Coronary Artery in Angiographic Images", 2019 IEEE VISUAL COMMUNICATIONS AND IMAGE PROCESSING (VCIP), IEEE, 1 December 2019 (2019-12-01), pages 1 - 4, XP033693934, DOI: 10.1109/VCIP47243.2019.8966015
YOO TAE KEUN ET AL: "A generative adversarial network approach to predicting postoperative appearance after orbital decompression surgery for thyroid eye disease", COMPUTERS IN BIOLOGY AND MEDICINE, NEW YORK, NY, US, vol. 118, 26 January 2020 (2020-01-26), XP086078629, ISSN: 0010-4825, [retrieved on 20200126], DOI: 10.1016/J.COMPBIOMED.2020.103628
Attorney, Agent or Firm:
STONE, Jonathan D. (US)
Download PDF: