US 12,144,638 B2
Methods and devices of generating predicted brain images
Yi-Tien Li, Taipei (TW); and Cheng-Yu Chen, Taipei (TW)
Assigned to TAIPEI MEDICAL UNIVERSITY, Taipei (TW)
Filed by TAIPEI MEDICAL UNIVERSITY, Taipei (TW)
Filed on Aug. 24, 2022, as Appl. No. 17/895,036.
Prior Publication US 2024/0065609 A1, Feb. 29, 2024
Int. Cl. A61B 5/00 (2006.01); A61B 5/055 (2006.01); A61B 5/103 (2006.01)
CPC A61B 5/4064 (2013.01) [A61B 5/055 (2013.01); A61B 5/103 (2013.01); A61B 5/742 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A method of generating a predicted brain image, comprising:
receiving a first brain image, the first brain image being generated by a magnetic resonance imaging (MRI) method;
encoding the first brain image to generate a latent vector; and
decoding the latent vector and one or more conditional features to generate the predicted brain image;
calculating a reconstruction loss between the predicted brain image and a known brain image, wherein the predicted brain image and the known brain image are for a same age; and
by a backpropagation method, adjusting encoding parameters and decoding parameters such that the reconstruction loss is minimized, the encoding parameters being used for encoding the first brain image and the decoding parameters being used for decoding the latent vector and the one or more conditional features,
wherein the one or more conditional features include at least one of: an age in future, a gender, previous brain images, omics features, and medical history, and
wherein the latent vector is multiplied by a first normal distribution.