US 12,073,539 B2
Systems and methods for denoising medical images
Yikang Liu, Cambridge, MA (US); Shanhui Sun, Lexington, MA (US); Terrence Chen, Lexington, MA (US); Zhang Chen, Brookline, MA (US); and Xiao Chen, Cambridge, MA (US)
Assigned to Shanghai United Imaging Intelligence Co., Ltd., Shanghai (CN)
Filed by Shanghai United Imaging Intelligence Co., Ltd., Shanghai (CN)
Filed on Dec. 29, 2021, as Appl. No. 17/564,348.
Prior Publication US 2023/0206401 A1, Jun. 29, 2023
Int. Cl. G06K 9/62 (2022.01); G06N 3/045 (2023.01); G06N 3/08 (2023.01); G06T 5/50 (2006.01); G06T 5/70 (2024.01); G16H 30/40 (2018.01)
CPC G06T 5/70 (2024.01) [G06N 3/045 (2023.01); G06N 3/08 (2013.01); G06T 5/50 (2013.01); G16H 30/40 (2018.01); G06T 2207/10064 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01)] 20 Claims
OG exemplary drawing
 
1. An apparatus, comprising:
one or more processors configured to:
receive an input medical image comprising noise;
receive a target noise level; and
generate an output medical image using a first artificial neural network (ANN) such that at least a portion of the noise is removed from the output medical image in accordance with the target noise level, wherein the first ANN is trained to generate the output medical image in accordance with the target noise level through a training process and wherein, during the training process, the first ANN is configured to:
receive a first pair of training images that includes a first noisy training image and a first target training image, wherein the first target training image is generated using a second ANN and includes a first level of noise;
determine the first level of noise associated with the first target training image;
generate, using the first ANN, a denoised version of the first noisy training image based on the first level of noise; and
adjust parameters of the first ANN based on a difference between the denoised version of the first noisy training image and the first target training image.