US 12,444,098 B2
Image processing device and image processing method
Fumio Hashimoto, Hamamatsu (JP)
Assigned to HAMAMATSU PHOTONICS K.K., Hamamatsu (JP)
Appl. No. 17/795,566
Filed by HAMAMATSU PHOTONICS K.K., Hamamatsu (JP)
PCT Filed Jan. 27, 2021, PCT No. PCT/JP2021/002809
§ 371(c)(1), (2) Date Jul. 27, 2022,
PCT Pub. No. WO2021/153604, PCT Pub. Date Aug. 5, 2021.
Claims priority of application No. 2020-012243 (JP), filed on Jan. 29, 2020.
Prior Publication US 2023/0102661 A1, Mar. 30, 2023
Int. Cl. G06T 11/00 (2006.01); A61B 6/00 (2024.01); G06T 5/70 (2024.01)
CPC G06T 11/005 (2013.01) [A61B 6/5258 (2013.01); G06T 5/70 (2024.01); G06T 2207/10081 (2013.01); G06T 2207/10088 (2013.01); G06T 2207/10104 (2013.01); G06T 2207/10108 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30016 (2013.01); G06T 2211/424 (2013.01)] 15 Claims
OG exemplary drawing
 
1. An image processing apparatus for removing noise from an m-th tomographic image created by dividing list data collected by a radiation tomography apparatus into M frames in a collection order and performing reconstruction processing based on the list data included in an m-th frame for each m of 1 or more and M or less, the apparatus comprising:
a feature extraction unit configured to input an input image to a feature extraction neural network and output an intermediate image from the feature extraction neural network;
a reconstruction unit configured to input the intermediate image to an m-th reconstruction neural network for each m of 1 or more and M or less and output an m-th output image from the m-th reconstruction neural network;
an evaluation unit configured to obtain an evaluation value based on a sum of differences between the m-th tomographic image and the m-th output image of each m of 1 or more and M or less; and
a control unit configured to repeatedly perform respective processes of the feature extraction unit, the reconstruction unit, and the evaluation unit and training of the feature extraction neural network and the m-th reconstruction neural network of each m of 1 or more and M or less based on the evaluation value, and output a plurality of m-th output images from the m-th reconstruction neural network of each m of 1 or more and M or less, wherein
M is an integer of 2 or more, the feature extraction neural network is a single neural network provided commonly for the first to M-th reconstruction neural networks, and the feature extraction neural network outputs the intermediate image representing common information extracted from the first to M-th tomographic images.