US 12,277,703 B2
Image processing method and device, electronic apparatus and readable storage medium
Dawei Tang, Beijing (CN); Jijing Huang, Beijing (CN); and Zhiliang Wang, Beijing (CN)
Assigned to BOE Technology Group Co., Ltd., Beijing (CN)
Appl. No. 17/785,173
Filed by BOE Technology Group Co., Ltd., Beijing (CN)
PCT Filed Sep. 10, 2021, PCT No. PCT/CN2021/117703
§ 371(c)(1), (2) Date Jun. 14, 2022,
PCT Pub. No. WO2022/089039, PCT Pub. Date May 5, 2022.
Claims priority of application No. 202011194009.2 (CN), filed on Oct. 30, 2020.
Prior Publication US 2023/0009202 A1, Jan. 12, 2023
Int. Cl. G06T 7/00 (2017.01); G06V 10/44 (2022.01); G06V 10/77 (2022.01); G06V 10/94 (2022.01)
CPC G06T 7/0012 (2013.01) [G06V 10/44 (2022.01); G06V 10/7715 (2022.01); G06V 10/94 (2022.01); G06T 2207/10116 (2013.01); G06T 2207/30096 (2013.01)] 14 Claims
OG exemplary drawing
 
1. An image processing method, wherein the image processing method is used for processing an X-ray image, comprising:
obtaining feature map data of an input image; wherein the input image is an image generated in the X-ray detection;
extracting a feature region in the feature map data in accordance with a size of a convolution kernel;
performing windowing processing on the feature region; and
obtaining a windowed feature map of the input image in accordance with the feature region obtained after the windowing processing;
wherein the performing the windowing processing on the feature region comprises inputting feature sub-data corresponding to the feature region in feature data into M groups of flip-flops sequentially to obtain a windowed feature sub-map corresponding to the feature region, wherein each group of flip-flops comprises N flip-flops, a value of M is the same as a size of the convolution kernel in a first direction, a value of N is the same as a size of the convolution kernel in a second direction, and the first direction is different from the second direction.