US 12,220,102 B2
Endoscopic image processing
Xinghui Fu, Shenzhen (CN); Zhongqian Sun, Shenzhen (CN); and Wei Yang, Shenzhen (CN)
Assigned to TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED, Shenzhen (CN)
Filed by TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED, Guangdong (CN)
Filed on Nov. 10, 2023, as Appl. No. 18/506,545.
Application 18/506,545 is a continuation of application No. 17/078,826, filed on Oct. 23, 2020, granted, now 11,849,914.
Application 17/078,826 is a continuation of application No. PCT/CN2019/112202, filed on Oct. 21, 2019.
Claims priority of application No. 201811276885.2 (CN), filed on Oct. 30, 2018.
Prior Publication US 2024/0081618 A1, Mar. 14, 2024
Int. Cl. A61B 1/00 (2006.01); G06N 3/045 (2023.01)
CPC A61B 1/0005 (2013.01) [A61B 1/000094 (2022.02); A61B 1/000096 (2022.02); G06N 3/045 (2023.01)] 20 Claims
OG exemplary drawing
 
1. An endoscopic image processing method, comprising:
receiving a current endoscopic image of an organ;
inputting the current endoscopic image into a deep convolutional network with a training parameter that is determined according to at least one first endoscopic image and at least one second endoscopic image transformed from the at least one first endoscopic image;
receiving organ type information from the deep convolutional network, the organ type information indicating an organ type corresponding to the current endoscopic image;
selecting an integration module of a plurality of integration modules associated with different organ types based on the organ type indicated by the organ type information, each of the plurality of integration modules being trained with different endoscopic images corresponding to the organ type associated with the respective integration module; and
processing the current endoscopic image of the organ using the selected integration module.