Stomach picture recognition method and device

Abstract

The invention provides a stomach picture recognition method and a device. The method comprises the following steps: dividing stomach video data into a plurality of video image sets with the same frame number according to a video frame sequence; the plurality of video image sets are input into a trained image recognition model, a stomach part recognition result corresponding to each frame of image in the stomach video data is obtained, the trained image recognition model is constructed by a convolutional neural network, a Transform network and a full connection layer, and the trained image recognition model is constructed by a convolutional neural network, a Transform network and a full connection layer; and training through a sample video image set marked with stomach part category labels. According to the method, the convolutional neural network and the Transform network are combined, so that when feature extraction is carried out on the stomach picture, the time sequence information of the picture features can be obtained, and the category of the stomach picture can be judged more accurately by combining the local picture information and the time sequence information.

Description

technical field [0001] The present invention relates to the technical field of image recognition, in particular to a stomach image recognition method and device. Background technique [0002] In existing endoscopic inspections, it is necessary to judge the stomach part where the video image was taken according to the video image captured by the endoscope. [0003] Although there have been technologies that use artificial intelligence image recognition to analyze gastroscope pictures, most of them simply use convolutional neural networks to extract and classify pictures, and only consider the local feature information of a single picture. Judging the part category of the picture, resulting in low accuracy of the recognition result. [0004] Therefore, there is an urgent need for a stomach picture recognition method and device to solve the above problems. Contents of the invention [0005] Aiming at the problems existing in the prior art, the present invention provides a s...

AI Technical Summary

Problems solved by technology

[0003] Although there have been technologies that use artificial intelligence image recognition to analyze gastroscope pictures, most of them simply use convolutional neural networks to extract and classify pictures, and only consider the local feature information of a single picture. Judging the part category of the picture, resulting in low accuracy of the recognition result

Images