A lower gastrointestinal endoscope image recognition method and system

Abstract

The present disclosure provides an image recognition method and system of a lower digestive tract endoscope. Wherein the method includes obtaining the lower digestive tract part image and marking the category to be recognized and the auxiliary category of the recognition interference image, and dividing it into a training set and a test set; using the training set and the test set to train and test the lower digestive tract part recognition model A and the lower digestive tract part Gastrointestinal part recognition model B; during the endoscopic operation, if the lower gastrointestinal part recognition model A recognizes that the type of the current lower gastrointestinal endoscopic image is the cecum in the first-level category, and there are consecutive N non-similar images If the probability of the cecum exceeds the preset threshold, the mirror withdrawal operation is started; N is a positive integer greater than or equal to 3; Real-time detection is performed on the endoscopic images of the tract, images belonging to the auxiliary category are excluded, and the first-level category and the second-level category to which the image belongs are output.

Description

technical field [0001] The disclosure belongs to the field of gastrointestinal endoscope image recognition, in particular to a lower gastrointestinal endoscope image recognition method and system. Background technique [0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art. [0003] In lower gastrointestinal endoscopy, it is often necessary to determine where a lesion is located in the colorectum. The inventors found that at present, most of the endoscopic examinations of the lower gastrointestinal tract are observed by human eyes, and the lesion site cannot be accurately judged. For example, multiple polyps in the lower gastrointestinal tract were found by visual observation, but it was not possible to accurately determine whether they were located in the ascending colon, transverse colon, descending colon, or sigmoid colon; [0004] In addition, some parts may be missed du...

AI Technical Summary

Problems solved by technology

The inventors found that at present, most of the endoscopic examinations of the lower gastrointestinal tract are observed by human eyes, and the lesion site cannot be accurately judged.

For example, multiple polyps in the lower gastrointestinal tract were found by visual observation, but it was not possible to accurately determine whether they were located in the ascending colon, transverse colon, descending colon, or sigmoid colon;

[0004] In addition, some parts may be missed during the lower gastrointestinal endoscopy, for example, the image quality of a certain part is poor due to the slipping of the mirror body or poor filling of the intestinal cavity, which affects the quality of the lower gastrointestinal endoscopy, resulting in Possibility of missed diagnosis

Images