Combined image analysis algorithm for quality control of colonoscopy procedure

The value of the colonoscopy procedure as an effective screening diagnostic method depends on the quality of the examination of the colon mucosa. In this case, it is necessary to bring the colonoscope to the dome of the cecum in more than 95% of cases. The presence of the dome area in the saved images and video sequences of the study indicates compliance with one of the important regulations for the quality of the colonoscopic procedure. The goal of this paper is to study algorithms based on deep machine learning methods for detecting the dome of the cecum in video images of a colonoscopic examination of the intestine. A database of 43 patients consisting of 4618 images of the dome of the cecum was collected for training and testing the algorithms. The developed combined ADCT-OP algorithm shows the advantage achieved through the use of the optical flow analysis algorithm, which is 0.14 according to the F1 metric. It is shown that the proposed algorithm can be used not only for post-processing of archived video data, but also for its application in real time. The proposed algorithm will allow for an objective procedure for quality control of colonoscopy examination in terms of reaching the dome of the cecum in accordance with medical regulations.