Adaptive Motion Compensated Spatial Temporal Filter of Colonoscopy Video

Colonoscopy is a popular procedure which used to detect an abnormality. Early diagnosis can help heal many patients. The purpose of this paper removing/reducing some artifacts improve the visual quality colonoscopy videos provide better information for physicians. This work complements series consisting three previously published papers. In paper, optic flow motion compensation, where number consecutive images are registered integrate create new image that has/reveals more than original one. Colon were classified into informative and noninformative by using deep neural network. Then, two different strategies treat images. Informative treated Lucas Kanade with adaptive temporal mean/median filter, whereas derivative Gaussian (LKDOG) median Comparison showed achieved results those state-of-the-art same degraded colon data set. proposed algorithm reduced error alignment factor about 0.3, 100% successful ratio. conclusion, approaches in case enhancing as shown section; also, it helped reveal from have very few details/no details.