Processing of endoscopic images

The group of Alexander Maksimenko, Ph.D.

The group includes:

The group is developing geometrical and topological methods of processing endoscopic images so that automatic image segmentation and classification by analyzing tissue’s cellular texture is possible. The main task is automatic extraction of vessel and pit patterns from the analyzed image, whose parameters describe tissue morphology. Both classical and modern image processing methods based on computational topology and use of persistence diagrams are used in processing.
At the present time, creation and filling in collaboration with professional endoscopists of Yaroslavl regional clinical cancer hospital of endoscopic image database, which corresponds to known types of cellular texture and verified by means of histological study, is in progress. It will allow a preliminary conclusion decision about lesion type to be given with the help of automatic vessel and pit pattern parameters classification.