Reducing redundancy in wireless capsule endoscopy videos

Area of Science:

• Medical Imaging
• Gastroenterology
• Computer Vision

Background:

• Wireless capsule endoscopy (WCE) generates vast amounts of video data.
• Redundant frames in WCE videos limit spatial coverage and diagnostic efficiency.
• Intestinal peristalsis causes repetitive forward and backward movements, leading to near-duplicate images.

Purpose of the Study:

• To develop and evaluate a method for eliminating similar frames in WCE videos.
• To maximize spatial coverage and minimize image redundancy.
• To improve the efficiency of small intestine examinations using WCE.

Main Methods:

• Implemented an intensity correction method.
• Utilized optical flow and feature-based techniques to detect near-duplicate images.
• Applied the combined method to wireless capsule endoscopy videos of the human intestines.

Main Results:

• The developed technique effectively detected and reduced near-duplicate images.
• A significant reduction of 52.3% in duplicate images was achieved in small intestine videos.
• The method addresses image redundancy caused by peristaltic egomotions.

Conclusions:

• The proposed frame elimination technique enhances WCE video analysis.
• This method improves spatial coverage and reduces data redundancy.
• It offers a valuable tool for more efficient gastrointestinal diagnostics.