Colonoscopy Image and Video Analysis for Performance Metrics

Colorectal Cancer is one of the leading causes of cancer related deaths in US. Due to significantly high miss rates in detection of polyps by doctors. A need was felt for automation of Polyp and Cancer detection in patients, one of which is finding camera motions inside the patient to make sure that doctor spends at least 6 minutes. This standard was set in 2006 by American Society for Gastrointestinal Endoscopy (ASGE) and American College of Gastroenterology (ACG). This book discusses various methods and techniques to estimate camera motion from RAW streaming Colonoscopy videos. This work includes analysis for various block search algorithms including fine tuning for frame skip rates, search window and approximations for different accuracy scenarios. Also some insight on implementation and analysis on parallel processing algorithms in CUDA (Compute Unified Device Architectures, NVIDIA). An analysis of performance can be judged by detailed section of experimental results. This work is a part of a major project which was funded by National Science Foundation and Mayo clinic, Rochester.