Digestive system diseases like Crohn’s disease, inflammatory bowel disease, and colorectal cancer, are affecting a large population across the world. Wireless Capsule Endoscopy (WCE) is a good alternative for screening colorectal cancer which is both relatively pain free and eliminates the fear of traditional colonoscopy among patients. WCE carries cameras on board, recording video of the gastro-intestinal (GI) system while traveling through it. WCE can, in principle, screen the entire GI system, requiring little effort from the patient, and the procedure does not require the presence of medical personnel, nor special facilities to perform the screening as the video is recorded continuously and transmitted to a remote system. With decreased cost related to hospital stays along with other benefits (e.g., relatively painless), WCE is seen as a promising alternative to early-stage colorectal cancer detection.

WCE image/video analysis comes with a set of unique challenges. Due to the passage mechanism through the digestive system, not all frames can be captured. Moreover, the images captured by WCE are of much lower spatial resolution than what most medical professionals are used to analyzing and machine learning experts are used to annotating for the purpose of creating training samples. The loss of temporal and spatial resolution need to be taken into account in downstream analysis of WCE image/video data.
Topics (not limited to):
Classification of pathologies
Segmentation of pathologies
Video summarization
Visualization
Image and video enhancement
Datasets
Structure from motion (rigid and non-rigid)
3D reconstruction of nonrigid anatomical structures