Capsule endoscopy (CE) is an established noninvasive diagnostic modality in investigating small bowel disease. CE has a pivotal role in assessing patients with suspected bleeding or identifying evidence of active Crohn’s disease in the small bowel. However, CE produces lengthy videos with at least eighty thousand frames, with a frequency rate of 2 frames per second. Gastroenterologists cannot dedicate 8 to 15 hours for reading the CE video frames to arrive at a diagnosis. This is why the issue of analyzing CE videos based on modern artificial intelligence techniques becomes a necessity. However, machine learning including deep learning has failed to report robust results because of the lack of large samples to train its neural nets. In this paper we are describing a thick data approach that learns from few anchor images provided by sound data sets like KVASIR and CrohnIPI to filter candidate frames that include interesting anomalies at any CE video and allow these candidate frames to a feature extraction process to provide representative measures on the anomaly, like the size of the anomaly and the color contrast compared to the image background, and later feed these features to a decision tree that can classify the candidate frames as having a condition like the Crohn's Disease. Our thick data approach as reported accuracy of detecting Crohn’s Disease based on availability of ulcer areas at the candidate frames for KVASIR was 89.9% and for the CrohnIPI was 83.3%. We are continuing our research to fine tune our approach by adding more thick data methods for enhancing the diagnosis accuracy.

Machine learning including deep learning cannot effectively result in reliable and robust accuracy in classifying medical images unless the neural net is trained on extremely large annotated samples. With many medical practices like the inflammatory bowel diseases (IBD), the annotated training data is relatively small requiring more additive techniques to bring focus to the context and then to predict important indicators like identifying acute ulcerative colitis regions of interest. This paper introduced the notion of thick data that can bring the required context by adding the expert heuristics in the form of anchor images and along with the use of Siamese neural network that uses triple ranking loss, it can rate the ulcerative colitis severity level according to the Mayo endoscopic (MES) scoring system. Many other thick data techniques can be added as cascading filters on this core model. We are anticipating many other results to follow in our future research.