Automatic Abnormality Detection System for Capsule Endoscopy

• Published in: Journal of Information Science and Engineering Vol. 36; no. 5; pp. 955 - 966
• Main Authors: Jani, Kuntesh K, Anand, Animesh, Srivastava, Subodh, Srivastava, Rajeev
• Format: Journal Article
• Language: English
• Published: Taipei 社團法人中華民國計算語言學學會 01.09.2020; Institute of Information Science, Academia Sinica
• Subjects: Abnormalities; Diagnosis; Electronic devices; Endoscopy; Hybrid systems; Image analysis; Image processing; Machine learning; Medical imaging; Performance evaluation; Physicians; deep learning; transfer learning; capsule endoscopy; medical image analysis; CAD
• ISSN: 1016-2364
• DOI: 10.6688/JISE.202009_36(5).0001

One-third of the world population suffers from diseases related to gastrointestinal (GI) tract. Capsule endoscopy (CE) is a non-sedative, hygienic, non-invasive and patient-friendly and particularly child-friendly technology to scan the entire GI tract. However, CE generates nearly 60000 images, which make the diagnosis process time consuming and tiresome for physicians. Also, the diagnosis is highly subjective and varies from person to person. Thus, a computer-aided diagnosis (CAD) system is a must. This study addresses a multi-class medical image analysis problem using image processing and machine learning techniques. It presents a CAD system based on the hybrid confluence of transfer learning and conventional machine learning technique for automatic abnormality detection in the GI tract. The system performs with an accuracy of 96.89%. The rigorous performance evaluation shows that the system is capable of fast and accurate diagnosis of GI tract abnormalities. Such a system can be beneficial to physicians and with the advancement of smart devices and IoT, such a system can prove to be a handy remote diagnosis tool for geographically distant locations where an expert of the subject may not be available.