RNN model-based classification of wireless capsule endoscopy bleeding images

Abstract

WCE (wireless capsule endoscopy) is a technique that may be used to diagnose gastrointestinal issues and provide painless gut imaging. Regardless, a variety of variables, such as effectiveness, tolerance, safety, and performance, make widespread use and modification challenging. Furthermore, automated analysis of the WCE data is essential for detecting anomalies. When a patient's digestive system is imaged using WCE, a vast amount of data is generated and these challenges have been addressed using a variety of computer assisted and vision-based technologies, but they do not achieve the essential level of precision, and further work is required. With this work, the goal is to create a system that can automatically analyze WCE images to identify problems and assist practitioners in making right diagnoses. Finally, a comparison of SODM-S1 with SODM shows that by modifying features to increase spatial dependency, our suggested technique may really improve model performance.