Dual-Model Ensemble Framework with Anatomical Context for Rare Event Detection in Capsule Endoscopy Videos
Abstract
This work presents a dual-model deep learning framework for rare event detection in capsule endoscopy videos. An ensemble-based rare event detection model is combined with an anatomy-aware gastrointestinal region detection model to improve reliability. Predictions from both models are merged during inference to reduce false positives and improve detection of rare abnormalities.
Fields of Study & Subjects
Machine Learning in Healthcare
Gastrointestinal Bleeding Diagnosis and Treatment
COVID-19 diagnosis using AI
Publication details
- Publisher
- Wiley
- Publication date
- 17 March 2026
- Document type
- Preprint