EndoOOD: Uncertainty-aware Out-of-distribution Detection in Capsule Endoscopy Diagnosis
Qiaozhi Tan, Long Bai, Guankun Wang, Mobarakol Islam, Hongliang Ren
TL;DR
EndoOOD addresses OOD detection in wireless capsule endoscopy by integrating uncertainty-aware mixup training, long-tailed in-distribution data calibration, and ViM-based calibrated inference. The approach decouples mixup to preserve calibration, reweights calibration for imbalanced ID data, and leverages ViM to differentiate ID and OOD signals at inference. Across Kvasir-Capsule and CIFAR-10 benchmarks, EndoOOD achieves superior OOD detection while boosting ID accuracy, demonstrating robust uncertainty estimation and reliable rejection of OOD inputs. This framework advances practical, uncertainty-aware OOD handling in medical imaging and could improve reliability in WCE diagnostics.
Abstract
Wireless capsule endoscopy (WCE) is a non-invasive diagnostic procedure that enables visualization of the gastrointestinal (GI) tract. Deep learning-based methods have shown effectiveness in disease screening using WCE data, alleviating the burden on healthcare professionals. However, existing capsule endoscopy classification methods mostly rely on pre-defined categories, making it challenging to identify and classify out-of-distribution (OOD) data, such as undefined categories or anatomical landmarks. To address this issue, we propose the Endoscopy Out-of-Distribution (EndoOOD) framework, which aims to effectively handle the OOD detection challenge in WCE diagnosis. The proposed framework focuses on improving the robustness and reliability of WCE diagnostic capabilities by incorporating uncertainty-aware mixup training and long-tailed in-distribution (ID) data calibration techniques. Additionally, virtual-logit matching is employed to accurately distinguish between OOD and ID data while minimizing information loss. To assess the performance of our proposed solution, we conduct evaluations and comparisons with 12 state-of-the-art (SOTA) methods using two publicly available datasets. The results demonstrate the effectiveness of the proposed framework in enhancing diagnostic accuracy and supporting clinical decision-making.