Influence of color correction on pathology detection in Capsule Endoscopy

2025-01-31

AI Summary

This study systematically investigates the impact of color correction on pathological object detection performance in wireless capsule endoscopy (WCE) images. We applied two color correction functions to preprocess images from the standard SEE-AI dataset and integrated them into RetinaNet and YOLOv5 detection frameworks for controlled evaluation. Results show that while color correction significantly increases predicted bounding box sizes and intersection-over-union (IoU) with ground truth, it fails to yield consistent improvements in F1-score or average precision at IoU=0.5 (AP₅₀); notably, false positive rates rise markedly for certain lesions—e.g., erythema and ulcers. To our knowledge, this work is the first to quantitatively refute the implicit assumption that color correction inherently enhances detection accuracy in WCE analysis. It reveals previously underappreciated risks associated with color correction in medical image analysis and provides critical empirical guidance for robust algorithm design and preprocessing strategy selection. The source code is publicly available.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Calibration & Uncertainty Quantification
• Search and Optimization: Evaluation and Analysis

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Web Mining and Content Analysis: Web data provenance, reliability, and authenticity
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Pathology detection in Wireless Capsule Endoscopy (WCE) using deep learning has been explored in the recent past. However, deep learning models can be influenced by the color quality of the dataset used to train them, impacting detection, segmentation and classification tasks. In this work, we evaluate the impact of color correction on pathology detection using two prominent object detection models: Retinanet and YOLOv5. We first generate two color corrected versions of a popular WCE dataset (i.e., SEE-AI dataset) using two different color correction functions. We then evaluate the performance of the Retinanet and YOLOv5 on the original and color corrected versions of the dataset. The results reveal that color correction makes the models generate larger bounding boxes and larger intersection areas with the ground truth annotations. Furthermore, color correction leads to an increased number of false positives for certain pathologies. However, these effects do not translate into a consistent improvement in performance metrics such as F1-scores, IoU, and AP50. The code is available at https://github.com/agossouema2011/WCE2024. Keywords: Wireless Capsule Endoscopy, Color correction, Retinanet, YOLOv5, Detection

Problem

• Color Calibration
• Wireless Capsule Endoscopy
• Disease Detection

Innovation

• Color Adjustment
• Wireless Capsule Endoscopy
• Disease Detection