Stacking Ensemble Learning for Accurate Polyp Segmentation

• Published in: International Conference on Bio-engineering for Smart Technologies (Online) pp. 1 - 5
• Main Authors: Chetoui, Mohamed, Akhloufi, Moulay A.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 14.05.2025
• Subjects: Accuracy; colonoscopy; Colorectal cancer; Computational modeling; convolutional neural network; Deep learning; Ensemble learning; Image segmentation; Polyp segmentation; Robustness; Scalability; Shape; Stacking
• ISSN: 2831-4352
• DOI: 10.1109/BioSMART66413.2025.11046145

Polyp segmentation in colonoscopy images is a crucial task for early detection and prevention of colorectal cancer. In this study, we propose an ensemble learning approach combining ResNet50-U-Net and ResNet50-Polyp, two fine-tuned deep learning models designed for gastrointestinal disease segmentation. The models were trained on the Kvasir-SEG dataset, and ensemble learning techniques were applied to enhance segmentation accuracy. Our approach demonstrates state-of-the-art performance, achieving an accuracy of 0.9689, precision of 0.9200, recall of 0.8800, F1-Score of 0.9065, and IoU of 0.7544. Comparative analysis with existing methods confirms the robustness and efficiency of our model in accurately identifying Polyp. The proposed ensemble model offers a reliable solution for Polyp segmentation, contributing to improved diagnostic in clinical settings.