Scholarlink

Peptic ulcers and other digestive tract disorders pose significant diagnostic challenges, often requiring expert analysis of endoscopic images. In this study, we aim to assist in the examination of gastrointestinal (GI) tract endoscopy images using advanced image classification techniques. Utilizing the Kvasir dataset, which comprises 3500 labeled images, we applied Contrast Limited Adaptive Histogram Equalization (CLAHE) for image pre-processing to enhance feature extraction. We implemented convolutional neural networks (CNNs) with InceptionV3 and ResNet50V2 architectures to classify endoscopic images into four categories: normal, esophagitis, polyps, and ulcerative colitis. Our model achieved a training accuracy of 100.00% and a validation accuracy of 92.71% with the ResNet50V2 model, demonstrating its effectiveness in identifying digestive tract disorders. This research highlights the potential of CNNs to aid medical professionals by providing a tool for the automatic classification of endoscopic images, thereby enhancing the diagnostic process for peptic ulcers and other GI tract conditions.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Development of a Deep Learning Model for Accurate Classification of Peptic Ulcer and Digestive Tract Disorders Using Endoscopic Images

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Pulak Islam,
  • Md. Asif Hasan,
  • Ishtiaq Hoque Farabi,
  • S. M. Sabbir Sobhani,
  • K. M. Safin Kamal,
  • Ahmed Wasif Reza

摘要

Peptic ulcers and other digestive tract disorders pose significant diagnostic challenges, often requiring expert analysis of endoscopic images. In this study, we aim to assist in the examination of gastrointestinal (GI) tract endoscopy images using advanced image classification techniques. Utilizing the Kvasir dataset, which comprises 3500 labeled images, we applied Contrast Limited Adaptive Histogram Equalization (CLAHE) for image pre-processing to enhance feature extraction. We implemented convolutional neural networks (CNNs) with InceptionV3 and ResNet50V2 architectures to classify endoscopic images into four categories: normal, esophagitis, polyps, and ulcerative colitis. Our model achieved a training accuracy of 100.00% and a validation accuracy of 92.71% with the ResNet50V2 model, demonstrating its effectiveness in identifying digestive tract disorders. This research highlights the potential of CNNs to aid medical professionals by providing a tool for the automatic classification of endoscopic images, thereby enhancing the diagnostic process for peptic ulcers and other GI tract conditions.