Scholarlink

Pan-cancer of the digestive system, which includes multiple organs and their accessory structures, presents significant challenges for clinical diagnosis due to its complexity and diversity. Although computer-aided diagnostic (CAD) techniques [1] are used by doctors to assist in diagnosis, the need for extensive labeled data limits their effectiveness. To address this, we propose a few-shot learning method that combines transfer learning and contrast learning to enhance the diagnosis of digestive system pan-cancer. Our model, tested on histopathology image datasets of the esophagus, stomach, and rectum, achieves accuracy rates of 67.46%, 92.67%, and 94.99% with only 15 samples per category. This method significantly advances the integration of few-shot learning with CAD techniques, supporting efforts in precision medicine.

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

An Extended Few-Shot Learning-Based Approach for Histopathological Image Classification of Pan-Cancer in the Digestive System

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Rui Li,
  • Md Mamunur Rahaman,
  • Xiaoyan Li,
  • Hongzan Sun,
  • Jinzhu Yang,
  • Minghe Gao,
  • Marcin Grzegozek,
  • Tao Jiang,
  • Xinyu Huang,
  • Chen Li

摘要

Pan-cancer of the digestive system, which includes multiple organs and their accessory structures, presents significant challenges for clinical diagnosis due to its complexity and diversity. Although computer-aided diagnostic (CAD) techniques [1] are used by doctors to assist in diagnosis, the need for extensive labeled data limits their effectiveness. To address this, we propose a few-shot learning method that combines transfer learning and contrast learning to enhance the diagnosis of digestive system pan-cancer. Our model, tested on histopathology image datasets of the esophagus, stomach, and rectum, achieves accuracy rates of 67.46%, 92.67%, and 94.99% with only 15 samples per category. This method significantly advances the integration of few-shot learning with CAD techniques, supporting efforts in precision medicine.