Scholarlink

We present a lesion-aware image captioning framework for ulcerative colitis (UC), integrating ResNet embeddings, Grad-CAM heatmaps, and CBAM-enhanced attention with a T5 decoder. Clinical metadata—including MES scores, bleeding, and vascular patterns—are incorporated as natural language prompts to guide caption generation. The resulting system produces structured, interpretable, and diagnostically aligned descriptions. Compared to previous approaches, our method improves both captioning quality and MES classification accuracy, offering a clinically meaningful tool for endoscopic reporting.

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

Lesion-Aware Visual-Language Fusion for Automated Image Captioning of Ulcerative Colitis Endoscopic Examinations

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Alexis Iván López Escamilla,
  • Gilberto Ochoa,
  • Sharib Ali

摘要

We present a lesion-aware image captioning framework for ulcerative colitis (UC), integrating ResNet embeddings, Grad-CAM heatmaps, and CBAM-enhanced attention with a T5 decoder. Clinical metadata—including MES scores, bleeding, and vascular patterns—are incorporated as natural language prompts to guide caption generation. The resulting system produces structured, interpretable, and diagnostically aligned descriptions. Compared to previous approaches, our method improves both captioning quality and MES classification accuracy, offering a clinically meaningful tool for endoscopic reporting.