Scholarlink

This research extends the use of non-contrast CT scans and deep learning for pancreatic cancer detection. CT scans from 82 patients, annotated for various cancer categories, were analyzed using You Only Look Once (YOLO) models. YOLOv7 achieved a validation accuracy of 94.7%, YOLOv8 maintained a similar performance of 93.1%, and YOLOv9 achieved the highest accuracy at 95.4%. These findings demonstrate that YOLOv7 and YOLOv9 effectively differentiate between normal and exocrine/neuroendocrine pancreatic cancers, indicating their reliability for medical image classification tasks.

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

Deep Learning and Non-Contrast CT for Widespread Pancreatic Cancer Detection

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • D. M. Rafiun Bin Masud,
  • Sayed Fuad Al Labib,
  • Arpon Bhattacharjee,
  • Md. Mottakin Rahat,
  • K. M. Safin Kamal,
  • Ahmed Wasif Reza

摘要

This research extends the use of non-contrast CT scans and deep learning for pancreatic cancer detection. CT scans from 82 patients, annotated for various cancer categories, were analyzed using You Only Look Once (YOLO) models. YOLOv7 achieved a validation accuracy of 94.7%, YOLOv8 maintained a similar performance of 93.1%, and YOLOv9 achieved the highest accuracy at 95.4%. These findings demonstrate that YOLOv7 and YOLOv9 effectively differentiate between normal and exocrine/neuroendocrine pancreatic cancers, indicating their reliability for medical image classification tasks.