在结肠镜检查中,用于人工智能辅助的多体检测的最佳标签方法

Yen-Po Wang1, Ying-Chun Jheng2, Ming-Chih Hou3

  • 1Endoscopy Center for Diagnosis and Treatment, Taipei Veterans General Hospital, Taiwan; Division of Gastroenterology, Taipei Veterans General Hospital, Taiwan; Institute of Brain Science, National Yang Ming Chiao Tung University School of Medicine, Taiwan; Faculty of Medicine, National Yang Ming Chiao Tung University School of Medicine, Taiwan.

概括

标准化结肠多标签对于准确的AI模型至关重要. 在注释中将边界框延长20%显著提高了AI多胞体检测准确性和性能.