ESAD (SARAS Endoscopic Surgeon Action Detection)

ESAD 是一个大型数据集，旨在解决内窥镜微创手术中的外科医生动作检测问题。 ESAD 旨在通过实际测试他们对外科医生执行的动作的意识来提高手术辅助机器人的有效性和可靠性。该数据集为前列腺切除术期间捕获的真实内窥镜视频帧上的 21 个动作类别提供边界框注释，并被用作最近 MIDL 2020 挑战的基础。

ESAD is a large-scale dataset designed to address the problem of surgeon action detection in minimally invasive endoscopic surgery. It aims to improve the effectiveness and reliability of surgical assistant robots by practically testing their awareness of the actions performed by surgeons. This dataset provides bounding box annotations for 21 action categories on real endoscopic video frames captured during prostatectomy, and served as the foundation for the recent MIDL 2020 Challenge.

搜集汇总

背景与挑战

背景概述

ESAD是一个大型内窥镜手术数据集，专注于检测外科医生在微创手术中的动作，以提升手术辅助机器人的性能。该数据集基于前列腺切除术的真实视频帧，提供了21个动作类别的边界框标注，并于2021年发布，曾作为MIDL 2020挑战的基准。