SLAM - Surgical LAparoscopic Motions

Laparoscopic surgery has been widely used in various surgical fields due to its minimally invasive and rapid recovery benefits. However, it requires surgeons to have exceptional technical expertise to overcome challenges such as limited surgical space, restricted visual feedback, and the need for high precision. The rapid development of computer vision and deep learning in surgical action recognition provides essential technical support for improving surgical navigation, training, and postoperative evaluation. Despite this progress, existing publicly available datasets have limited scale, high homogeneity, and inconsistent labeling quality, making it challenging to meet the needs of deep learning models. To address the above issues, we developed the SLAM dataset (Surgical LAparoscopic Motions), which encompasses various surgical types. The dataset includes annotations for seven key actions: Abdominal Entry, Use Clip, Hook Cut, Suturing, Panoramic View, Local Panoramic View, and Suction. Experienced medical experts meticulously reviewed the annotations to ensure high quality and clinical accuracy. The SLAM dataset aims to promote the development of laparoscopic surgical action recognition and artificial intelligence-driven surgery, supporting intelligent surgical robots and surgical automation.

腹腔镜手术因具备微创性与术后快速恢复的优势，已在各类外科领域得到广泛应用。但该手术要求外科医师具备卓越的专业技术水平，以应对手术空间狭小、视觉反馈受限、操作精度要求高等挑战。计算机视觉与深度学习在手术动作识别领域的快速发展，为手术导航、术者培训及术后评估的优化提供了关键技术支撑。尽管取得了上述进展，但当前公开可用的手术数据集普遍存在规模有限、同质化程度高、标注质量参差不齐等问题，难以满足深度学习模型的训练需求。为解决上述问题，本研究构建了SLAM（Surgical LAparoscopic Motions，手术腹腔镜运动）数据集，该数据集涵盖多种手术类型。该数据集包含7类核心手术动作的标注：腹部入路（Abdominal Entry）、施夹（Use Clip）、钩切（Hook Cut）、缝合（Suturing）、全景视野（Panoramic View）、局部全景视野（Local Panoramic View）与抽吸（Suction）。本研究邀请资深医学专家对所有标注结果进行了细致审核，以确保数据集的高质量与临床准确性。本SLAM数据集旨在推动腹腔镜手术动作识别与人工智能驱动外科的发展，为智能手术机器人及手术自动化技术的研发提供支撑。