Offline surgical phase recognition

Offline surgical phase recognition is a subtask in the field of computer vision, aiming to automatically identify and annotate different phases of a surgical procedure by analyzing surgical video data. This task uses the first 40 videos for training and the last 40 videos for testing to evaluate the accuracy and robustness of the model. Its goal is to enhance the standardization and traceability of surgical processes, optimize operating room management, support postoperative quality control, and training education, making it of significant clinical application value.