Online surgical phase recognition

Online surgical phase recognition is an important subtask in the field of computer vision, aiming to automatically detect and identify different phases during surgery through real-time analysis of surgical videos. The goal of this task is to enhance the standardization and automation level of surgical procedures, optimize operating room management and resource allocation, while assisting surgeons in decision support and training evaluation. Using the first 40 video segments for training and the last 40 video segments for testing ensures the accuracy and robustness of the model.