Surgeons-Robotic Interaction Scene Classification Dataset

The current medical industry faces challenges of insufficient precision and lack of smooth human-robot collaboration in the application of surgical robots. Existing solutions often lack high-quality training data, which limits the intelligent decision-making and operational capabilities of robots. This dataset aims to solve the recognition and judgment problems of robots in complex surgical environments by providing rich images of interaction scenes between surgeons and robots. Data collection uses high-resolution cameras in real surgical environments to ensure scene authenticity. Multiple rounds of annotation and expert review ensure data quality and consistency. The data storage format is JPEG, organized by time and scene to facilitate subsequent analysis and model training. The dataset contains a total of 5000 images, with a file size of approximately 1.5G. The core advantage of this dataset lies in its high annotation accuracy and consistency, with annotation accuracy exceeding 95%. Additionally, the use of new data augmentation technologies enhances the model's adaptability in different scenes, expected to increase the recognition rate of surgical robots by 20% during operations. Applying this dataset can effectively improve the intelligence level of medical robots in surgical operations to meet actual clinical needs.