Summary of Ossar: Towards Open-set Surgical Activity Recognition in Robot-assisted Surgery, by Long Bai et al.
OSSAR: Towards Open-Set Surgical Activity Recognition in Robot-assisted Surgery
by Long Bai, Guankun Wang, Jie Wang, Xiaoxiao Yang, Huxin Gao, Xin Liang, An Wang, Mobarakol Islam, Hongliang Ren
First submitted to arxiv on: 10 Feb 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Robotics (cs.RO)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper presents a novel framework, Open-Set Surgical Activity Recognition (OSSAR), for robotic surgical activity recognition in open-set scenarios. Existing algorithms struggle with real-world challenges, as they are designed for closed-set paradigms and fail to recognize test samples from unseen classes during training. OSSAR leverages hyperspherical reciprocal points to enhance distinction between known and unknown classes and refines model calibration to prevent misclassification of unknown classes as known ones. The framework is evaluated on two public datasets: JIGSAWS and a novel dataset for endoscopic submucosal dissection. Results show significant outperformance over state-of-the-art approaches, highlighting OSSAR’s effectiveness in addressing real-world surgical challenges. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine robots doing surgery without human help. To make this happen, we need to teach the robot what to do during surgery. The problem is that most robots are trained for specific tasks and can’t handle unexpected situations. This paper proposes a new way to recognize robotic surgical activities in these unpredictable situations. Our approach uses a special strategy to tell known and unknown actions apart and calibrates our model to avoid mistakes. We test our method on two public datasets and show that it performs better than existing methods. This means our solution can be used in real-world surgery scenarios. |
Keywords
» Artificial intelligence » Activity recognition
