Correlation filters tissue tracking with application to robotic minimally invasive surgery

Area of Science:

• Robotics
• Computer Vision
• Surgical Technology

Background:

• Vision-based tissue tracking is crucial for autonomous surgical robots.
• Challenges include occlusion, deformation, and appearance changes.

Purpose of the Study:

• To propose a novel correlation filter tissue tracking framework for minimally invasive surgery.
• To enhance the robustness and accuracy of autonomous surgical robot systems.

Main Methods:

• Developed a correlation filter framework with synthetic features and a bi-branch design.
• Integrated an incrementally learned detector with updating and trigger schemes for re-detection.

Main Results:

• Validated the framework on public tracking benchmark datasets.
• Developed a surgical tissue tracking dataset from the Cholec80 dataset for intra-operative scenes.

Conclusions:

• The proposed framework achieves outstanding performance, surpassing existing methods.
• Demonstrates the feasibility of correlation filters for effective tissue tracking in surgery.