Transfer Learning for Surgical Task Segmentation

In this paper, we present a novel approach for surgical task segmentation. A segmentation policy learns the correlations between features and segmentation points from manually labeled data. The most correlated features and rules for segmenting them are identified and learned. These form a complete set of segmentation policy. The proposed approach is developed to segment new but similar tasks through transfer learning. It is verified through applying the segmentation rule learned from the labeled data to segment other tasks. The performance of the proposed algorithm was evaluated by comparing the results against the ground truths. Experimental results demonstrate that our approach can achieve high segmentation rates with an accuracy of between 68. 8% - 81. 8%.

Authors

• Ya-Yen Tsai
• Bidan Huang
• Yao Guo 0002
• Guang-Zhong Yang

Keywords

• Task analysis
• Motion segmentation
• Trajectory
• Feature extraction
• Needles
• Surgery
• Transfer Learning
• Segmentation Task
• Surgical Tasks
• Segmentation Approach
• Manual Labeling
• Segmentation Points
• Training Data
• Training Dataset
• Transition State
• Public Databases
• Precision And Recall
• Segmentation Algorithm
• Segmentation Results
• Labeled Data
• Automatic Segmentation
• Unsupervised Algorithm
• Angular Speed
• Motion Trajectory
• Candidate Features
• Motion Primitives
• Model-free Approach
• Inverse Reinforcement Learning
• Linear Speed
• Needle Holder
• Average Recall
• Dirichlet Process
• Degrees Of Freedom
• Kinematic
• Multiple Frames