Related Experiment Video
Updated: Jun 25, 2025
09:41
Estimation of Contact Regions Between Hands and Objects During Human Multi-Digit Grasping
Published on: April 21, 2023
1.6K
ERegPose: An explicit regression based 6D pose estimation for snake-like wrist-type surgical instruments.
Jinhua Li1, Zhengyang Ma2, Xinan Sun2
1Key Lab for Mechanism Theory and Equipment Design of Ministry of Education, School of Mechanical Engineering, Tianjin University, Tianjin, China.
Summary
Estimating the 6D pose of flexible surgical instruments is difficult. ERegPose, a new deep learning strategy, accurately determines instrument pose, improving surgical precision.
Area of Science:
- Robotics
- Computer Vision
- Medical Technology
Background:
- Accurate 6D pose estimation of snake-like wrist-type surgical instruments is complex due to their intricate kinematics and flexible nature.
- Existing methods struggle with the unique challenges posed by these advanced surgical tools.
Purpose of the Study:
- To develop a precise and reliable method for estimating the 6D pose of snake-like wrist-type surgical instruments.
- To introduce ERegPose, a novel strategy combining a deep neural network and a specialized dataset.
Main Methods:
- ERegPoseNet, a deep neural network, was designed for explicit 6D pose regression.
- An in-house dataset of simulated surgical operations was annotated for training and validation.
- A Single Shot multibox Detector (SSD)-like detector was utilized to capture rotational features by generating instrument tip bounding boxes.
Main Results:
- ERegPoseNet achieved a 3D translation error of 1.056 mm and a 3D rotation error of 0.073 rad.
- The average distance (ADD) metric was 3.974 mm, indicating high overall spatial transformation accuracy.
- Experimental validation confirmed the necessity of the SSD-like detector and L1 loss for optimal performance.
Conclusions:
- ERegPose significantly outperforms existing methods in 6D pose estimation for snake-like surgical instruments.
- The proposed strategy offers a promising solution for enhancing precision in various surgical applications.
- Accurate pose estimation is crucial for advancing robotic-assisted surgery and improving patient outcomes.
