Related Experiment Video
Updated: Jul 3, 2025
05:57
Long-term Video Tracking of Cohoused Aquatic Animals: A Case Study of the Daily Locomotor Activity of the Norway Lobster Nephrops norvegicus
Published on: April 8, 2019
6.9K
Self-supervised context-aware correlation filter for robust landmark tracking in liver ultrasound sequences.
Lin Ma1, Junjie Wang1, Shu Gong2
1College of Computer Science and Engineering, Chongqing University of Technology, Chongqing, China.
Biomedizinische Technik. Biomedical Engineering
|February 14, 2024
Summary
This study introduces a self-supervised method for accurate liver landmark tracking in ultrasound images, crucial for image-guided radiation therapy. The novel approach achieves high speed and accuracy, outperforming existing methods.
Area of Science:
- Medical Imaging
- Radiation Therapy
- Machine Learning
Background:
- Respiratory motion causes organ displacement, challenging accurate landmark tracking in image-guided radiation therapy.
- Liver landmark tracking accuracy is particularly affected by these motion-induced displacements.
Purpose of the Study:
- To develop a self-supervised method for robust landmark tracking in long liver ultrasound sequences.
- To mitigate the impact of speckle noise and artifacts on ultrasonic image tracking.
Main Methods:
- A Siamese-based context-aware correlation filter network was employed.
- Training utilized consistency loss between forward tracking and back verification using labeled and unlabeled data.
- A fusion strategy for template patch features enhanced appearance information for tracking.
Main Results:
- The proposed method achieved a mean tracking error of 0.79 ± 0.83 mm.
- A high frame rate of 118.6 fps was reached, indicating a superior speed-accuracy trade-off.
- The method secured 5th place in the CLUST2015 2D point-landmark tracking task.
Conclusions:
- Extensive experiments validated the effectiveness of the proposed self-supervised approach.
- The technique demonstrated top-tier performance on the CLUST2015 leaderboard at submission.
- The method offers a robust solution for liver landmark tracking in challenging ultrasound sequences.
