Related Experiment Video
Updated: Feb 19, 2026
10:25
Deep Learning-Based Segmentation of Cryo-Electron Tomograms
Published on: November 11, 2022
10.9K
Automatic Segmentation of Ultrasound Tomography Image
Shibin Wu1,2, Shaode Yu1,2, Ling Zhuang3
1Institute of Biomedical and Health Engineering, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, 1068 Xueyuan Avenue, Shenzhen University Town, Shenzhen 518055, China.
Biomed Research International
|October 31, 2017
Summary
An automated GrabCut (AUGC) algorithm significantly improves ultrasound tomography (UST) image segmentation for breast density and cancer analysis. This method offers high accuracy and speed, benefiting large-scale medical studies.
Area of Science:
- Medical imaging
- Image processing
- Computational anatomy
Background:
- Ultrasound tomography (UST) image segmentation is crucial for medical applications like breast density estimation.
- Current segmentation methods are often slow and require extensive manual input.
- There is a need for automated and efficient UST image segmentation techniques.
Purpose of the Study:
- To develop an automated algorithm for UST image segmentation.
- To improve the efficiency and accuracy of segmentation compared to existing methods.
- To enable large-scale studies using UST imaging for breast cancer screening and pathological quantification.
Main Methods:
- An automated algorithm based on GrabCut (AUGC) was developed.
- Automated GrabCut initialization was designed for incomplete labeling.
- The algorithm was accelerated using multicore parallel programming.
- AUGC was applied to segment 32 in vivo UST volumetric images.
Main Results:
- AUGC achieved high accuracy with Dice coefficient (D) = 0.9275, Jaccard (J) = 0.8660, and False Positive (FP) = 0.0077.
- The algorithm processed a volumetric image in approximately 4 seconds on average.
- AUGC outperformed Confidence Connected Region Growing (CCRG), watershed, and Active Contour based Curve Delineation (ACCD).
Conclusions:
- The proposed AUGC algorithm provides an accurate and efficient solution for UST image segmentation.
- AUGC's speed and automation make it suitable for large-scale medical studies.
- This advancement supports improved breast cancer screening and pathological quantification using UST imaging.
