Related Experiment Video
Updated: Mar 15, 2026
Three-dimensional Location Approach with Silk Thread Guided Laparoscopic Segmentectomy for Liver Tumor
Published on: May 23, 2025
Liver Tumor Segmentation from MR Images Using 3D Fast Marching Algorithm and Single Hidden Layer Feedforward Neural
Trong-Ngoc Le1, Pham The Bao2, Hieu Trung Huynh3
1Faculty of Information Technology, Industrial University of Ho Chi Minh City, 12 Nguyen Van Bao, Go Vap District, Ho Chi Minh City, Vietnam; Faculty of Information Technology, University of Science, 227 Nguyen Van Cu, District 5, Ho Chi Minh City, Vietnam.
This study presents a new computerized method for segmenting liver tumors in MRI scans, achieving accurate boundary detection and volume estimation. The developed scheme offers a promising tool for improving liver tumor analysis in clinical settings.
Area of Science:
- Medical Imaging
- Computer-Aided Diagnosis
- Oncology
Background:
- Accurate liver tumor segmentation is crucial for diagnosis and treatment planning.
- Manual segmentation is time-consuming and subject to inter-observer variability.
Purpose of the Study:
- To develop and evaluate an automated scheme for liver tumor segmentation in T1-weighted MR images.
- To compare the automated segmentation results with manual segmentation by a radiologist.
Main Methods:
- A four-stage computerized scheme involving region of interest extraction, noise reduction, boundary enhancement, and a single hidden layer feedforward neural network (SLFN) for voxel classification.
- Utilized a 3D fast marching algorithm for initial region generation and postprocessing for boundary refinement.
Main Results:
- The scheme demonstrated a mean volumetric overlap error of 27.43% and a mean percentage volume error of 15.73%.
- Surface distance metrics included mean average surface distance of 0.58 mm, root mean square surface distance of 1.20 mm, and maximal surface distance of 6.29 mm.
Conclusions:
- The proposed computerized scheme provides accurate liver tumor segmentation in MR images.
- The method shows potential for assisting radiologists in liver tumor analysis and reducing segmentation variability.
