Related Experiment Video
Updated: Jan 16, 2026
06:39
Three-dimensional Location Approach with Silk Thread Guided Laparoscopic Segmentectomy for Liver Tumor
Published on: May 23, 2025
413
Liver Tumor Segmentation Based on Multi-Scale Deformable Feature Fusion and Global Context Awareness
Chenghao Zhang1, Lingfei Wang1, Chunyu Zhang2
1College of Intelligent Systems Science and Engineering, Harbin Engineering University, Harbin 150001, China.
Biomimetics (Basel, Switzerland)
|September 26, 2025
Summary
This study introduces a novel framework for liver tumor segmentation, improving accuracy for irregular tumors using advanced feature fusion and context modeling. The method enhances segmentation performance and clinical reliability.
Area of Science:
- Medical Imaging
- Computer Vision
- Artificial Intelligence
Background:
- Liver tumor segmentation is challenging due to heterogeneous and irregular morphologies.
- Accurate segmentation is crucial for diagnosis, treatment planning, and monitoring.
Purpose of the Study:
- To develop an advanced framework for robust liver tumor segmentation.
- To improve the adaptability and accuracy of automated segmentation models for complex tumor structures.
Main Methods:
- Proposed a framework integrating Deformable Large Kernel Attention (D-LKA) for irregular feature adaptation.
- Incorporated a Context Extraction (CE) module for enhanced global semantic modeling.
- Utilized a Dual Cross Attention (DCA) mechanism for improved cross-scale feature fusion via modified skip connections.
Main Results:
- Achieved superior segmentation accuracy and generalization performance compared to state-of-the-art models on combined LiTS, MSD Task08, and 3D-IRCADb01 datasets.
- Demonstrated enhanced Intersection over Union (IoU) and other key segmentation metrics.
- Feature visualizations provided insights into model interpretability and boundary delineation.
Conclusions:
- The proposed framework offers a novel and practical solution for precise liver tumor segmentation.
- The approach shows strong potential for clinical application and real-world deployment due to its accuracy and reliability.
