Related Experiment Video
Updated: Aug 13, 2025
13:01
Industrialized, Artificial Intelligence-guided Laser Microdissection for Microscaled Proteomic Analysis of the Tumor Microenvironment
Published on: June 3, 2022
3.8K
An Intelligent Auxiliary Framework for Bone Malignant Tumor Lesion Segmentation in Medical Image Analysis
Xiangbing Zhan1, Jun Liu2, Huiyun Long1
1State Key Laboratory of Public Big Data, College of Computer Science and Technology, Guizhou University, Guiyang 550025, China.
Diagnostics (Basel, Switzerland)
|January 21, 2023
Summary
Accurate segmentation of bone malignant tumors is vital for patient survival. A new deep learning framework, SEAGNET, enhances medical image analysis for precise tumor detection and improved diagnostic efficiency.
Area of Science:
- Oncology
- Medical Imaging
- Artificial Intelligence
Background:
- Bone malignant tumors are aggressive with poor prognoses, necessitating rapid and accurate diagnosis for effective treatment and improved survival rates.
- Accurate segmentation of tumorous lesions in medical images is challenging due to complex backgrounds and indistinct boundaries.
- Existing deep learning research for bone tumor segmentation is limited, particularly in handling complex image characteristics.
Purpose of the Study:
- To develop an intelligent auxiliary framework for segmenting bone malignant tumor lesions in medical images.
- To improve the accuracy and efficiency of diagnosing bone malignant tumors through advanced deep learning techniques.
- To address the limitations in current research concerning the segmentation of bone tumors with complex imaging features.
Main Methods:
- Proposed a supervised edge-attention guidance segmentation network (SEAGNET) for medical image segmentation.
- Incorporated a boundary key points selection module to enhance edge attention learning and retain fine-grained edge information.
- Utilized instance segmentation networks and mixed attention mechanisms to precisely locate tumors and capture contextual information for boundary ambiguity.
Main Results:
- Achieved superior performance compared to state-of-the-art segmentation methods.
- Obtained a Dice similarity coefficient of 0.967, precision of 0.968, and accuracy of 0.996.
- Demonstrated the framework's effectiveness in real-world medical data.
Conclusions:
- The SEAGNET framework significantly contributes to improving diagnostic accuracy and clinical efficiency in identifying bone malignant tumors.
- The proposed method effectively handles complex backgrounds and blurred boundaries in medical images.
- This intelligent auxiliary framework shows great potential in assisting clinicians with precise tumor diagnosis and treatment planning.
