Related Experiment Video
Updated: Dec 30, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
3.3K
Automatic lesion segmentation and classification of hepatic echinococcosis using a multiscale-feature convolutional
Shenghai Xin1,2, Huabei Shi1, A Jide2
1Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing, 100084, China.
Medical & Biological Engineering & Computing
|January 18, 2020
Summary
This study introduces an AI network for precise diagnosis of hepatic echinococcosis (HE). The system accurately segments and classifies HE lesions from CT scans, improving diagnostic capabilities for this parasitic liver disease.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Parasitology
Background:
- Hepatic echinococcosis (HE) is a severe parasitic liver disease necessitating accurate diagnosis and treatment.
- Current diagnostic methods for HE can be challenging, requiring precise identification and characterization of lesions.
Purpose of the Study:
- To develop and evaluate a novel automatic network for segmenting and classifying hepatic echinococcosis (HE) lesions.
- To improve the accuracy and efficiency of HE diagnosis using deep learning techniques.
Main Methods:
- A novel automatic HE lesion segmentation and classification network was proposed, featuring lesion region positioning (LRP) and lesion region segmenting (LRS) modules.
- A convolutional neural network (CNN) was employed for classifying HE lesion types.
- The system was trained and validated on CT slices from 160 patients, with expert radiologist delineation.
Main Results:
- The automatic segmentation achieved a high Dice score of 89.89%.
- Classification performance metrics included Dice scores of 80.32% (cystic vs. alveolar) and 82.45% (calcified vs. noncalcified).
- High sensitivity, specificity, NPV, PPV, and ROC AUC values demonstrated the network's diagnostic accuracy.
Conclusions:
- The proposed automatic network effectively segments and classifies hepatic echinococcosis lesions from CT images.
- This AI-driven approach shows significant potential for enhancing the precision and reliability of HE diagnosis.
