Computer-Aided Diagnosis of Hepatocellular Carcinoma Cells Based on MRI Imaging

• Published in: 2025 2nd International Conference on Intelligent Computing and Robotics (ICICR) pp. 1234 - 1241
• Main Authors: Yuan, Letian, Yuan, Yuchun
• Format: Conference Proceeding
• Language: English
• Published: IEEE 16.05.2025
• Subjects: Accuracy; Classification algorithms; Clustering algorithms; Computer aided diagnosis; Design automation; Feature extraction; Hepatocellular Carcinoma; Magnetic resonance imaging; Malignancy grading; MRI; Multi-classifier fusion algorithm; Nearest neighbor methods; Reliability; Robots
• DOI: 10.1109/ICICR65456.2025.00219

This paper presents a comprehensive study on the Computer-Aided Diagnosis (CAD) of Hepatocellular Carcinoma (HCC) cells based on Magnetic Resonance Imaging (MRI). Given the high incidence of HCC, particularly in China, and the significant recurrence rate, there is a pressing need for accurate and non-invasive diagnostic methods. The study focuses on the development of a stratified strategy-based adaptive weighted multi-classifier fusion algorithm. This algorithm assigns multiple single-layer classifiers for Region of Interest (ROI) image multi-layer features, enabling automatic malignancy grading of HCC. The proposed multi-classifier fusion algorithm, based on K-nearest neighbors and clustering principles, allocates different weights to each layer classifier. The results demonstrate that the feature extraction module design and the hierarchical adaptive weighted multi-classifier fusion algorithm significantly improve the diagnostic accuracy. In conclusion, the HCC malignancy grading CAD system provides a highly reliable auxiliary diagnosis for assessing the malignancy level of HCC.