Multi-View Feature Transformation Based SVM+ for Computer-Aided Diagnosis of Liver Cancers With Ultrasound Images

• Published in: IEEE journal of biomedical and health informatics Vol. 27; no. 3; pp. 1512 - 1523
• Main Authors: Zhang, Huili, Guo, Lehang, Wang, Jun, Ying, Shihui, Shi, Jun
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.03.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Classification algorithms; Cognitive tasks; Diagnosis; Image contrast; Image enhancement; Knowledge management; learning using privileged information; Liver; Liver cancer; Machine learning; Machine learning algorithms; Medical diagnosis; Medical imaging; Multi-view transfer learning; radius-margin; Support vector machines; Task analysis; Transfer learning; Transformations (mathematics); Ultrasonic imaging; Ultrasound; ultrasound imaging
• ISSN: 2168-2194; 2168-2208; 2168-2208
• DOI: 10.1109/JBHI.2022.3233717

It is feasible to improve the performance of B-mode ultrasound (BUS) based computer-aided diagnosis (CAD) for liver cancers by transferring knowledge from contrast-enhanced ultrasound (CEUS) images. In this work, we propose a novel feature transformation based support vector machine plus (SVM+) algorithm for this transfer learning task by introducing feature transformation into the SVM+ framework (named FSVM+). Specifically, the transformation matrix in FSVM+ is learned to minimize the radius of the enclosing ball of all samples, while the SVM+ is used to maximize the margin between two classes. Moreover, to capture more transferable information from multiple CEUS phase images, a multi-view FSVM+ (MFSVM+) is further developed, which transfers knowledge from three CEUS images from three phases, i.e., arterial phase, portal venous phase, and delayed phase, to the BUS-based CAD model. MFSVM+ innovatively assigns appropriate weights for each CEUS image by calculating the maximum mean discrepancy between a pair of BUS and CEUS images, which can capture the relationship between source and target domains. The experimental results on a bi-modal ultrasound liver cancer dataset demonstrate that MFSVM+ achieves the best classification accuracy of 88.24±1.28%, sensitivity of 88.32±2.88%, specificity of 88.17±2.91%, suggesting its effectiveness in promoting the diagnostic accuracy of BUS-based CAD.