Paper
Study on Support Vector Machine for Image Reconstruction Algorithm of Electrical Capacitance Tomography
- Authors:
- SHEN Lili; HUANG Jingluo; Han Yuhui
- Abstract
- Support vector machine (SVM) is based on the special small samples theory with strong generalization ability, and is selected as an optimal theory for small samples classify problem. Electrical capacitance tomography (ECT) is a typical small samples and nonlinear mapping problem. In this paper the ECT image reconstruction algorithm based on SVM is proposed and a novel training method is proposed to improve the efficiency of SVM classifier by selecting active penalty parameters. Programming software based on Matlab6.5 and VC6.0, the simulation and experiment results indicates this algorithm has stronger space resolution and generalization ability, overall performance of the algorithm is better than some classic reconstruction algorithm such as LBP, etc. However, the algorithm requires the higher quality of training samples, otherwise less real-time, loss of application value.
- Keywords
- support vector machine; electrical capacitance tomography; image reconstruction; space resolution; generalization ability
- StartPage
- 55
- EndPage
- 58
- Doi