Image classification using segmentation, functional principal component analysis and sparse sufficient dimension reduction
Wenqian Liang · Digital Commons-TMC (Texas Medical Center) · 2016
With the development of digital technologies, image classification has become an important method for improving the accuracy and reliability of disease diagnosis. Automatic image classification can be a powerful tool for computer-aided disease diagnosis which provides fast and reliable disease diagnosis. We proposed a novel method for automatic image classification which combines image segmentation, three-dimensional functional principal component analysis (FPCA) and sparse sufficient dimension reduction (SDR). To evaluate the performance of the proposed method for image classification analysis, we applied it to a total of 118 volunteers’ Magnetic resonance imaging (MRI) data and classify them into ischemic group and normal group. The proposed method was compared with the Wavelet PCA-based method and sparse logistic regression. The results showed that the proposed method outperformed the Wavelet PCA-based method and sparse logistic regression.