Kernel-aligned multi-view canonical correlation analysis for image recognition.pdfVIP

Infrared Physics Technology 78 (2016) 233–240 Contents lists available at ScienceDirect Infrared Physics Technology journal homepage: /locate/infrared Kernel-aligned multi-view canonical correlation analysis for image recognition Shuzhi Su, Hongwei Ge ?, Yun-Hao Yuan Key Laboratory of Advanced Process Control for Light Industry (Ministry of Education), School of Internet of Things Engineering, Jiangnan University, Wuxi, China highlights  We propose a kernel-aligned multi-view canonical correlation analysis method.  Each feature vector is transformed into a feature matrix by kernel alignment.  An optimized combination of multiple basis kernels is automatically learned.  Extensive experiments have shown the superiority of the method. article info Article history: Received 21 June 2016 Revised 25 August 2016 Accepted 25 August 2016 Available online 27 August 2016 Keywords: Canonical correlation analysis Multi-view feature learning Kernel alignment Image recognition abstract Existing kernel-based correlation analysis methods mainly adopt a single kernel in each view. However, only a single kernel is usually insuf?cient to characterize nonlinear distribution information of a view. To solve the problem, we transform each original feature vector into a 2-dimensional feature matrix by means of kernel alignment, and then propose a novel kernel-aligned multi-view canonical correlation analysis (KAMCCA) method on the basis of the feature matrices. Our proposed method can simultaneously employ multiple kernels to better capture the nonlinear distribution information of each view, so that correlation features learned by KAMCCA can have well discriminating power in real-world image recognition. Extensive experiments are designed on ?ve real-world image datasets, including NIR face images, thermal face images, visible face images, handwritten digit images, and object images. Promising experimental results on the datasets have manifested the effectiveness of our proposed meth