Abstract：Considering the different contributions of different facial components to face analysis,e.g. eyes,mouth etc.,a face analysis based on multi-component sparse coding is proposed. Firstly,some facial components which play important role to face analysis are selected. Then,the dictionaries of multiple components are learnt by using multi-view sparse coding algorithm,and the sparse codes of each face image are computed based on the dictionary. The final decision is made through pooling the sparse codes into support vector machines and least squares classifiers. Face analysis experiments include face recognition,facial expression recognition,face recognition with occlusion,and facial expression recognition with occlusion. The experimental results show that the proposed method based on multi-component sparse coding learns optimal weights of different facial components and outperforms single facial component method and simple multi-component fusion method.
Key words： Face Component Face Analysis Sparse Coding Face Recognition Facial Expression Recognition
Sparse coding represents a signal sparsely by using an overcomplete dictionary, and obtains promising performance in practical computer vision applications, especially for signal restoration tasks such as image denoising and image inpainting. In recent years, many discriminative sparse coding algorithms have been developed for classification problems, but they cannot naturally handle visual data represented by multiview features. In addition, existing sparse coding algorithms use graph Laplacian to model the local geometry of the data distribution. It has been identified that Laplacian regularization biases the solution towards a constant function which possibly leads to poor extrapolating power. In this paper, we present multiview Hessian discriminative sparse coding (mHDSC) which seamlessly integrates Hessian regularization with discriminative sparse coding for multiview learning problems. In particular, mHDSC exploits Hessian regularization to steer the solution which varies smoothly along geodesics in the manifold, and treats the label information as an additional view of feature for incorporating the discriminative power for image annotation. We conduct extensive experiments on PASCAL VOC'07 dataset and demonstrate the effectiveness of mHDSC for image annotation.