张量框架下的多线性人脸识别技术的分析.pdfVIP

Abstract Tensor-based multilinear approaches regard human face as the composite consequence of geometries, viewpoints and illuminations, and use tensor decomposing algebra to get factor (i.e viewpoint, illumination, person, etc) subspaces for recognition. The algorithm allows for bigger changes in viewpoints and illuminations than linear subspace, and is less computationally expensive than non-linear methods. Based on the method, the major work done by the paper is summarized as follows: