Total Variation Algorithm Model with SOCP-based Partial Differential Equation
In order to reduce influence of various illumination conditions for face recognition, Total variation (TV) algorithm model is proposed. Mix SOCP-based partial differential equation technique with itto improve the recognition rate of TV model. First, use Radial Basis Functions (RBF) deformation technique to align key facial features of the reference 2D face with those of the target face. Then, employ optimized TV model to factorize each of the two aligned face images to an illumination-dependent componentand an illumination-invariant component. Finally, the illumination-dependent component of reference face replaces that of the target face, this can realize illumination transferring. This method does not require any prior information regarding the lighting conditions or the 3D geometries of the underlying faces. The optimized model smoothing filter preserves the edges better and can removes halo efficiently, parameter selection is also simpler.
Face Recognition Total Variation Model Partial Differential Equation Second-Order Cone Program
Kezheng Lin Chen Xin Ying Xu
Harbin University of Science and Technology, Harbin 150080, China
国际会议
哈尔滨
英文
221-225
2010-08-01（万方平台首次上网日期，不代表论文的发表时间）