Extensive numerical experiments were implemented on AR, Extended Yale B and CMU-PIE databases to verify the performance of VFRC for image-to-image SSPP problem under the complication conditions, broad changing in illumination conditions and change of pose and expression conditions, respectively.