no code implementations • 10 Jul 2017 • Changyong Yu, Yunde Jia
We use the anisotropic diffusion to enhance the edges and boundary locations of a face image, and the kernel matrix model to extract face image features which we call the diffusion-kernel (D-K) features.
