A novel spatial information recovery algorithm based on fuzzy clustering in H.264

As no reference frame can be obtained in Intra-frame error concealment, reducing the degradation of spatial reconstructed image is one of the key techniques in error concealment. In this paper, a novel spatial error concealment algorithm based on fuzzy clustering for Intra-frame in H.264 is proposed. This proposed algorithm gives a new selection of eigenvectors and similarity measurement for fuzzy clustering. The experimental results show that the proposed algorithm can improve the edge details, achieve a better subjective quality of the recovered image and increase PSNR (Peak Signal-to-Noise Ratio) as well.