* Indian Statistical Institute, Kolkata ** University of Liverpool, UK *** Indian Institute of Engineering Science and Technology, Shibpur
This work presents a novel multiple pyramid-based image inpainting method using local patch distribution and steering kernel feature based sparse representation. This helps to maintain texture consistency and structure coherence in the inpainted region. The algorithm has two steps. In the first step, each patch in the target region (region to be inpainted) is approximated by statistically dominant local candidate patches to preserve local consistency. Then each approximated patch is refined by sparse representation of candidate patches based on local steering kernel (LSK) feature to retain texture quality. In the second step, we apply a multiple pyramid-based approach to generate several inpainted versions of the input image, one for each of the pyramids. Finally, we combine the inpainted images by gradient-based weighted average to produce the final inpainted image. This approach helps to maintain structure coherence and to remove artifacts which may appear in the inpainted images due to different initial scales of the individual pyramids.
Target Image Xu's patch-sparsity Li's MDF Ours