Global Image Completion with Joint Sparse Patch Selection and Optimal Seam Synthesis

Shiming Ge,Kaixuan Xie,Shuying Li,Rui Yang
DOI: https://doi.org/10.1016/j.sigpro.2015.09.031
IF: 4.729
2015-01-01
Signal Processing
Abstract:This paper presented a global exemplar-based image completion method for filling large missing or unwanted regions in an image. Based on three proposed completion rules, image completion problem is formulated as a global discrete optimization problem with a well-defined energy function. The energy function can evaluate image consistency globally and is minimized with an expectation-maximization (EM) like algorithm, which considers patch matching and patch synthesis in a unified way. In the algorithm, M step and E step are achieved by sparse patch subspace searching and optimal seam synthesis respectively. The patch subspace is learned with the statistics of geometric transformation relationships of similar patches. Moreover, E step combines image patch synthesis and coherent correction simultaneously. We analyzed the proposed global energy function and optimization method in theory. Simulation comparisons with other state-of-the-art methods show the superiority of our proposed method in ensuring global coherent and avoiding image blurring.
What problem does this paper attempt to address?