Evolutionary best neighborhood matching algorithm for error concealment of block-based image coding

Best - Neighborhood - Matching (BNM) is a novel error concealment algorithm by employing long-range blockwise similarity in the image, which has proved to be a significant improvement over traditional approaches [10]. However, one of the major drawbacks of the BNM algorithm is its exhaustive searching procedure, which leads to a very high computation burden. This paper proposes an improved version of the BNM method by combining Evolutionary Strategy (ES) into the searching procedure. Compared with the basic BNM algorithm, our new scheme, named ES-BNM algorithm, can obtain similar error concealment results in much shorter time. ES-BNM can be also viewed as a new successful application that demonstrates the powerful searching ability of ES.