KNN classification based nonlinear noise delineation in color image using optimized BM3D filter

Abstract

Image enhancement using optimized methods along with optimized filters is need of current era. A novel technique is proposed in this paper, which have KNN based pixels classifications strategy. This classification is used to identify noise level and only noisy pixels are processed using optimized BM3D using particle swarm optimization. The outside group pixels are brought back into group thereby removing the noise. The process is further followed by resolution enhancement and Retinex for dynamic presentation purpose. The experimentation results are also included in the paper which shows optimum performance.