A novel similarity measure for fuzzy peer group based removal of mixed noise

In general, raw image may suffer from various uncertain distortions arising due to many factors like low dynamic range of imaging device, low illumination, bad weather conditions. In this paper, a new idea of fuzzy peer group-based image denoising algorithm has been presented with a novel similarity measure function. Unlikely the Gaussian membership function based conventional similarity measure; here proposed method hired the local spatial information induced similarity measure together with Gaussian membership function by exploiting the significance of both. Using the proposed similarity measure, fuzzy peer group of similar pixels contrived and color image having contamination of mixed noise i.e., impulse noise and Gaussian noise is decontaminated with refashioned form of bilateral filter. The algorithm proposed has been compared with some state-of-the-art methods in studies and other techniques. The experimental performance shows that proposed approach is having leading filtering outcome with edge preserving and color information from existing methods.