Image salt and pepper noise self-adaptive suppression algorithm based on similarity function

Through summarizing the existing detail-preserving salt and pepper noise suppression methods, a new similarity function self-adaptive weighted algorithm is proposed. It analyzes and overcomes the shortcoming of the local extremum misjudgment of the Maximum-minimum noise detector by using a similarity function self-adaptive weighted algorithm. The local window noise probability is estimated by applying extremum trimming operation to select a suitable filtering window (recursive window or non-recursive window). Thus the proposed algorithm realizes self-adaptive suppression of different salt and pepper noise probabilities using a 3 × 3 filtering window. Experiments show that the results of salt and pepper noise suppression, detail-preserving and computation efficiency are satisfactory.