An extremely fast adaptive high-performance filter to remove salt and pepper noise using overlapping medians in images

Abstract

A new fast adaptive high-performance filter (FAHPF) has been proposed to remove salt-and-pepper noise in images. ‘Maximize the speed without compromising denoising performance’ is the fundamental intention to build up the FAHPF algorithm. Among diverse phases of filtering employed in the FAHPF, overlapping medians (OM) is our newly proposed frame-based filtering concept which is the basis for speed of FAHPF and running averages embedded with OM is an idea behind excellent denoising performance of FAHPF at the same pace. Simulation experiments have been conducted and denoising results of FAHPF has been investigated against very recently developed filtering methods. It is proved that the FAHPF excellently outperforms many of state-of-the-art filters considered for comparison, in terms of peak signal to noise ratio, structural similarity index, and visual representation and requires the extremely shortest execution time among all, which could make it as a real time filter.

Keywords:

• Denoising
• Overlapping medians
• Real time filter
• Running averages
• Salt-and-pepper noise