Development of Algorithm for Dynamic Hybrid Mean/Median Filter using Adaptive Window Selection Approach to Eliminate Salt & Pepper Noise
Keywords:Adaptive, Filter, Multimedia Data, Noisy Data, Salt Pepper Noise.
The research work aimed to develop a new Dynamic Hybrid Mean/Median Filter (DHMMF) algorithm to eliminate salt & pepper noise. The proposed DHMMF algorithm decides window size dynamically during runtime, also adaptively adjusts window size based on the non-noisy pixels present in a local window. Window size is limited to 9 X 9. This reduces blurring and computational complexity. Filter is designed with two stages, noise detection succeeded by filtering strategy. During the noise detection stage, if pixel intensity value is in-between 1 to 254, it is classified as a non-noisy pixel and left unchanged. Pixel having 0 or 255 intensity value is classified as a noisy pixel. During the filtering stage, a noisy pixel is replaced with mean, median or trimmed values within a local window depending on various algorithmic conditions. Performance of DHMMF algorithm is compared with various existing methods. Performance is tested for low, medium and high density noise. Simulation results demonstrate that image quality is retained by preserving fine details and edges which results in better visual quality. Quantitative and qualitative analysis is carried out using PSNR and SSIM respectively.
 Yung-Yue Chen, Ching-Ta Lu, Pei-Yu Chang, â€œEnhancement of Salt-and-Pepper Noise Corrupted Images Using Fuzzy Filter Designâ€, Frontier Computing, Vol. 422, (2017), pp. 691â€“701, available online: https://doi.org/10.1007/978-981-10-3187-8_65,
 Erkan, UÄŸur & GÃ¶krem, Levent & Enginoglu, Serdar, â€œDifferent applied median filter in salt and pepper noiseâ€, Computers & Electrical Engineering, Vol. 70, (2018), pp. 789-798, , available online: https://doi.org/10.1016/j.compeleceng.2018.01.019
 Mousavi, Seyed Mojtaba, Syed Ab Rahman, â€œA robust medical image watermarking against salt and pepper noise for brain MRI imagesâ€, Multimedia Tools and Applications, Vol.76, (2016), pp.10313-10342, available online: https://doi.org/10.1007/s11042-016-3622-9.
 MÃºjica-Vargas, Dante & Rubio, Jose de Jesus & Kinani, Jean Marie, â€œAn efficient nonlinear approach for removing fixed-value impulse noise from grayscale imagesâ€, Journal of Real-Time Image Processing, Vol. 14, (2017), pp. 617â€“633, , available online: https://doi.org/10.1007/s11554-017-0746-8.
 UÄŸur Erkan, Levent GÃ¶krem, â€œA new method based on pixel density in salt and pepper noise removalâ€, Turkish Journal of Electrical Engineering & Computer Sciences, Vol. 26, (2017), pp. 162-171, , available online: https://doi.org/10.3906/elk-1705-256.
 Esakkirajan, S & Thangaraj, Veerakumar & N. Subramanyam, Adabala & PremChand, C.H., â€œRemoval of High Density Salt and Pepper Noise Through Modified Decision Based Unsymmetric Trimmed Median Filterâ€, IEEE Signal Processing Letters, Vol. 18,(2011), pp.287-290, available online: https://doi.org/10.1109/LSP.2011.2122333.
 UÄŸur Erkan, Adem Kilicman, â€œTwo new methods for removing salt-and-pepper noise from digital imagesâ€, Science Asia, Vol. 42, (2016), pp. 28-32, , available online: https://doi.org/10.2306/scienceasia1513-1874.2016.42.028.
 Kenny Toh, Nor Ashidi Mat Isa, â€œNoise Adaptive Fuzzy Switching Median Filter for Salt-and-Pepper Noise Reductionâ€, IEEE Signal Processing Letters, Vol. 17, (2010), pp. 281-284, , available online: https://doi.org/10.1109/LSP.2009.2038769.
 Igor DjuroviÄ‡, â€œBM3D filter in salt-and-pepper noise removalâ€, Image Video Processing, (2016), Vol.13, available online: https://doi.org/10.1186/s13640-016-0113-x.
 Ke TuHongbo Li, Fuchun Sun, â€œA statistical learning based image denoising approachâ€, Frontiers of Computer Science, Vol.9, No.5, (2015), pp.713â€“719, available online: https://doi.org/10.1007/s11704-015-4224-9.
 Abdol Hamid Pilevar, Soudeh Saien, Mina Khandel, â€œA new filter to remove salt and pepper noise in color imagesâ€, Signal, Image and Video Processing, (2015), Vol.9, No.4, pp.779â€“786, available online: https://doi.org/10.1007/s11760-013-0514-6.
 B. K., Shreyamsha Kumar, â€œImage denoising based on non-local means filter and its method noise thresholdingâ€, Signal, Image and Video Processing, Vol.7, (2013), pp.1211-1227, available online: https://doi.org/10.1007/s11760-012-0389-y.