Engineering, Technology & Applied Science Research Vol. 7, No. 6, 2017, 2288-2292 2288 www.etasr.com Banerjee et al.: Random Valued Impulse Noise Removal Using Region Based Detection Approach Random Valued Impulse Noise Removal Using Region Based Detection Approach Shubhendu Banerjee Narula Institute of Technology Kolkata West Bengal, India shankushubhendu@gmail.com Aritra Bandyopadhyay Supreme Knowledge Foundation Group of Institutions Mankundu West Bengal, India aritra.d90@gmail.com Avik Mukherjee Tata Consultancy Services Ltd Digital Interactive IOU in TCS Pune, India mukherjee.avik852@gmail.com Atanu Das Netaji Subhash Engineering College Garia, Kolkata West Bengal, India atanudas75@yahoo.co.in Rajib Bag Supreme Knowledge Foundation Group of Institutions Mankundu West Bengal, India rajib.bag@gmail.com Abstract—Removal of random valued noisy pixel is extremely challenging when the noise density is above 50%. The existing filters are generally not capable of eliminating such noise when density is above 70%. In this paper a region wise density based detection algorithm for random valued impulse noise has been proposed. On the basis of the intensity values, the pixels of a particular window are sorted and then stored into four regions. The higher density based region is considered for stepwise detection of noisy pixels. As a result of this detection scheme a maximum of 75% of noisy pixels can be detected. For this purpose this paper proposes a unique noise removal algorithm. It was experimentally proved that the proposed algorithm not only performs exceptionally when it comes to visual qualitative judgment of standard images but also this filter combination outsmarts the existing algorithm in terms of MSE, PSNR and SSIM comparison even up to 70% noise density level. Keywords-random valued inpulse noise; image filtering; region based; detection I. INTRODUCTION Random valued impulse noise removal is a stringent task. This type of noise is introduced in the digital images through transmission and acquisition [1]. A significant characteristic of this type of noise is that only fractions of the pixels are degraded [2]. A variety of filters [4-22] are being proposed in past years to remove random valued impulse noise. The main motive of using these filters is to detect and reduce the noise as well as preserving the image information. Median filters were rapidly used to remove this type of noise. Author in [3] introduced a median filter which was able to contain noise up to some level. It was a good filter, which preserved edges greatly. But this filter was not fulty effective as noiseless pixels also got changed during restoration. Successively some other median filters were proposed [4-6]. These filters provided noise restoration to a medium level but they were not so effective for the same reason, so the image got blurred and edges were modified. Afterwards a variety of condition based filters were introduced. But those filters failed to restore the noise effectively. Author in [7] introduced a two phase swap filter which was able to preserve details at a low noise density. A tri state median filter that used a standard median and center weighted median filter to detect the noise but was not able to preserve minute details at medium noise densities was proposed in [8]. In [9], author proposed another adaptive center weighted median filter which sighted a slight improvement in preserving details but that also in lower noise densities. Author in [10] introduced a filter which was able to preserve minute details at medium noise density. Though the detection cabability was not that impressive. To enhance its capability a generic programming filter was proposed in [11]. This filter had a two stage cascading detector which enchanced the detection rate. A new dimension in the filtering is introduced in [12]. This was a PDE based technique which used anisotropic diffusion to filter impulsive noise. The method performed better than the previous disscussed schemes. In 2012 ROR [13] (robust outlyingness ratio) was proposed. This filter first calucated a robust outlyingness ratio to find out the impulsive pixels then by the ROR value the pixels are divide into four clusters. It was followed by a two stage detection process. Use of fuzzy filters had also shown great result. In 2013 a robust direction based detector [14] was proposed which used standard deviation based detection procedure to filter the noise. Afterwards, TDWM [15] was proposed. This filter used directional threshold based approach to remove noise. This filter was good at medium noise density level. Recently, a new detection method was introduced in [16]. This method used standard deviation and threshold concept together to approximate neighborhood pixels. This filter was successful at 60% noise density but failed once the noise was higher than that. So, removal of impulse at higher noise density was still a