LITERATURE SURVEYIn3,Al-Araji introduced an auto selection method for reduction of impulse noisein wireless communication systems. The selection was made on the basis of estimation of rate of impulsenoise present in the system. The impulse noise is detected by comparing it witha fixed reference. Their system used subtraction-gating for low occurrence rateand directly sends the incoming signal to the output. If impulse is absent. Thesimulation results for QAM and FSK and real time implementation on an FPEG werepresented and the bit error rate was estimated.
In4, Chen put forward the idea of an adaptive pixel correlation filter(ACPF) toeliminate impulse noise, with an adaptive threshold designed based on the correlations between apixel and its neighbour. The filter was designed with a novel adaptive workingwindow and weighted function for impulse noise detection and preservingnoise-free pixels based on the fact that both horizontal and verticalcorrelations for a pixel are more significant than other orientations.In5, Chen proposed an adaptive working window to remove impulse noise. There wasa correlation between the neighbouring pixels.
The pixel correlation was usedto determine the intensity of impulse noise . They introduced a simple rule forimpulse noise reduction for various median-based filter with adaptive workingwindow. The switching median filter and multi-state median filter are designedwith this simple rule and the simulations showed that this method enhances the reliability.
In6, Divyajyothi proposed a new integrated fuzzy filter for the reduction ofadditive noise and impulse noise from digital color images. They used animpulse noise detector in the filter to detect the presence of impulse. Thedetector divides the set of pixels into affected points and clean points. Aselection filter module was employed to select the appropriate filter to matchthe type of noise. The output of integrated filter contains the enhanced imageafter noise removal. Their method combined the advantage of both additive noiseand impulse noise filters. Their results showed that the approach effectivelyremoved the integrated noise even at high noise levels.
In7,Rezvanian presented an efficient method to reduce impulse noise in twopasses. During first pass impulse noise detection using ANFIS, and at thesecond pass the impulse noise estimation, that corrupted noise pixel replacedwith new value based on neural networks. This method was experimented onpopular greyscale test images and compared with other methods using subjectiveand objective measures. Their results showed that the method works efficientlyin reducing impulse noise.
In8, Tao Chan proposed a tristate median filter for image denoising .Thispreserves image details while effectively suppressing impulse noise. Thisproposed filter outperforms other median filter by balancing the tradeoffs’between noise reduction and detail preservation. To achieve better results, acamera calibrations procedure may be placed before our systems. For differentvalues of noise density, optimum threshold range for yielding smallest MSEvalues and good visual quality can be obtained through similar simulationexperiments.In9, Zhou Wang and David zhang proposed a new median based filter i.e. progressiveswitching median filter to restore images corrupted by salt and pepper noise.
Thisalgorithm consists of two main concepts: switching scheme based filter and progressivemethod impulse detection and nose filtering procedures are progressively throughseveral iterations. The simulation results shows that the proposed algorithm iseffective than traditional median based filters. This method is particularlyuseful for the cases where the images are very highly corrupted ratios rangingfrom 5% to 70% . The MSE is also better than other median based methods, evenwhen noise ratios are high. It can remove most of the noisy pixels as well as preserve image details .
In10 Tao Chen and Hong Ren vu introduced a novel adaptive median-filter thatconsists the switching scheme based on the impulse detection mechanism i.e.adaptive impulse detection using centre weighted median filters.
The extensivesimulations shows that the proposed filter consistently works well in suppressingboth fixed and random valued impulses with different noise ratios. While, stillpossessing a computational structure.In11 ,Tao Chen and Hong Ren vu introduced a generalized framework of medianbased switching schemes. It is also called multi-state median filter.
By usingsimple threshold logic outputs, Senk and Trpovski proposed a robust estimatorof the variance, MAD, is modified and used to efficiently separate noisy pixelsfrom the image details and therefore has no sensitivity to image contents. The complexity of proposed algorithm is equivalent tothat of the median filter. The pixelwise MAD concept is straight forward, lowin complexity and achieves high filtering performance.
In12, Raymond Chan,Chen Hu and Mila Nikolova proposed a two stage iterativemethod for removing random valued impulse noise. In the first phase, anadaptive centre weighted median filter(ACWM) is used to identify corrupted pixels . In thelatter phase, these noise candidates are restored using a detail preservingregularization method which allows edges and noise free pixels to be preserved.These two phases are applied alternatively. This simulation results indicatethat the proposed method is significantly better than those using just nonlinear filters or regulization only.
This can be done very fast. The timing canbe enhanced by better implementations of minimization routines for solving therobust statistics with applications to early vision.In13, Hancheng Yu,Li Zhao and Haixian Wang proposed an efficent algorithmmethod uses a statistics of rank ordered relative differences to identifypixels which are likely to be corrupted .
It consists of two methods 1.RORDimpulse detector into many existing techniques, allowing them to detect andproperly handle impulse like pixels. 2. A sample weighted mean filter (SWMF) byusing the RORD detector and the reference image to suppress impulse noise, whilepreserving image details.
Although our algorithm is applied iteratively muchfaster, especially when the noise is high, the sample weighted filter offersgood filtering performance while its implementation complexity is lower thanothers.In14, Francisco Estard, David Fleet and Allan Jepson proposed a probabilisticalgorithm for image noise removal.ie., Stochastic Image Denoising .
Theseproposed algorithm for image denoising based on simulated random walks on imagespace and also very simple. Random walk produce stable estimates even for fewtrials and the overall behaviour of the random walks approximate that of morecomputationally expensive blur kernels. Stochastic denoising will become auseful tool for noise removal.In15, Jain-Feng Cai,Raymond H.Chan and Mila Nikolova(2010) proposed a fast two –phase image de blurring under impulse noise. This proposed algorithm consists ofa two- phase approach to restore image corrupted by blur and impulse noise,which is much simpler.
In the first phases, accurate detection of a location ofoutlier using a median-type filter and second phase edge -preservingrestoration that de-blur using only those data samples that are note noisecandidates. The PSNR of the restoration by our method is about 1 to 3 dB higherthan that by the variation method. Even for blurred images corrupted by 55%random valued noise, proposed method can give a very good restoration result.Comparing the two –phase methods. Image De-blurring under impulse noise withthe two- phase methods for de-blurring images corrupted by impulse Gaussiannoise.
The proposed method produces more computationally efficient and takesonly about 1/8 CPU.In16,P.Etrahanias and A.N.
Venetsanopoulos (1993) proposed a Vector DirectionalFilter-A New Class of Multichannel Image Processing Filters. These proposedfilters separate the processing of vector-valued signals into directionalprocessing and magnitude processing. This provide a link between single-channelimage processing, where only magnitude processing is essentially performed, andmultichannel image processing where both the direction and the magnitude of theimage vectors play an important role in the resulting image. VDF perform atleast as good and in most cases and also better chrominance estimate than VMFand this justifies their employment in colour image processing. VDF can achievevery good filtering results for various noise source models.In17, Yiqiu Dong, Raymond H, Chan and Shufang Xu (2007) proposed a DetectionStatistic for Random -Valued Impulse Noise.
This proposed technique fordetecting random –valued impulse noise based on image statistic. By thisstatistic, we can identify most of the noisy pixels in the corrupted images.Combining it with an edge preserving regularization, we obtain a powerful two-stage method for de-noising random valued impulse noise even for noise levelas high as 60%. These propose a new local image statistic ROLD, by which we canidentify more noisy pixels with less false-hits. We combine kit with the edge preservingregularization in the two- stage iterative method.In18 , Yiqiu Dong and Shufang Xu proposed a New Directional weighted medianfilter for removal of RVIN. Theyproposed a new impulse detector .
based on difference between current pixels andits neighbours aligned with four main directions. Then we combine it with theweighted median filter to get a new directional weighted median filter. Thisproposed method suppresses noise level and preserve image details too, includingthin lines. In addition it can beextended to restore color images corrupted by impulse noises.In19, Roman Garnett, timothy Huegerich and charles chui introduced a universalnoise removal algorithm, which consists of an impulse detector. This proposes alocal image statistics for finding corrupted pixels. The simulation result iscapable of reducing both Gaussian and impulse noise .This method is extended toremove any mix of Gaussian and impulse noise.
In20, vector directional filters (VDF) for multichannel image processing were introduced.VDF separates the processing of vector valued signal into directionalprocessing and magnitude processing. This provides a link between singlechannel image processing, where only magnitude processing is essentiallyperformed and multichannel image processing where both the direction and magnitudeof the image vectors play a vital role in the processed image.
The experimentalresult shows that in case of color images, VDF achieved good results for various noise source models. In21, shih-Chang Hsia proposed an efficient noise removal algorithm using anadaptive digital image processing approach. Simulations have demonstrated thatthe new adaptive algorithm could effectively reduce noise impulse even incorrupted images. In order to achieve real time implementations, a costeffective architecture is proposed using parallel structure and pipelinedprocessing. The proposed processor can achieve throughput rate of 45M pixels/susing only 4k gates and two line buffers. Unlike median filtering chips, thisprocessor provides better filtering quality and its circuit is much lesscomplex.
In22, Deepa Et Al proposed an efficient low cost VLSI architecture for the edgepreserving impulse noise removal technique. The architecture comprises of twoline buffers, register banks, impulse noise detector, edge oriented noisefilter and impulse arbiter. The proposed algorithm involves only fixed sizewindow instead of variable window size. The storage space required is a twoline buffer. Both, reduces storage requirement as well as computationcomplexity. The implemented edge preserving algorithm results in better visualquality and pipelined architecture results in better visual quality fordenoised image.CONCLUSIONThispaper surveys different common median filtering techniques.
Each technique hasits own advantages, and disadvantages. From literature, we found that most ofthe recent median filtering based methods employ two or more than two of theseframe work in order to obtain an improved impulse noise reduction and enhancingpicture quality.