A Review of Histogram Equalization Based Image Contrast Enhancement Methods

Contrast enhancement is considered as one of the critical trademark in the field of image processing and histogram equalization is a straightforward and understood strategy for image contrast upgrade, this technique utilizes the histogram of image in its preparing. In spite of the fact that the ordinary histogram equalization technique is broadly acknowledged however this strategy experiences the drawback of "meanshift" issue, i.e. mean brightness of processed image will always be the middle gray level regardless the mean brightness of the input image. So it is not considered as the best technique for balance improvement with brightness protection. A few other histogram equalization based techniques have been introduced to overcome the issue of mean shift problem. In this paper we will survey generally utilized histogram equalization strategies utilized for differentiation upgrade and brightness conservation.


1-INTRODUCTION
In the field of advanced image preparing contrast enhancement is for the most part utilized as a part of computer vision and it is likewise utilized for development of pictorial data for human visual perception.It is generally used for medical image processing, speech recognition, texture synthesis and various other applications.A few techniques have been presented in the field of image contrast enhancement [2]- [10].Let us review some of the important methods.Low contrast advanced image investigation is a testing issue in computerized image preparing.Low differentiation computerized images decrease the capacity of observer to analyze the image.Histogram equalization (HE) is the one of the prominent strategy for image contrast improvement [1].The histogram of the discrete dark level image speaks to the recurrence of event of every dim level in the image [5].This technique is predominantly utilized on account of its straightforwardness of usage and computationally less cost than different strategies.In spite of the fact that the HE technique is broadly acknowledged however many time histogram equalization image experiences the "mean-shift" issue [5], i.e. it shifts the mean intensity value to the middle gray level of the intensity range.So this procedure is not helpful in circumstances, where mean brightness conservation is required.To defeat this issue various varieties of HE technique have been proposed.In this paper we will give a precise survey on generally utilized HE based techniques.The organization of this work is as follows.After providing a brief introduction in section 1, in section 2 we will cover the histogram equalization method in detail.Section 3 covers in details all the widely used HE based algorithms.Finally section 4 concludes the entire context 2-THE HISTOGRAM EQUALIZATION METHOD Histogram equalization is an outstanding technique for image contrast improvement that utilizations histogram of image in its preparing.HE is a spatial space method in which alteration of pixels power qualities is done specifically which prompts upgrade of image.

D-Equal Area Dualistic Sub-Image Histogram Equalization
The BBHE technique solved the mean-shift issue up to some degree; however this strategy was not ready to protect fine subtle elements display in the information image.To take care of this issue Equal Area Dualistic Sub-Image Histogram Equalization (DSIHE) was proposed [6].This strategy takes after an indistinguishable approach from that of the BBHE with the exception of DSIHE sections the image histogram in light of the median of the information image.

E-Recursive Mean Separate Histogram Equalization
This technique is an expansion of BBHE; the RMSHE [7] gives brighter preservation than BBHE.In BBHE the mean based division is preformed just once yet in RMSHE this division is performed more than once.This recursive nature of the RMSHE strategy infers adaptable conservation which is exceptionally valuable in purchaser electronic items.Not able to preserve image fine details.
5 DSIHE This method is able to deal with mean-shift problem.
Not good as it is not able to maintain more mean brightness in the processed image.6 RMSHE An extension of BBHE, this method perform histogram segmentation recursively.
More frequent grey levels are overenhanced and less frequent grey levels are less enhanced.7 RSWHE An extension of RMSHE, the method was developed to overcome disadvantages of RMSHE method.
No disadvantage.
(a) shows input image having low contrast, (b) shows histogram of input image, (c) shows histogram equalized image and (d) shows histogram of processed image.3-HISTOGRAM EQUALIZATION BASED METHODS A-Adaptive Histogram Equalization Method Adaptive histogram equalization (AHE) [2] is a system of image differentiation upgrade utilized as a part of computerized image handling.It is not the same as ordinary histogram adjustment procedure in the way that the AHE figures different histogram balances freely, each of which has a place with various areas of images.The main advantage of AHE is that it can upgrade the nearby complexity of image and consequently, preserve more brightness.Conventional histogram adjustment utilizes a similar change to change all pixels.HE demonstrates better in situations where pixels qualities are conveyed consistently in the image.Be that as it may, flops in situations where image comprises of various brighter and darker locals as it is not ready to improve the complexity of image.Therefore not able to protect splendor of brightness of image.AHE conquers the downside of HE by improving nearby differentiation of image.AHE improve this change work by changing every pixel values.In light of the histogram of square, which is encompassed by pixel esteem every pixel esteem is changed as appeared in figure.The pixel estimation of neighborhood is corresponding for every change work and aggregate dispersion work.The customary histogram adjustment is like the change work got from every pixel esteem.

Fig. 4
Fig. 4 shows the concept of BBHE, here the input image histogram is divided into two sub-histograms (source [7]).D-Equal Area Dualistic Sub-Image Histogram EqualizationThe BBHE technique solved the mean-shift issue up to some degree; however this strategy was not ready to protect fine subtle elements display in the information image.To take care of this issue Equal Area Dualistic Sub-Image Histogram Equalization (DSIHE) was proposed[6].This strategy takes after an indistinguishable approach from that of the BBHE with the exception of DSIHE sections the image histogram in light of the median of the information image.

Fig. 5
Fig.5shows the concept of RMSHE, here the input image histogram is divided into four sub-histograms(ref [7]).F-Recursively Separated and Weighted Histogram EqualizationThe basic thought of RSWHE [8] is to fragment an information histogram into at least two subhistograms recursively, to alter the sub-histograms by methods for a weighting procedure in view of a normalized power law work, and to perform histogram equalization out on the weighted subhistograms autonomously.

Fig
Fig. 6intermediate results of RSWHE: (a) input image, (b) input histogram, (c) segmented histogram, (d) weighted and normalized PDF, (e) output image, and (f) output histogram (ref [8]).Next we show a tabular comparison of each method discussed in this work.4-CONCLUSION This paper gives the assessment of various histogram equalization strategies.The investigation of different histogram equalization strategies demonstrates that brightness is not protected productively by histogram equalization strategy.Different strategies like BBHE and DSIHE are acquainted which tries with expel the drawback of histogram equalization up to some degree.RMSHE conquers the downside of BBHE and DSIHE.