International Journal of Computer Trends and Technology (IJCTT) - volume4Issue4 April 2013

Combine Approach of Enhancing Images Using Histogram Equalization and Laplacian Pyramid
Rucha A. Deshmukh#1, Dr. Siddharth A. Ladhake*2 M.E. 4th Sem Computer Science and Engineering SIPNA C.O.E.T. Amravati. Near Nemani Godown,Amravati-444605,Maharashtra,India Sant Gadge Baba Amravati University, India
#

Abstract Image enhancement aims at improving some characteristics of image to make it visually better one. One of the effective contrast enhancement techniques is histogram equalization (HE). But image enhancement methods based on histogram equalization often fail to improve local information and sometimes leading to over enhancement. In order to overcome these drawbacks, we propose a method based on modified Laplacian pyramid framework that will decompose image so that its local as well as global information is improved. HE is proposed to provide a well balanced mapping function in order to reduce quantum jump regarding global contrast. Keywords contrast enhancement, detail Laplacian Pyramid, histogram equalization. enhancement,

techniques is histogram equalization (HE). HE is an effective technique to transform a narrow histogram by spreading the gray-level clusters in the histogram and it is adaptive since it is based on histogram of a given image. The classic HE provides the best visual performance in certain conditions compared to other state-of-the-art techniques, but has the fatal flaw of over enhancement when a quantum jump occurs in histogram of an image. To overcome these shortcomings, we propose an image enhancement method based on a modified Laplacian pyramid framework. II. RELATED WORK Several contrast enhancement techniques have been used to improve histogram of an image. Histogram equalization (HE) [1], the representative method for contrast enhancement, has been developed to satisfy humans with its resultant images. HE is an effective technique to transform a narrow histogram by spreading the gray-level clusters in the histogram and it is adaptive since it is based on histogram of an image. The classic HE provides the best visual performance in certain conditions compared to other state-of-the-art techniques but has disadvantage of over enhancement. A technique named Contrast-limited adaptive histogram equalization [2] inspired some post-HE methods separates an image into tiles, equalizes them and interpolates their boundaries. Method is successful in improving contrast enhancement but adds up the result into unnatural look. Inspired from bright preserving bi-histogram equalization (BBHE) [2], some HE methods started to divide a histogram into two or more parts. BBHE individually equalizes two sub histograms based on the mean value of the image[3-4]. Similar to BBHE, dualistic sub-image histogram equalization [5] separates an image into two sub-images based on the median brightness value of an image and equalizes each subimage. Partially overlapped sub-block histogram equalization [6] improves the computational complexity and speed of the blocked-overlapped HE. The clipping technique of CLAHE [7] was modified by the gain controllable clipped histogram equalization (GCCHE). After a clipping step in the histogram, a portion of the residual is redistributed uniformly to all the bins and the rest is added locally to low and high intensity bins. However, its local process contributes less to the overall performance.

I. INTRODUCTION Image enhancement is an important step in almost every image processing application. The objective of image enhancement is to increase the visual perception of the image so that they are more suitable for human viewers or machine vision applications. It is well known in the image processing society that there is no unifying or general theory for image enhancement algorithms. Thus an enhancement algorithm that is suitable for some application may not work in other applications. In image processing technology, image enhancement means improving image quality through a broad range of techniques such as contrast enhancement, color enhancement, dynamic range expansion, edge emphasis, and so on. The critical technique affecting the human visual system (HVS) is to improve the contrast of an image, because HVS is more sensitive to luminance than to other components such as color information. Contrast enhancement plays a crucial role in image processing applications such as digital photography, medical image analysis, remote sensing. Contrast enhancement techniques modify the image through some pixel mapping such that the histogram of the processed image is more spread than that of the original image. These techniques either enhance the contrast globally or locally. If a single mapping derived from the image is used then it is a global method; if the neighbourhood of each pixel is used to obtain a local mapping function then it is a local method. Using a single global mapping cannot (specifically) enhance the local contrast. One of the most popular global contrast enhancement

ISSN: 2231-2803

http://www.ijcttjournal.org

Page 669

International Journal of Computer Trends and Technology (IJCTT) - volume4Issue4 April 2013
Other contrast enhancement technique allows decomposition of image into spatial frequency bands in which individual bands are manipulated. One decomposition scheme first proposed for lossless image compression was Laplacian pyramid [8]. Among them CIELP showed interesting performance but showed halo effect. Instead of Laplacian pyramid a wavelet-based algorithm was used which was based on contrast sensitivity function (CSF) considering both color colorfulness index (CCI) and color natural index (CNI). Some retinex-theory based algorithm like MSRCR [12] and McCann-99 [14] were also used but were too much expensive and required well calibrated parameters respectively. We are proposing an enhancement technique using merits of two methods, namely, histogram equalization and Laplacian pyramid. The target is to deliver an image with high improved quality preserving its natural look. III. PROPOSED FRAMEWORK The main focus of HE-based algorithm is to improve global contrast of an image which means very little attention is payed on improving its local contrast and edges. This is a limitation of HE methods. So we construct a framework to decompose images so as to separate image into several frequency component in spatial domain. The framework will work as follows: For an input RGB image transformation is performed to obtain I0 of the luminance. The I0 is decomposed into band-pass images. The RGB input and I0 are later reused in final step of color restoration. The full-scale image pyramid [10] can prevent halo effect generated during construction process. The decomposition involves decomposing images into In Gaussian low-pass filtered images and Dn Laplacian images by subtracting a high-level In+1 from a low-level In, *g is the convolution Gaussian low-pass filter image and n is the index of decomposition layer (n>=1) . This thread is common to previous idea by Fatal et al. [15] who stated a reconstruction equation as I0 = I N + Where N is the decomposition layer. (1)

Fig 1: Block diagram of proposed framework

Fig 2: Full-scale image pyramid decomposition

A. Contrast Enhancement There are several reasons for an image to have a poor contrast like poor quality of used imaging device, lack of expertise of the operator, and the adverse external condition at the time of acquisition. These effects result in under-

ISSN: 2231-2803

http://www.ijcttjournal.org

Page 670

International Journal of Computer Trends and Technology (IJCTT) - volume4Issue4 April 2013
utilization of the offered dynamic range. As a result, such images and videos may not reveal all the details in the captured scene, and may have a washed-out and unnatural look. Contrast enhancement targets to eliminate these problems, thereby to obtain a more visually-pleasing or informative image or both. Typical viewers describe the enhanced images as if a curtain of fog has been removed from the picture. 1) Generating the histogram: The histogram with luminance levels in the range K [0, L-1] is a discrete function as h (lk) = nk (2) Where lk is the kth luminance level in K and nk represents the number of pixels having luminance level lk.
Fig 6: Boosted minor areas

Second, the ridges between valleys are searched and boosted. Ridge boundary is defined as the bins between the first point of the positive slope and the last point of the negative slope.

Fig 7: Boosting local areas in smoothed histogram. Fig 3: Generating histogram

2) Smoothing the histogram: In the histogram, a ridge shape with some consecutive luminance levels can be regarded as the feature area of an image. To globally distinguish between ridges and valleys and remove their ripples, we smooth the histogram [11-13] like as follows: hg ( lk) = h (lk) * g (lk) (3)

In the fig 7, we can see that the small ridge of minor areas are boosted in the ridge boundary after the local maximum value Lmax and the local minimum value Lmin (i.e the higher of the two values Lmin1 and Lmin2) are found. 4) Slantwise clipped bins & redistributing the residual: The clipping technique is used as it effectively suppresses the quantum jump and the proposed algorithm gathers the residual from local and global clipping.

Fig 4: Smoothing histogram

Fig 8: Clipping bins

Fig 9: Redistribute the residual

Where g (x) = e-x^2 where g(x) is a Gaussian function, x is the corresponding location to a bin of the histogram, and coefficients of the Gaussian filter are normalized. 3) Boosting small ridges: Boosting minor areas, i.e. small ridges, in the histogram is one of the key strategies of the proposed contrast enhancement to suppress the quantum jump. First, the peak value in the smoothed histogram hg ( lk) is found as p (K) =

5) Generate luminance mapping function: This is the last step in which luminance function is obtained as shown

( )}

(4)

ISSN: 2231-2803

http://www.ijcttjournal.org

Page 671

International Journal of Computer Trends and Technology (IJCTT) - volume4Issue4 April 2013
Fig 10: Luminance mapping function

B. Detail Enhancement HE-based methods have inherent limitations on improving the local information of an image. To solve this problem, we employ detail enhancement through the proposed framework which initially separates an image into band-pass images. With this framework, we can deal with the original details of high-pass images. The proposed detail enhancement has the effect of distinctively emphasizing reliable details. An example of the detail enhancement scheme is shown below:

framework aims to provide an output image with well natural look without over-enhancement or severe failure. ACKNOWLEDGMENT I express my sincere gratitude to Resp. Dr. A. D. Gawande Head of the Department, Computer Science & Engineering & Resp. Dr. S. A. Ladhake for providing their valuable guidance and necessary facilities needed for the successful completion of this seminar throughout. Lastly, but not least, I thank all my friends and wellwishers who were a constant source of inspiration. REFERENCES
[1] [2] R. C. Gonzalez, and R. E. Woods, Digital Image Processing, 3rd ed., Upper saddle River, N.J., Prentice Hall, 2008. K. Yeong-Taeg, Contrast enhancement using brightness preserving bihistogram equalization, IEEE Trans. Consum. Electron., vol. 43, no. 1, pp. 1-8, 1997. S. D. Chen, and A. R. Ramli, Contrast enhancement using recursive mean separate histogram equalization for scalable brightness preservation, IEEE Trans. Consum. Electron., vol. 49, no. 4, pp. 1301-1309, Nov, 2003. C. Soong-Der, and A. R. Ramli, Minimum mean brightness error bihistogram equalization in contrast enhancement, IEEE Trans. Consum. Electron., vol. 49, no. 4, pp. 1310-1319, 2003. W. Yu, C. Qian, and Z. Baeomin, Image enhancement based on equal area dualistic sub-image histogram equalization method, IEEE Trans. Consum. Electron., vol. 45, no. 1, pp. 68-75, 1999. J. Y. Kim, L. S. Kim, and S. H. Hwang, An advanced contrast enhancement using partially overlapped sub-block histogram equalization, IEEE Transactions on Circuits and Systems for Video Technology, vol. 11, no. 4, pp. 475-484, Apr, 2001. K. Tae Keun, P. Joon Ki, and K. Bong Soon, Contrast enhancement system using spatially adaptive histogram equalization with temporal filtering, IEEE Trans. Consum. Electron., vol. 44, no. 1, pp. 82-87, 1998. P. J. Burt, and E. H. Adelson, The Laplacian Pyramid as a Compact Image Code, IEEE Trans. Commun., vol. 31, no. 4, pp. 532-540, 1983. S. M. Pizer, E. P. Amburn, J. D. Austin, R. Cromartie, A. Geselowitz et al., Adaptive Histogram Equalization and Its Variations, Computer Vision Graphics and Image Processing, vol. 39, no. 3, pp. 355-368, Sep, 1987. Y. M. Baek, H. J. Kim, J. A. Lee, S. G. Oh, and W. Y. Kim, Color image enhancement using the Laplacian pyramid, Advances in Multimedia Information Processing - PCM 2006, Proceedings, vol. 4261, pp. 760-769, 2006. T. Arici, S. Dikbas, and Y. Altunbasak, A Histogram Modification Framework and Its Application for Image Contrast Enhancement, IEEE Trans. Image Process., vol. 18, no. 9, pp. 1921-1935, 2009. D. J. Jobson, Z. U. Rahman, and G. A. Woodell, A multiscale retinex for bridging the gap between color images and the human observation of scenes, IEEE Trans. Image Process., vol. 6, no. 7, pp. 965-976, Jul, 1997. S. S. Agaian, B. Silver, and K. A. Panetta, Transform Coefficient Histogram-Based Image Enhancement Algorithms Using Contrast Entropy, IEEE Trans. Image Process., vol. 16, no. 3, pp. 741-758, 2007. J. McCann, Lessons learned from Mondrians applied to real images and color gamuts, in Seventh Color Imaging Conference: Color Science,Systems and Applications, pp. 1-8, 1999. R. Fattal, M. Agrawala, and S. Rusinkiewicz, Multiscale shape and detail enhancement from multi-light image collections, ACM Trans. Graph., vol. 26, no. 3, 2007.

[3]

[4]

[5] Fig 3: Concept of detail enhancement

[6]

C. Color restoration Using (1), the enhanced luminance image is obtained as I = IN + D (5) Where the right hand components are obtained by combining the enhanced images. The concept here used is that we combine all the contrast enhanced and decomposed images together along with luminance and get the enhanced luminance image. IV. EXPERIMENTAL RESULTS The most important part of our technique is about the number of layers of image to be decomposed. In the proposed framework, a new HE algorithm was proposed to enhance contrast information. Practically, the proposed HE algorithm alone can improve image quality. Robustness is an important factor for practical use. The allowable degree of overenhancement and the well performed degree of quality heavily depend on the viewers subject. Fewer enhancements is preferable in some difficult conditions like a flat background. Another difficult task occurs when a narrow band of the histogram dominates an image. Especially, when we observe the results processed under dominant dark regions. V. CONCLUSIONS The image enhancement technique used by joint effort of histogram equalization and Laplacian pyramid aims to overcome the quantum jump that conventional HE-based algorithms are likely to suffer from and for supplementing the improved local details that those methods cannot provide due to their inherent limitations. The proposed combination

[7]

[8] [9]

[10]

[11]

[12]

[13]

[14]

[15]

ISSN: 2231-2803

http://www.ijcttjournal.org

Page 672