Contrast Enhancement Techniques For Image

Abstract

The Enhancement of images is done to get clear information about the details of the digital image. Contrast Enhancement is an important technique in Image Enhancement for improving the contrast of the image. Various contrast enhancement techniques like the histogram equalization, genetic algorithms and fuzzy methods are available for the betterment of the contrast of the images. Low contrast of the images tends to reduce the ability of the observer in analyzing the image. High contrast of the images may result in over enhancement and feature loss problems which may lead to unnatural look in the processed images. This paper presents an analysis of the state of the art techniques that are based on the high contrast and low contrast of the images.

Keywords: Image Enhancement; Contrast Enhancement; Low Contrast; High Contrast.

Introduction

The objective of Image Enhancement is enhancing the input image, thereby making the observer understand the image. Image Enhancement techniques work both in spatial and frequency domains. These techniques improve the brightness, contrast and also edge sharpening and smoothing. Most of the methods are implemented in spatial domain. In order to represent the image in a compressed format, the frequency domain is used [1]. Digital images are used in different applications and it is necessary to use the most expertise method for the image enhancement on the degraded images. The cause of the degradation of the image could be the lack of contrast that may result in low or high lighting scenario [2].

Contrast Enhancement

Contrast improvement is one of the important approaches to both the video and image, which can efficiently upgrade the quality of human visual perception and recognition. Various images like medical images, aerial images, satellite images and even real life photographs may suffer from noise and poor contrast due to the inappropriate lighting during image acquisition and/or wrong setting of aperture size and shutter speed of a camera, so it is necessary to enhance the contrast and remove the noise to increase image quality. Contrast Enhancement aims at enhancing the global or local contrast of an image [3].

Contrast is the ratio between the white and the black, or the light and the dark parts of a scene. High-contrast images display a full range of tones, from bright highlights to dark shadows. Low-contrast images have a much smaller, shallower range of tones.

Contrast of an image can be revealed by its Histogram. Histogram of a monochrome image with L possible gray levels f = 0 1L-1possible gray levels, f=0, 1,.., L-1 P(l) = n/ n,n is the number of pixels with gray level l

• nl is the number of pixels with gray level l
• n is the total number of pixels in the image.

Various algorithms are proposed for the study of these low contrast and high contrast images and changing them to a user friendly mode so that the observer gains the exact information out of the image.

Literature Survey

To overcome the disadvantages of low contrast, an infrared based image enhancement algorithm is presented. A bilateral filter is implemented on the original infrared image which then yields the base layer and detail layer. The histogram equalization technique is applied for the global contrast enhancement. Gamma correction was then used for the detail layer enhancement. The proposed algorithm could solve the bleaching problems and showed better enhancement in contrast [4].

Various real time images, satellite images may face the contrast problems due to lack of proper lightening. To overcome these the author proposed a fuzzy rule based contrast enhancement along with intensification(INT) operator and fuzzy expected value FEV. Index of fuzziness along with processing time is used in the analysis of these techniques [5]. In order to differentiate the darker and brighter portions of the image, a fuzzy image enhancement is proposed for low contrast and non uniform illumination. Power law transformation is used to enhance the image. Proposed method performs better on non uniform, low contrast images [6].

For the enhancement of non uniform illumination and low contrast images, the author proposed a method that uses an enhancer for the dark and bright regions separately with the help of a mean filter. Results are compared with different histogram based methods and the proposed method outperforms these methods [7]. This paper proposed a fuzzy method for the transform domain. Sequency based mapped Real Transform (SMRT) domain is analyzed. Results are compared with other methods like Fuzzy INT operator (FINT).Enhancement metrics are also compared. The method shows good performance in enhancing the low contrast regions [1].

The author discussed the problems that arise in computer graphics when the graphic images have to be similar to actual ones. In order to overcome this problem, a local contrast enhancement algorithm is presented that uses histogram equalization and stretching. The tone mapping is also used to solve the high contrast images. Good results are shown [8].

The visual appearance of the low contrast images are improved by altering the global and local contrast. This paper suggests a modified histogram based method along with the saturation and local contrast adjustment. Better results are displayed compared to the other existing methods [9]. Small sized objects that have low contrast are considered and the contrast enhancement is applied on these objects based on the histograms. Contrast kernels are used in the estimation of the parameters in the proposed method [10]

The author proposed a generalized algorithm for contrast enhancement that could enhance the color image as well as preserve the hue. It also used the scaling and shifting method with grey scale transformation technique [11]. A non linear image enhancement technique is presented for high contrast images where in the digital images that are captured in different non uniform lighting situation are considered. The contrast enhancement, color restoration, intensity enhancement issues were considered and results showed good clarity of the image [12]

Conclusion

Contrast enhancement always plays a crucial part in the Image Enhancement. The contrast of a grayscale image indicates how easily objects in the image can be distinguished. The High contrast image has many distinct intensity values and the Low contrast image uses few intensity values. Many techniques that considered the low contrast and high contrast of images were surveyed. These methods convert the low or high contrast image into quality image with balanced contrast for the viewer’s readability.