A comparison between different colour image contrast enhancement algorithms

In this paper, we study different colour image contrast enhancement algorithms to bring about the comparison between their performance which shows the improvement in the visual quality of the colour images captured under poor illumination and/or other color illumination conditions. The motivation behind this work is to improve the features in the image that are not clearly visible due to the influence of the illumination conditions and to achieve colour rendition and dynamic range compression for the low contrast images. Hence main objective is to remove the illumination effect present in the image. The methods discussed here acts like the pre processing techniques for images which are used for further image processing applications. The Retinex and Homomorphic methods acts like filtering techniques which are used to remove the low frequency illumination components of the images while retaining the high frequency reflectance components. The main goal of the Retinex theory is to compensate for local brightness in the images. It tries to flatten (reduce) the gap between the local bright and dark region, and provides with a better perception effect, especially on the fine features of region of interest (ROI). In the Retinex method, the surround functions acts like high pass filter, and in the present study we have used Gaussian, Laplacian and Gamma distribution functions as surround functions. We report the performance of each these surround functions. In the Homomorphic filtering method a high pass filter is used as a surround function. Here the illumination and the reflectance values of the pixels are separated based on illumination-reflectance model. The main idea of Homomorphic filtering enhancement is to remove the illumination in image. The Homomorphic filtering destroys some part of the image which does not require enhancement. That part is recovered here by using a threshold after applying the Homomorphic technique to the image. We also compare the retinex and H omomorphic method in the context of image contrast enhancement.