Adaptive color enhancement based on multi-scaled Retinex using local contrast of the input image

Methods of contrast enhancement and color correction are currently among the most active fields in imaging research. The tone curve or histogram of an image is generally used to improve contrast and detail, even though this is often unsatisfactory because the intensity and chromaticity of the illumination vary with geometric position. This has led to the development of the multi-scaled Retinex algorithm in which the in fluence of non-uniform illumination is reduced by partitioning the original image using local average images that are estimated based on Gaussian filtering. However, this algorithm produces color distortion because of the dominant chromaticity of the illumination. To solve this problem, this study introduces a new color correction method for digital images inspired by the multi-scale Retinex algorithm. The method includes selecting appropriate parameters for the Gaussian filters because this is the most important factor in reducing the artifacts introduced by the conventional multi-scale Retinex algorithm. Measures of visual contrast and halo are used to check the condition of artifacts for different values of the Gaussian parameters. The illuminant component is estimated to correct its chromaticity. A saturation compensation method based on preserving the chroma ratio is also used to compensate for the possible lack of saturation resulting from the modified multi-scale Retinex model.