Color Image Enhancement by Logarithmic Transformation in Fuzzy Domain

A new fuzzy logic and logarithmic transformation method for enhancing of low contrast color images is proposed. The input RGB image is converted into HSV (Hue, Saturation, Value) color space to preserve the chromatic information contained in the original image. To enhance the image, only the saturation component S and intensity value component V are transformed in fuzzy domain by using S-type membership functions with parameters calculated by solution of the maximum of fuzzy entropy optimization problem. Next, the fuzzy domain is transformed to the logarithmic one. An enhancement of local contrasts by the method of unsharp masking is fulfilled in a logarithmically transformed fuzzy domain for S and V components. After that, a number of inverse transformations is performed: from the logarithmic domain to the fuzzy one and afterwards to the spatial domain. The method of quantitative assessment of quality of color images is proposed. The examples show the effectiveness of the developed method for improving the quality of color images.