We introduce a theoretical framework for measuring the information content in the edges extracted from a color image. The main difficulty in estimating the amount of information (differential entropy [1, 2]) in an image signal is to fit an appropriate probability mass function to the trichromatic image data. To estimate the amount of information in the edges extracted from a color image, we first convolve the image with a derivative filter. By fitting a Kotz-Type probability distribution to the convolved image, we then estimate the differential entropy of the edge coefficients as a measure of the uncertainty involved in the edge content of a postreceptoral chromatic image. The proposed estimation of differential entropy provides an efficient means of processing the edge content information under a variety of natural illuminations, which might be further used as a quantitative measure for evaluating color constant image retrieval.