Image sources separation for color mixtures

Abstract

In this paper, we have a mixture of two added images and it is required to separate the two image sources from the combined one after receiving the combined image and one of the two source images. The Discrete Wavelet Transform (DWT) is applied on the combined image. Also, it is applied on the corrupted and received image source. The high frequency components of the combined image and the image source are correlated. The result of correlation is used in classifying the pixel whether belongs to image source or zero. The experiment is applied on the other image source and the combined image. The performance is measured and compared to a recent paper in the same topic and the proposed method is proved to have better results than the results of the recent paper. In the recent paper, the color mixture is not considered, it is considered in our work and gave very good results.