Here we proposed a novel method for the separation of two images. Initially we fused the two Images. From the fused Image we are finding the gradient of fused image. The gradient is finding the edge of the image. The gradient value will be 0 and the edges of the image value will differ from zero. After that we are finding the matched layers from the gradient image. From this layer we are separating the original source layer. We develop a sparse blind separation algorithm to estimate both parameterized motions and mixing coefficients. Then, a novel reconstruction approach is presented to recover layers, by utilizing not only the mixing model but also the statistical properties of natural images. The whole method can handle more general motions than translations, including scaling, rotations and other transformations. The separation method contains two main parts, one of which is to estimate the mixing model and the other is to reconstruct the layers according to the mixing model.