Research on Blind Separation of Image Sources against Rotation Based on Convex Optimization

Blind source separation is how to recover a set of signals from a set of their observations, without any priori knowledge of sources. In this paper, a novel blind source separation algorithm of image signals against rotation based on the convex analysis of mixtures of non-negative sources is proposed. This new method firstly preprocesses the observations, and then extracts the rotation invariant factor, according to the special assumption called local dominance which is showed in the convex analysis of mixtures of non-negative sources algorithm, the issue of blind separation of image sources which is influenced by rotation turns into a solvable convex optimization, through which the mixing matrix can be determined. Finally by solving the mixing equation group to obtain the image sources. Experimental results demonstrate that this novel algorithm is quite effective for blind separation of image sources against rotation and shows 80 percent increase in the performance index compared to ICA, NMF and CAMNS algorithms.