Blind Image Separation Using Composite Criteria based on Statistical Information and Sparse Representation
Fouad Boudjenouia, Karim Abed‐Meraim, Aladine Chetouani, Rachid Jennane · 2023
This paper deals with blind image separation by exploiting the statistical characteristics of the mixtures (information related to the sources independence) with the sparsity of the signals. More precisely, we investigate and compare the gain that can be reached by considering two different signal properties, namely the statistical independence and the sparse representation. These properties are considered in a ‘dual’ as well as in a ‘hybrid’ mode. Both approaches are compared, simulation and experimental results demonstrate the efficiency of the proposed methods.