Multifocus image restoration by fusion methods

This paper compares the results of two different modified fusion schemes for multifocus images. The first method is based on the algorithm proposed by De and Chanda that uses the multiresolution signal decomposition by morphological wavelets. This paper proposes a new strategy for choosing the signal detail of the fused image in the reconstruction step. The second fusion scheme is pixel based and combines the information level measures proposed by Maruthi and Sankarasubramnian with the edge information obtained from the Canny edge filter. A comparison of results obtained using the original and improved methods are also presented.