Boxes

Grayscale input images originate from a digital camera.

On the first picture is the right box in focus, on the second picture is the left box in focus.

first input image	second input image

There are illustrated some of our experimental results. We show here the SA outputs with m, n =200 and note that it is comparable (measures of quality are almost equal) to the CA. Because we did not have a whole, ideal image at our disposal, we can only compare ''ideal'' parts of the input images with their respective parts in the fused image. It is easily seen that the quality of fusion in the background region (left box) is better than that in the foreground region (right box). This was caused by differences in reflections from the surfaces of the boxes. It seems that the initial setting values did not play a significant role in this application. Note that the application of the simplest SA, with m, n = 1, produced a very good fusion.

SA with n,m=1	SA with n,m=200	CA with nstart,mstart=1,kmax=8,step=3

© 2007 Institute for Research and Applications of Fuzzy Modeling, University of Ostrava, Czech Republic