Can you provide more information about the images you have now that define the different classes? How are the classes defined, exactly? For example, are they vector layers? Or are they raster images with a certain pixel value indicating a particular class? And exactly what do you want as a result? Do you want a single vector dataset with all class vectors included in that dataset? Or do you want something like ENVI's classification format, which is a raster image in which each pixel is given a pixel value that indicates the class to which it is assigned. - Peg Exelis VIS Tech Support

I have a Quick Bird image and i use feature extraction to extract 8 classes of interest (grass, trees, buldings, roads, ect). I apply this function 8 time for 8 classes that i need with other tipes of segmentation, masks and ancilary data. Now i have 8 classification image and i need to combine the class of interest from each classification to one classification image. The final result must be a classification image because i need to use confusion matrix to see if is the best method for classification. It is a little bit urgent to know how can i solve this problem. I can't use only one feature extraction baed on rules for all classification.

Before i get an answer here, i got a solution somewere else. Thank you for reply Peg.