Hello. I have a 3-band ASTER image that I classified. I exported the training and testing pixels as ASCII and used LIBSVM to do the classification. I got close to 95% accuracy. LIBSVM also gave me the proper SVM parameters (slack and penalty). What I want to do now is bring these back into ENVI somehow so that I can get a classified image to compare with the maximum likelihood and neural net classification images that I derived completely with ENVI. How can I go about using my LIBSVM work back in ENVI to create this classified image? Thank you.