how to directly recognize image using my own trained CNN?

I have trained CNN using my own data set of text images. there are 1500 classes in my network. Now I want recognize images one-by-one trough the trained network. I am calling 'classify' to recognize a single image but produces the following error:

label = classify(convnet,I,1500) % convnet is my trained network, I is image to recognize, 1500 is number of groups

Error using classify (line 145) The length of GROUP must equal the number of rows in TRAINING.

How can I figure it out?