Active Object Recognition with a Space-Variant Retina

Figure 9

The total number of units learned for a gnostic set as a function of the number of fixations extracted from images labeled with the gnostic set’s category. The black dot indicates the number of units learned for each Caltech-256 category with 50 training images per category and 500 fixations per image, so each gnostic set contained 1239 units learned from 25000 fixations.