A Recurrent Temporal Model for Semantic Levels Categorization Based on Human Visual System

Figure 13

Some features related to the car, fish, helicopter, and dolphin categories are obviously shown in which these features represent the wheels for car categories and the dorsal, ventral, and caudal fins for fish categories. The image inverting does not affect these features. However, selecting features like the cabin’s overall shape and the dorsal fin for helicopter and dolphin categories, respectively, indicate the situation where the input inverted image cannot be recognized.