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

Figure 5

Choose the best number of feedback feeds and the best number of feature neurons participating in the feedback. To select the number of feedback loops and the number of feature neurons, the model has executed ranges from 1 to 20 feedback loops for the number of feature neurons between 1 to 8 neurons. There was no difference in accuracy increment after running the number of 8 feedback loops and the number of 6 feature neurons, and only the computational load of the model increased.