Our target here is to see how decidenet performs on Categorized Crowd counting Tasks.
Like other traditional models, decidenet also fails to distill sitting and standing crowd. It is sure now that traditional models are of no good use for categorization