Exemplar Free Class Agnostic Counting

We tackle the task of Class Agnostic Counting, which aims to count objects in a novel object category at test time without any access labeled training data for that category. All previous class agnostic counting methods cannot work fully automated setting, and require computationally expensive adaptation. To address these challenges, we propose visual counter operates setting does not Our proposed approach first identifies exemplars from repeating an image, then counts objects. region proposal network identifying exemplars. After exemplars, obtain corresponding by using density estimation based Visual Counter. evaluate our on FSC-147 dataset, show it achieves superior performance compared existing approaches. code models are available at: https://github.com/Viresh-R/ExemplarFreeCounting.git .