Implementation of a CNN based object counting algorithm on bi-i cellular vision system

Object counting has been used in many areas such as medical and industrial applications. It is a challenging problem to count the target objects in high speed. It is useful to implement image processing applications using the high capability computational power offered by Cellular Neural Network type analog processor named as ACE16k. In this paper, we implement an efficient object counting algorithm working on ACE16k chip. Our results have proved that the proposed algorithm can count objects on a given image rapidly and accurately.