A bio-inspired tracking camera system

We propose a bio-inspired reconfigurable tracking camera system using FPGAs. In the system, a wide-lens camera captures an entire image, while a zoom-lens camera tracks a target in the image and magnifies it. In the current system, a probabilistic neural network (PNN) and mosaic processing are used for the image recognition and preprocessing, respectively. Thanks to the FPGA-based design, not only the PNN and the mosaic processing, but also other recognition and preprocessing algorithms can be implemented onto the system to adapt various images and targets.