Implementation and Comparison of a Fast Shape Recognition Algorithm using Different FPGA Platforms

In the present work, we implement and compare a fast shape recognition algorithm in two FPGA platforms: ZYBO and VIRTEX-5. We successfully implement our algorithm for several image resolutions in both hardware. We have found that VIRTEX-5 provides 14-19% faster processing than ZYBO, making it ideal for real time execution and higher resolution algorithms for computer vision, however, this comes at a cost of 8-10% more power consumption with respect to ZYBO platform. These results show that both platforms are good candidates for implementing this kind of algorithms depending on the focus of the hardware: performance or power saving.