Application of Object Detection Algorithm on SoC-FPGA using OpenCL

Nowadays, object detection algorithms have been widely applied in various aspects of life. However, along with the advancement of these algorithms comes high computational complexity, which requires significant processing time. One approach to reduce processing time is to enhance computational speed by implementing Convolutional Neural Networks (CNNs) on Field-Programmable Gate Arrays (FPGAs). However, using FPGAs presents a challenge, as it requires hardware description languages to design complex CNNs, which can be time-consuming. To overcome this challenge, this paper proposes a novel hardware design approach leveraging the Intel® FPGA SDK (Software Development Kit) for $\mathrm{O}\mathrm{p}\mathrm{e}\mathrm{n}\mathrm{C}\mathrm{L}^{\mathrm{T}\mathrm{M}}$ to implement hardware using the C language with the OpenCL framework. This approach involves implementing a CNN on the DE1-SoC board specifically for object detection tasks.