An Energy-Efficient Reconfigurable CNN Accelerator on PYNQ-Z2

This paper presents an FPGA-based hardware accelerator designed for object detection using Convolutional Neural Networks (CNNs). The Baseline system leverages the parallel processing capabilities of FPGAs to efficiently execute deep learning operations required for object detection tasks. A CNN model trained on standard datasets is implemented and mapped onto an FPGA using high-level synthesis tools. The design achieves a balance between speed and resource usage, making it suitable for applications that require low latency and power efficiency. Performance is evaluated in terms of resource utilization, processing time, and detection accuracy, confirming the effectiveness of the hardware solution in embedded object detection applications.