Real-time HOG+SVM based object detection using SoC FPGA for a UHD video stream

TL;DR

This paper presents a real-time pedestrian detection system using HOG and SVM on an FPGA-based SoC, capable of processing UHD video at 60 fps, demonstrating the effectiveness of reprogrammable devices for embedded vision.

Contribution

The paper introduces a hardware implementation of HOG+SVM for real-time pedestrian detection on a UHD video stream using an FPGA-based SoC, enabling high-resolution processing at 60 fps.

Findings

Achieved real-time processing of 4K UHD video at 60 fps.

Demonstrated the suitability of FPGA-based reprogrammable devices for embedded vision.

Confirmed high detection accuracy with a single-scale detector.

Abstract

Object detection is an essential component of many vision systems. For example, pedestrian detection is used in advanced driver assistance systems (ADAS) and advanced video surveillance systems (AVSS). Currently, most detectors use deep convolutional neural networks (e.g., the YOLO -- You Only Look Once -- family), which, however, due to their high computational complexity, are not able to process a very high-resolution video stream in real-time, especially within a limited energy budget. In this paper we present a hardware implementation of the well-known pedestrian detector with HOG (Histogram of Oriented Gradients) feature extraction and SVM (Support Vector Machine) classification. Our system running on AMD Xilinx Zynq UltraScale+ MPSoC (Multiprocessor System on Chip) device allows real-time processing of 4K resolution (UHD -- Ultra High Definition, 3840 x 2160 pixels) video for 60 frames per second. The system is capable of detecting a pedestrian in a single scale. The results obtained confirm the high suitability of reprogrammable devices in the real-time implementation of embedded vision systems.