GPU-based Pedestrian Detection for Autonomous Driving

TL;DR

This paper presents a real-time pedestrian detection system optimized for Nvidia Tegra X1, combining LBP, HOG, sliding window, and SVM, achieving significant speedup and energy efficiency improvements over desktop solutions.

Contribution

It introduces a GPU-accelerated pedestrian detection pipeline tailored for embedded platforms, demonstrating real-time performance and energy efficiency benefits.

Findings

8x speedup on Nvidia Tegra X1

Better performance/watt ratio than desktop CUDA platforms

Effective real-time pedestrian detection on embedded hardware

Abstract

We propose a real-time pedestrian detection system for the embedded Nvidia Tegra X1 GPU-CPU hybrid platform. The pipeline is composed by the following state-of-the-art algorithms: Histogram of Local Binary Patterns (LBP) and Histograms of Oriented Gradients (HOG) features extracted from the input image; Pyramidal Sliding Window technique for candidate generation; and Support Vector Machine (SVM) for classification. Results show a 8x speedup in the target Tegra X1 platform and a better performance/watt ratio than desktop CUDA platforms in study.