Edge-GPU Based Face Tracking for Face Detection and Recognition Acceleration

2025-05-07

AI Summary

To address low throughput and high power consumption in real-time face detection and recognition for public scenarios on edge GPUs, this paper proposes a hardware-software co-optimization framework targeting the NVIDIA Jetson AGX Orin platform. We introduce, for the first time on Orin, holistic hardware orchestration across CPU, DLA, CUDA, and ISP units, coupled with a lightweight KCF–DeepSORT fusion tracking module that dynamically triggers recognition only upon stable face trajectory establishment—eliminating redundant per-frame inference. Further optimized via TensorRT acceleration and an adaptive ROI-based recognition pipeline, the system achieves 290 FPS on 1920×1080 video streams (average 6 faces/frame) while reducing power consumption by 800 mW. The approach significantly outperforms CPU- or GPU-only baselines in energy efficiency. Our core contributions are: (1) a unified hardware-parallel scheduling framework exploiting all available Orin accelerators, and (2) a tracking-driven sparse recognition mechanism that balances accuracy, latency, and energy.

Technology Category

• Machine Learning: Hardware-aware ML
• Computer Vision: Motion & Tracking
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Cost-effective machine vision systems dedicated to real-time and accurate face detection and recognition in public places are crucial for many modern applications. However, despite their high performance, which could be reached using specialized edge or cloud AI hardware accelerators, there is still room for improvement in throughput and power consumption. This paper aims to suggest a combined hardware-software approach that optimizes face detection and recognition systems on one of the latest edge GPUs, namely NVIDIA Jetson AGX Orin. First, it leverages the simultaneous usage of all its hardware engines to improve processing time. This offers an improvement over previous works where these tasks were mainly allocated automatically and exclusively to the CPU or, to a higher extent, to the GPU core. Additionally, the paper suggests integrating a face tracker module to avoid redundantly running the face recognition algorithm for every frame but only when a new face appears in the scene. The results of extended experiments suggest that simultaneous usage of all the hardware engines that are available in the Orin GPU and tracker integration into the pipeline yield an impressive throughput of 290 FPS (frames per second) on 1920 x 1080 input size frames containing in average of 6 faces/frame. Additionally, a substantial saving of power consumption of around 800 mW was achieved when compared to running the task on the CPU/GPU engines only and without integrating a tracker into the Orin GPU'92s pipeline. This hardware-codesign approach can pave the way to design high-performance machine vision systems at the edge, critically needed in video monitoring in public places where several nearby cameras are usually deployed for a same scene.

Problem

• Optimizing face detection and recognition on edge GPUs
• Reducing power consumption in real-time face tracking systems
• Improving throughput using combined hardware-software approaches

Innovation

• Utilizes all NVIDIA Jetson AGX Orin hardware engines
• Integrates face tracker to reduce redundant processing
• Achieves 290 FPS with 800 mW power savings