Cooperative Inference for Real-Time 3D Human Pose Estimation in Multi-Device Edge Networks

2025-04-03

AI Summary

In resource-constrained edge environments, achieving both high accuracy and low latency for 3D human pose estimation remains challenging. Method: This paper proposes a multi-device collaborative inference framework. Its core innovations include: (1) a novel dual-confidence-threshold-driven dynamic offloading mechanism, where the edge device identifies ambiguous samples based on local confidence and selectively offloads them to the edge server for re-estimation; and (2) modeling MPJPE minimization as a joint accuracy–latency optimization problem across devices, solved via non-convex problem decomposition and a low-complexity algorithm to determine optimal thresholds and latency allocation. The framework integrates a lightweight on-device model, confidence-aware sample filtering, and collaborative inference. Contribution/Results: Under strict end-to-end latency constraints, it significantly reduces MPJPE. Experiments demonstrate an effective accuracy–latency Pareto trade-off and robust high performance across diverse multi-access edge computing (MEC) scenarios.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Computer Vision: 3D Computer Vision
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

Accurate and real-time three-dimensional (3D) pose estimation is challenging in resource-constrained and dynamic environments owing to its high computational complexity. To address this issue, this study proposes a novel cooperative inference method for real-time 3D human pose estimation in mobile edge computing (MEC) networks. In the proposed method, multiple end devices equipped with lightweight inference models employ dual confidence thresholds to filter ambiguous images. Only the filtered images are offloaded to an edge server with a more powerful inference model for re-evaluation, thereby improving the estimation accuracy under computational and communication constraints. We numerically analyze the performance of the proposed inference method in terms of the inference accuracy and end-to-end delay and formulate a joint optimization problem to derive the optimal confidence thresholds and transmission time for each device, with the objective of minimizing the mean per-joint position error (MPJPE) while satisfying the required end-to-end delay constraint. To solve this problem, we demonstrate that minimizing the MPJPE is equivalent to maximizing the sum of the inference accuracies for all devices, decompose the problem into manageable subproblems, and present a low-complexity optimization algorithm to obtain a near-optimal solution. The experimental results show that a trade-off exists between the MPJPE and end-to-end delay depending on the confidence thresholds. Furthermore, the results confirm that the proposed cooperative inference method achieves a significant reduction in the MPJPE through the optimal selection of confidence thresholds and transmission times, while consistently satisfying the end-to-end delay requirement in various MEC environments.

Problem

• Real-time 3D human pose estimation in resource-constrained edge networks
• Optimizing confidence thresholds and transmission times for accuracy
• Balancing inference accuracy and end-to-end delay in MEC

Innovation

• Cooperative inference with dual confidence thresholds
• Optimal thresholds and transmission time optimization
• Lightweight models filter images for edge re-evaluation