DGHMesh: A Large-scale Dual-radar mmWave Dataset and Generalization-focused Benchmark for Human Mesh Reconstruction

2026-04-19

AI Summary

This work addresses the lack of benchmark datasets for evaluating generalization in millimeter-wave radar-based human mesh reconstruction, which hinders fair performance comparison under varying configurations. To this end, we introduce DGHMesh, a large-scale dual-modality (FMCW/SFCW) millimeter-wave dataset that, for the first time, provides synchronized raw I/Q data from dual radars alongside high-fidelity 3D human mesh annotations. We further establish a comprehensive generalization benchmark encompassing diverse configuration shifts, including human position, orientation, subarray scale, and cross-subject scenarios. Concurrently, we propose mmPTM, a query-based multi-radar fusion framework that jointly models point clouds and imaging tubes for efficient feature integration. Experiments demonstrate that mmPTM achieves superior accuracy and strong generalization across multiple sub-benchmarks, validating the effectiveness of both the proposed dataset and methodology.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Computer Vision: Multi-modal Vision

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Security and Privacy: Large-scale security measurements

Abstract

Millimeter-wave (mmWave) radar has shown great potential for contactless, privacy-preserving, and robust human sensing, yet existing mmWave-based human mesh reconstruction (HMR) studies are still limited by the lack of benchmarks for generalization analysis under configuration shifts and fair comparison of different algorithms. To address the limitation, we present DGHMesh, a large-scale dual-radar mmWave dataset and generalization-focused benchmark for HMR. It contains data from 15 subjects performing 8 actions, with 360,000 synchronized frames collected from FMCW radar, SFCW radar, RGB images, and high-precision 3D HMR annotations. In addition, the dataset provides synchronized raw I/Q data from both radar modalities and accurately calibrated radar spatial positions. The benchmark is designed to evaluate HMR methods under diverse measurement configurations, including human position shifts, human orientation shifts, subarray size variations, and cross-subject settings. Based on DGHMesh, we also propose mmPTM, a query-based multi-radar fusion framework that jointly exploits point clouds and imaging tubes for HMR. Extensive experiments are conducted against representative baselines under different settings. The results demonstrate that mmPTM consistently achieves outstanding accuracy and competitive generalization capability across multiple sub-benchmarks, validating the effectiveness of multi-radar fusion and the practical value of the proposed dataset and benchmark for mmWave-based HMR research. DGHMesh and mmPTM are publicly available at https://github.com/SPIresearch/DGHMesh.(The complete benchmark and code will be released after paper publication)

Problem

• mmWave radar
• human mesh reconstruction
• generalization benchmark
• configuration shifts
• dual-radar

Innovation

• mmWave radar
• human mesh reconstruction
• generalization benchmark
• multi-radar fusion
• dual-radar dataset