Multi-Head Adaptive Graph Convolution Network for Sparse Point Cloud-Based Human Activity Recognition

2025-04-03

AI Summary

To address human activity recognition (HAR) from sparse, noisy mmWave radar point clouds in privacy-sensitive and low-light scenarios, this paper proposes the Multi-head Adaptive Graph Convolutional Network (MA-GCN). Unlike conventional graph convolutional networks with fixed kernel functions, MA-GCN introduces a novel Multi-head Adaptive Kernel (MAK) module that dynamically generates multiple geometry-aware convolutional kernels for each local neighborhood, enabling fine-grained modeling of local geometric structure while preserving global spatial context consistency. The framework jointly integrates sparse point cloud geometric representation, graph convolution, multi-head attention, and adaptive kernel learning into an end-to-end trainable architecture. Extensive experiments on multiple benchmark datasets demonstrate state-of-the-art performance, significantly improving recognition accuracy under severe sparsity and noise. The source code is publicly available.

Technology Category

• Machine Learning: Hardware-aware ML
• Planning, Routing, and Scheduling: Activity and Plan Recognition
• Computer Vision: Multi-modal Vision

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Security and Privacy: Large-scale security measurements

Abstract

Human activity recognition is increasingly vital for supporting independent living, particularly for the elderly and those in need of assistance. Domestic service robots with monitoring capabilities can enhance safety and provide essential support. Although image-based methods have advanced considerably in the past decade, their adoption remains limited by concerns over privacy and sensitivity to low-light or dark conditions. As an alternative, millimetre-wave (mmWave) radar can produce point cloud data which is privacy-preserving. However, processing the sparse and noisy point clouds remains a long-standing challenge. While graph-based methods and attention mechanisms show promise, they predominantly rely on"fixed"kernels; kernels that are applied uniformly across all neighbourhoods, highlighting the need for adaptive approaches that can dynamically adjust their kernels to the specific geometry of each local neighbourhood in point cloud data. To overcome this limitation, we introduce an adaptive approach within the graph convolutional framework. Instead of a single shared weight function, our Multi-Head Adaptive Kernel (MAK) module generates multiple dynamic kernels, each capturing different aspects of the local feature space. By progressively refining local features while maintaining global spatial context, our method enables convolution kernels to adapt to varying local features. Experimental results on benchmark datasets confirm the effectiveness of our approach, achieving state-of-the-art performance in human activity recognition. Our source code is made publicly available at: https://github.com/Gbouna/MAK-GCN

Problem

• Recognizing human activities from sparse noisy point clouds
• Overcoming privacy and low-light limitations of image-based methods
• Developing adaptive graph kernels for dynamic local feature extraction

Innovation

• Multi-Head Adaptive Kernel for dynamic feature capture
• Graph Convolutional Network adapts to local geometries
• Millimetre-wave radar for privacy-preserving point clouds