Parkinson’s disease (PD) is a neurodegenerative disorder that causes gait abnormalities. Early and accurate recognition of PD gait is crucial for effective treatment. Recently, deep learning techniques have shown promising results in PD gait analysis. However, most existing methods focus on severity estimation and frozen gait detection, while the recognition of Parkinsonian gait and normal gait from the forward video has not been reported. In this paper, we propose a novel spatiotemporal modeling method for PD gait recognition, named WM–STGCN, which utilizes a Weighted adjacency matrix with virtual connection and Multi-scale temporal convolution in a Spatiotemporal Graph Convolution Network. The weighted matrix enables different intensities to be assigned to different spatial features, including virtual connections, while the multi-scale temporal convolution helps to effectively capture the temporal features at different scales. Moreover, we employ various approaches to augment skeleton data. Experimental results show that our proposed method achieved the best accuracy of 87.1% and an F1 score of 92.85%, outperforming Long short-term memory (LSTM), K-nearest neighbors (KNN), Decision tree, AdaBoost, and ST–GCN models. Our proposed WM–STGCN provides an effective spatiotemporal modeling method for PD gait recognition that outperforms existing methods. It has the potential for clinical application in PD diagnosis and treatment.

Zero-shot 3D visual grounding (ZS-3DVG) refers to the task of grounding an object within a 3D scene based on a natural-language description, particularly for categories unseen during training. In this task, color serves as a crucial cue, but its use is hindered by challenges in both accuracy and speed. Accuracy is degraded by the instability of the RGB color space under varying illumination and the semantic gap between the standard color definitions and human perception. Speed is hampered by the computational overhead of rendering 3D point clouds into multiple 2D views. Although more views can improve color recognition, this incurs substantial latency, making real-time applications infeasible. To address these challenges, we propose a lightweight module that enhances both speed and accuracy. With respect to speed, our module bypasses the rendering pipeline and operates directly on HSV histograms extracted from 3D point clouds, significantly reducing the computation. For accuracy, our module leverages a key property of the HSV color space: chromatic colors concentrate along the hue axis, whereas achromatic colors lie on the value axis. Based on this, we mitigate the RGB instability by routing the input data to specialized classifiers focusing on the relevant axis. Furthermore, we resolved the semantic mismatch by grouping perceptually similar colors. The experimental results show that our method achieves higher accuracy and faster inference than RGB- or CLIP-based baselines, enabling robust and efficient color recognition in ZS-3DVG.