Fig. 12 | Scientific Reports

Fig. 12

From: Enhanced spatiotemporal skeleton modeling: integrating part-joint attention with dynamic graph convolution

Fig. 12

Ablation study showing the contribution of the Part-Joint Attention (PJA) and Dynamic Graph Convolution Network (DGCN) modules. The x-axis represents prediction horizons (80 ms, 160 ms, 320 ms, 400 ms), and the y-axis shows the Mean Per Joint Position Error (MPJPE) in millimeters. Lower MPJPE indicates better performance. The figure demonstrates that PJA contributes more to short-term predictions, while DGCN improves long-term motion forecasting accuracy.

Back to article page