Skip to main content
. 2025 Aug 11;16:1596408. doi: 10.3389/fneur.2025.1596408

Figure 8.

Radar charts comparing performance metrics on AFLW and PoseTrack datasets. Each chart includes Recall, F1 Score, Accuracy, and AUC. Four models are represented: w/o GBTL, w/o MMDF, w/o DCO, and Ours, with Ours achieving the highest values on both datasets.

Analysis of the ablation study for our approach on the AFLW and PoseTrack Datasets. Graph-based temporal learning(GBTL), multi-modal data fusion (MMDF), domain-constrained optimization (DCO).