Activity Recognition on VAST
93.5AccuracyLogi-PAR
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Logi-PARCategory=Neuro-Sym, Backbone=Swin-B2026.03 | 93.5 | 91.2 | 92.4 | 91.8 | 0.96 | 4 | |
| Dual-CausalCategory=Causal, Backbone=ResNet-1012026.03 | 90.5 | 88.9 | 90.2 | 89.5 | 0.94 | 6 | |
| InternVideo2Category=VLM, Backbone=InternViT-6B2026.03 | 90.1 | 91.5 | 78.2 | 84.3 | 0.9 | 18 | |
| LanguageBindCategory=VLM, Backbone=CLIP-ViT-L2026.03 | 89.8 | 90.1 | 79.5 | 84.5 | 0.9 | 16 | |
| ContextGPTCategory=Neuro-Sym, Backbone=LLMs2026.03 | 89.5 | 87.2 | 88.1 | 87.6 | 0.92 | 8 | |
| UniFormerV2Category=Vis-T, Backbone=UniFormer-L2026.03 | 89.1 | 86.8 | 84.5 | 85.6 | 0.91 | 10 | |
| VideoLLaMA2Category=VLM, Backbone=Mistral-7B2026.03 | 88.6 | 90.4 | 74.2 | 81.5 | 0.87 | 21 | |
| VideoMAE-V2Category=Vis-T, Backbone=ViT-g2026.03 | 88.4 | 85.2 | 82.1 | 83.6 | 0.89 | 12 | |
| NS-VQA (Med)Category=Neuro-Sym, Backbone=ViT2026.03 | 87.8 | 85.5 | 89.4 | 87.4 | 0.91 | 7 | |
| Swin-UNETRCategory=Medical-T, Backbone=Swin-B2026.03 | 86.5 | 82.4 | 85.1 | 83.7 | 0.88 | 9 | |
| Med-flamingo (Few-Shot)Category=VLM, Backbone=Openflamingo-9b2026.03 | 85.3 | 92.8 | 68.4 | 78.7 | 0.85 | 25 | |
| ClinActNetCategory=Multimodal, Backbone=Transformer2026.03 | 84.1 | 81.5 | 80.2 | 82 | 0.86 | 19 | |
| ActionCLIPCategory=Vis-VLM, Backbone=ViT-B/162026.03 | 82.3 | 80.1 | 78.5 | 79.4 | 0.85 | 22 | |
| SlowFastCategory=Vis-CNN, Backbone=ResNet-502026.03 | 79.5 | 76.2 | 74.1 | 75.8 | 0.82 | 28 |