Audio-Visual Video Parsing on LLP (test)
81.7Audio Segment ScoreEAR
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| EAR2026.05 | 81.7 | 75.1 | 75.5 | 71.3 | 68 | 80.8 | 75.1 | 69.5 | — | 72.9 | 73.2 | |
| EAR2026.05 | 81.7 | 75.1 | 75.5 | 71.3 | 68 | 62 | 75.1 | 69.5 | 80.8 | 72.9 | 73.2 | |
| VAPLAN2026.05 | 80.4 | 71.7 | 73 | 68.3 | — | — | — | — | — | — | — | |
| VALOR2026.05 | 80.4 | 72.2 | 71.8 | 65.9 | 63.7 | 79.7 | 72 | 64.6 | — | 68 | 69.4 | |
| VAPLAN2026.05 | 80.4 | 71.7 | 73 | 68.3 | — | — | — | — | — | — | — | |
| VALOR2026.05 | 80.4 | 72.2 | 71.8 | 65.9 | 63.7 | 55.6 | 72 | 64.6 | 79.7 | 68 | 69.4 | |
| UWAV2026.05 | 78.4 | 71.1 | 74.5 | 69.6 | 65.5 | 78.4 | 72.8 | 66.1 | — | 69 | 70.3 | |
| UWAV2026.05 | 78.4 | 71.1 | 74.5 | 69.6 | 65.5 | 57.7 | 72.8 | 66.1 | 78.4 | 69 | 70.3 | |
| TeMTGVisual features=CLIP, Audio features=CLAP2026.05 | 74.4 | 61.9 | 72.9 | 69 | 62 | 53.2 | 69.8 | 61.4 | 74.1 | 62.2 | 66.1 | |
| EARVisual features=CLIP, Audio features=CLAP2026.05 | 71.3 | 66.4 | 71.8 | 68.1 | 66.2 | 61.4 | 69.8 | 65.3 | 70.1 | 63.9 | 67.4 | |
| NREPVisual features=CLIP, Audio features=CLAP2026.05 | 70.2 | 62.8 | 70.9 | 67.3 | 64.4 | 57.6 | 68.5 | 62.6 | 68.8 | 61.1 | 65.4 | |
| LINKVisual features=CLIP, Audio features=CLAP2026.05 | 69.7 | 63.4 | 69 | 64.9 | 62.1 | 55.7 | 66.9 | 61.3 | 68.5 | 60.8 | 64.2 | |
| MM-CSEVisual features=CLIP, Audio features=CLAP2026.05 | 69.5 | 63 | 71.3 | 67.5 | 64.2 | 57.8 | 68.3 | 62.7 | 68.9 | 61.1 | 65.4 | |
| TIPNetVisual features=CLIP, Audio features=CLAP2026.05 | 69.2 | 63 | 70.3 | 66.9 | 64.2 | 58.1 | 67.9 | 62.7 | 68.3 | 61.2 | 65.2 | |
| VAPLANVisual features=CLIP, Audio features=CLAP2026.05 | 69 | 61.9 | 70.2 | 66.4 | 63.5 | 56.9 | 67.6 | 61.7 | 67.9 | 60.1 | 64.5 | |
| UWAVVisual features=CLIP, Audio features=CLAP2026.05 | 68.9 | 63.5 | 72.3 | 68.7 | 65.6 | 59.6 | 68.9 | 63.9 | 68.3 | 62.4 | 66.2 | |
| LSLDVisual features=CLIP, Audio features=CLAP2026.05 | 68.7 | 61.5 | 71.3 | 67.4 | 63.4 | 55.9 | 67.8 | 61.6 | 68.2 | 60.6 | 64.6 | |
| VALORVisual features=CLIP, Audio features=CLAP2026.05 | 68.1 | 61.2 | 68.4 | 64.7 | 61.9 | 55.5 | 66.2 | 60.4 | 66.8 | 59 | 63.2 | |
| PPLVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 65.9 | 57.3 | 66.7 | 64.3 | 61.9 | 54.3 | 64.8 | 59.9 | 63.7 | 57.9 | 61.7 | |
| EARVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 65.9 | 60.2 | 69.8 | 66.5 | 63.8 | 58.8 | 66.5 | 61.8 | 64.8 | 58.5 | 63.7 | |
| PPLVisual features=ResNet-152 + 3D ResNet, Audio features=VGGish2026.05 | 65.9 | 57.3 | 66.7 | 64.3 | 61.9 | 54.3 | 64.8 | 59.9 | 63.7 | 57.9 | 61.7 | |
| EARVisual features=ResNet-152 + 3D ResNet, Audio features=VGGish2026.05 | 65.9 | 60.2 | 69.8 | 66.5 | 63.8 | 58.8 | 66.5 | 61.8 | 64.8 | 58.5 | 63.7 | |
| MM-CSEVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 65 | 59.1 | 66.8 | 64.1 | 60 | 54.9 | 63.9 | 59.4 | 64.2 | 57.6 | 61.5 | |
| MM-CSEVisual features=ResNet-152 + 3D ResNet, Audio features=VGGish2026.05 | 65 | 59.1 | 66.8 | 64.1 | 60 | 54.9 | 63.9 | 59.4 | 64.2 | 57.6 | 61.5 | |
| LEAP (MM-pyramid)Visual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 64.8 | 59.2 | 67.7 | 64.9 | 61.8 | 56.5 | 64.8 | 60.2 | 63.6 | 57.4 | 62.1 | |
| Mettle2026.05 | 64.3 | 56.1 | 55.9 | 52.2 | 50.9 | 45.1 | 57 | 51.1 | 61.2 | 53.1 | — | |
| CMPAEVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 64.2 | 56.6 | 66.4 | 63.7 | 59.2 | 51.8 | 63.3 | 51.9 | 62.8 | 55.7 | 59.6 | |
| UWAVVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 64.2 | 58.6 | 70 | 66.7 | 63.4 | 57.5 | 65.9 | 60.9 | 63.9 | 57.4 | 62.8 | |
| UWAVVisual features=ResNet-152 + 3D ResNet, Audio features=VGGish2026.05 | 64.2 | 58.6 | 70 | 66.7 | 63.4 | 57.5 | 65.9 | 60.9 | 63.9 | 57.4 | 62.8 | |
| SDDPVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 64 | 57.2 | 66.8 | 62.5 | 59.6 | 52.6 | 63.5 | 57.4 | 63.2 | 55.2 | 60.2 | |
| SDDPVisual features=ResNet-152 + 3D ResNet, Audio features=VGGish2026.05 | 64 | 57.2 | 66.8 | 62.5 | 59.6 | 52.6 | 63.5 | 57.4 | 63.2 | 55.2 | 60.2 | |
| CAE-AV2026.02 | 63.8 | 53.3 | 59.6 | 56 | 53.9 | 47 | 59 | 52.8 | 60.4 | 51.6 | — | |
| CoLeafVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 63.5 | 56.2 | 64.5 | 62 | 58.6 | 52.1 | 62.1 | 56.7 | 61.8 | 54.5 | 59.2 | |
| CoLeafVisual features=ResNet-152 + 3D ResNet, Audio features=VGGish2026.05 | 63.5 | 56.2 | 64.5 | 62 | 58.6 | 52.1 | 62.1 | 56.7 | 61.8 | 54.5 | 59.2 | |
| CMRNVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 63.2 | 55.7 | 63.9 | 60.4 | 57 | 49.3 | 61.4 | 55.1 | 61.4 | 54.2 | 58.2 | |
| Rachavarapu et al.Visual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 63.1 | 54.1 | 63.5 | 60.3 | 57.7 | 51.5 | 61.4 | 55.2 | 60.6 | 52.3 | 58 | |
| VALORVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 62.8 | 57.1 | 66.7 | 63.9 | 60 | 54.4 | 63.2 | 58.5 | 62.3 | 55.9 | 60.5 | |
| VALORVisual features=ResNet-152 + 3D ResNet, Audio features=VGGish2026.05 | 62.8 | 57.1 | 66.7 | 63.9 | 60 | 54.4 | 63.2 | 58.5 | 62.3 | 55.9 | 60.5 | |
| LEAP (HAN)Visual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 62.7 | 56.4 | 65.6 | 63.1 | 59.3 | 54.1 | 62.5 | 57.8 | 61.8 | 55 | 59.8 | |
| LSLDVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 62.7 | 55.7 | 67.1 | 64.3 | 59.4 | 52.6 | 63.1 | 57.6 | 62.2 | 55.2 | 60 | |
| LSLDVisual features=ResNet-152 + 3D ResNet, Audio features=VGGish2026.05 | 62.7 | 55.7 | 67.1 | 64.3 | 59.4 | 52.6 | 63.1 | 57.6 | 62.2 | 55.2 | 60 | |
| HAN+PPL2026.05 | 62.5 | 55.4 | 55.3 | 51.1 | 52.3 | 58.3 | 56 | 50.9 | — | 50.6 | 53.9 | |
| HAN+PPL2026.05 | 62.5 | 55.4 | 55.3 | 51.1 | 52.3 | 46.9 | 56 | 50.9 | 58.3 | 50.6 | 53.9 | |
| VAPLANVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 62.4 | 55.7 | 66.7 | 63.3 | 60.3 | 53.7 | 63.1 | 57.6 | 61.4 | 54.3 | 59.9 | |
| VAPLANVisual features=ResNet-152 + 3D ResNet, Audio features=VGGish2026.05 | 62.4 | 55.7 | 66.7 | 63.3 | 60.3 | 53.7 | 63.1 | 57.6 | 61.4 | 54.3 | 59.9 | |
| LGFNetVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 62.2 | 54.4 | 64 | 60.3 | 56.9 | 49.8 | 61 | 54.8 | 60.5 | 52.9 | 57.7 | |
| AVMoE2026.02 | 62.1 | 51.8 | 60 | 55.7 | 54.4 | 47.6 | 58.8 | 51.7 | 59 | 50.2 | — | |
| Xu et al.Visual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 61.9 | 53.9 | 64.8 | 61.6 | 57.6 | 50.2 | 61.4 | 55.2 | 60.9 | 53.1 | 58.1 | |
| CM-PIEVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 61.7 | 53.7 | 55.2 | 51.3 | 50.1 | 43.6 | 55.7 | 49.5 | 56.8 | 51.3 | 52.9 | |
| DHHNVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 61.7 | 54.8 | 63.2 | 60.4 | 56.8 | 51.1 | 60.6 | 55.4 | 59.7 | 53.3 | 57.7 | |
| MA+PPL2026.05 | 61.7 | 55.4 | 61.8 | 57.9 | 57.5 | 59.4 | 60.6 | 55 | — | 52.6 | 57.4 | |
| MA+PPL2026.05 | 61.7 | 55.4 | 61.8 | 57.9 | 57.5 | 51.6 | 60.6 | 55 | 59.4 | 52.6 | 57.4 | |
| AV-UnifiedBackbone=CLIP ViT-L-142026.03 | 61.6 | 54.8 | 66.9 | 63 | 59.8 | 52.2 | 62.8 | 56.7 | 61.5 | 53.9 | — | |
| JoMoLDVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 61.6 | 53.5 | 63.4 | 59.8 | 57 | 50 | 60.5 | 54.4 | 60 | 52.2 | 57.2 | |
| JoMoLDVisual features=ResNet-152 + 3D ResNet, Audio features=VGGish2026.05 | 61.6 | 53.5 | 63.4 | 59.8 | 57 | 50 | 60.5 | 54.4 | 60 | 52.2 | 57.2 | |
| DHHN2026.03 | 61.4 | 54.6 | 63.4 | 60.8 | 56.8 | 51.1 | 60.5 | 55.5 | 59.5 | 53.3 | — | |
| AVFASVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 61.4 | 54.2 | 64.5 | 61.9 | 58.4 | 51.9 | 61.4 | 56 | 59.9 | 53 | 58.3 | |
| MM-Pyramid2026.03 | 61.1 | 53.8 | 60.3 | 56.7 | 55.8 | 49.4 | 59.7 | 54.1 | 59.1 | 51.2 | — | |
| MM-pyramidVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 61.1 | 53.8 | 60.3 | 56.7 | 55.8 | 49.4 | 59.7 | 54.1 | 59.1 | 51.2 | 56.1 | |
| CVCMS2026.03 | 60.8 | 53.8 | 63.5 | 58.9 | 57 | 49.5 | 60.5 | 54 | 59.5 | 52.1 | — | |
| MGN2026.03 | 60.8 | 51.1 | 55.4 | 52.4 | 50.4 | 44.4 | 55.5 | 49.3 | 57.2 | 49.1 | — | |
| MGNVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 60.8 | 51.1 | 55.4 | 52.4 | 50.4 | 44.4 | 55.5 | 49.3 | 57.2 | 49.1 | 52.6 | |
| CVCMSVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 60.8 | 53.8 | 63.5 | 58.9 | 57 | 49.5 | 60.5 | 54 | 59.5 | 52.1 | 57 | |
| MGN2026.02 | 60.7 | 51 | 55.5 | 52.4 | 50.6 | 44.4 | 55.6 | 49.3 | 57.2 | 49.2 | — | |
| MGN2026.05 | 60.7 | 51 | 55.5 | 52.4 | 50.6 | 44.4 | 55.6 | 49.3 | 57.2 | 49.2 | — | |
| AVMoE2026.05 | 60.5 | 50.8 | 58.9 | 54.9 | 53.2 | 46.3 | 57.5 | 50.7 | 57.8 | 49.4 | — | |
| MA2026.03 | 60.3 | 53.6 | 60 | 56.4 | 55.1 | 49 | 58.9 | 53 | 57.9 | 50.6 | — | |
| MAVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 60.3 | 53.6 | 60 | 56.4 | 55.1 | 49 | 58.9 | 53 | 57.9 | 50.6 | 55.5 | |
| MAVisual features=ResNet-152 + 3D ResNet, Audio features=VGGish2026.05 | 60.3 | 53.6 | 60 | 56.4 | 55.1 | 49 | 58.9 | 53 | 57.9 | 50.6 | 55.5 | |
| CMBSVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 60.2 | 51.1 | 54.3 | 50.8 | 50 | 43.7 | 54.8 | 48.5 | 55.7 | 48.3 | 51.7 | |
| HAN2026.02 | 60.1 | 51.3 | 52.9 | 48.9 | 48.9 | 43 | 54 | 47.7 | 55.4 | 48 | — | |
| HAN2026.03 | 60.1 | 51.3 | 52.9 | 48.9 | 48.9 | 43 | 54 | 47.7 | 55.4 | 48 | — | |
| HANVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 60.1 | 51.3 | 52.9 | 48.9 | 48.9 | 43 | 54 | 47.7 | 55.4 | 48 | 51 | |
| HANVisual features=ResNet-152 + 3D ResNet, Audio features=VGGish2026.05 | 60.1 | 51.3 | 52.9 | 48.9 | 48.9 | 43 | 54 | 47.7 | 55.4 | 48 | 51 | |
| HAN2026.05 | 60.1 | 51.3 | 52.9 | 48.9 | 48.9 | 43 | 54 | 47.7 | 55.4 | 48 | — | |
| DG-SCT2026.02 | 59 | 49.2 | 59.4 | 56.1 | 52.8 | 46.1 | 57.1 | 50.5 | 57 | 49.1 | — | |
| DG-SCT2026.05 | 59 | 49.2 | 59.4 | 56.1 | 52.8 | 46.1 | 57.1 | 50.5 | 57 | 49.1 | — | |
| TB-AVA2026.05 | 56.4 | 48.3 | 61.1 | 56.7 | 55.6 | 48.6 | 57.7 | 51.2 | 54.9 | 47.9 | — | |
| AVEL2026.02 | 49.9 | 43.6 | 37.3 | 32.4 | 37 | 32.6 | 41.4 | 36.2 | 43.6 | 37.4 | — | |
| AVEVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 49.9 | 43.6 | 37.3 | 32.4 | 37 | 32.6 | 41.4 | 36.2 | 43.6 | 37.4 | 39.1 | |
| AVE2026.05 | 49.9 | 43.6 | 37.3 | 32.4 | 37 | 32.6 | 41.4 | 36.2 | 43.6 | 37.4 | — | |
| AVSDN2026.02 | 47.8 | 34.1 | 52 | 46.3 | 37.1 | 26.5 | 45.7 | 35.6 | 50.8 | 37.7 | — | |
| AVSDN2026.03 | 47.8 | 34.1 | 52 | 46.3 | 37.1 | 26.5 | 45.7 | 35.6 | 50.8 | 37.7 | — | |
| AVSDNVisual Backbone=ResNet-152 + 3D ResNet, Audio Backbone=VGGish2026.05 | 47.8 | 34.1 | 52 | 46.3 | 37.1 | 26.5 | 45.7 | 35.6 | 50.8 | 37.7 | 41.4 | |
| AVEL2026.03 | 47.2 | 40.4 | 37.1 | 34.7 | 35.4 | 31.6 | 39.9 | 35.5 | 41.6 | 36.5 | — | |
| LanguageBind with AV2ATraining-free=true, Foundation Models=LanguageBind, AV2A module=true2025.03 | 40.9 | 35.9 | 57.4 | 54.4 | 59.1 | 52.3 | 52.4 | 47.5 | 43.4 | 38.9 | — | |
| CLIP + CLAP with AV2ATraining-free=true, Foundation Models=CLIP + CLAP, AV2A module=true2025.03 | 32.7 | 28.5 | 43.5 | 41.5 | 55.1 | 48.2 | 43.8 | 39.4 | 34.7 | 31.2 | — | |
| LanguageBindTraining-free=true, Foundation Models=LanguageBind2025.03 | 20.3 | 17.5 | 21.3 | 20 | 24.3 | 21 | 22 | 19.5 | 20.8 | 18.4 | — | |
| CLIP + CLAPTraining-free=true, Foundation Models=CLIP + CLAP2025.03 | 17.6 | 15.1 | 18.8 | 17.8 | 30 | 25.6 | 22.1 | 19.5 | 18.3 | 16.2 | — |