Single-source sound localization on VGGSound single-source (test)
60.2IoU@0.5Generation-Analysis-Refinement
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Generation-Analysis-RefinementN (Number of iterations in Stage 2)=52026.04 | 60.2 | 60.5 | 55.2 | |
| Generation-Analysis-RefinementN (Number of iterations in Stage 2)=32026.04 | 60.1 | 60.2 | 55 | |
| T-VSLMethod Category=Ours2024.04 | 53.7 | 48.1 | 52.9 | |
| AlignmentMethod Category=Current SOTA Based Methods, Pre-trained Encoders=AudioCLIP2024.04 | 50.8 | 45.3 | 50.2 | |
| AVGNMethod Category=Current SOTA Based Methods, Pre-trained Encoders=AudioCLIP2024.04 | 49.6 | 44.1 | 49.5 | |
| FNACMethod Category=Current SOTA Based Methods, Pre-trained Encoders=AudioCLIP2024.04 | 48.4 | 43.3 | 49.1 | |
| AudioCLIPMethod Category=CLIP-Based Baseline Methods, Pre-trained Encoders=AudioCLIP2024.04 | 47.4 | 42.8 | 48.5 | |
| OA-SSL2026.04 | 47.3 | 51.7 | 44.9 | |
| MarginNCEMethod Category=Current SOTA Based Methods, Pre-trained Encoders=AudioCLIP2024.04 | 46.9 | 42.5 | 48.3 | |
| DSOLMethod Category=Current SOTA Based Methods, Pre-trained Encoders=AudioCLIP2024.04 | 46.8 | — | 47.9 | |
| Mix-and-LocalizeMethod Category=Current SOTA Based Methods, Pre-trained Encoders=AudioCLIP2024.04 | 45.6 | 41.8 | 47.4 | |
| EZ-VSLMethod Category=Current SOTA Based Methods, Pre-trained Encoders=AudioCLIP2024.04 | 45.1 | 40.6 | 47.2 | |
| LVSMethod Category=Current SOTA Based Methods, Pre-trained Encoders=AudioCLIP2024.04 | 44.6 | 39.7 | 45.3 | |
| Wav2CLIPMethod Category=CLIP-Based Baseline Methods, Pre-trained Encoders=AudioCLIP2024.04 | 43.6 | 39.3 | 44.7 | |
| OTSMethod Category=Current SOTA Based Methods, Pre-trained Encoders=AudioCLIP2024.04 | 43.2 | 38.9 | 43.7 | |
| CoarsetoFineMethod Category=Current SOTA Based Methods, Pre-trained Encoders=AudioCLIP2024.04 | 43.1 | 39.4 | 44.8 | |
| Proposed MethodVisual Backbone=ResNet-18, Audio Backbone=ResNet-182024.03 | 41.4 | 46.2 | 41.2 | |
| NoPrior2026.04 | 41.4 | 46.2 | 41.2 | |
| AVGN2024.03 | 40.8 | 35.3 | 42.3 | |
| AVGN2026.04 | 40.8 | 33.2 | 42.3 | |
| Qwen2.5-Omni2026.04 | 39.4 | 43.6 | 41.8 | |
| EZ-VSL2024.03 | 38.9 | 31.3 | 39.5 | |
| EZ-VSL2026.04 | 38.9 | 31.3 | 39.5 | |
| Mix-and-Localize2024.03 | 36.3 | 32.5 | 38.9 | |
| Mix-and-Localize2026.04 | 36.3 | 32.5 | 38.9 | |
| DSOL2024.03 | 35.7 | — | 37.2 | |
| DSOL2026.04 | 35.7 | — | 37.2 | |
| LVS2024.03 | 34.4 | 29.6 | 38.2 | |
| LVS2026.04 | 34.4 | 29.6 | 38.2 | |
| OTS2024.03 | 32.8 | 29.8 | 35.7 | |
| OTS2026.04 | 32.8 | 29.8 | 35.7 | |
| CoarseToFine2024.03 | 29.1 | 28.2 | 34.8 | |
| CoarseToFIne2026.04 | 29.1 | 28.2 | 34.8 | |
| MiniCPM-o2026.04 | 24.9 | 40.9 | 32.1 | |
| DMC2024.03 | 23.9 | — | 27.6 | |
| DMC2026.04 | 23.9 | — | 27.6 | |
| InteractiveOmni2026.04 | 21 | 36.4 | 16.4 | |
| Attention10k2024.03 | 19.2 | — | 30.6 | |
| Attention 10k2026.04 | 19.2 | — | 30.6 |