Research map / Video / Video generation
AlignSep: Temporally-Aligned Video-Queried Sound Separation with Flow Matching
Xize Cheng, Chenyuhao Wen, Slytherin Wang, Yongqi Wang, Zehan Wang, Rongjie Huang, Tao Jin, Zhou Zhao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add flow matching instead of diffusion (ablation) | AlignSep · VGGSound-Hard | TA-V (alignment acc): 92.28 -> 95.76 | Improved |
| add flow matching instead of diffusion (ablation) | AlignSep · VGGSound-Clean | SA-A (CLAP): 64.12 -> 73.38 | Improved |
| add CAVP temporal visual encoder (ablation) | AlignSep · VGGSound-Hard | TA-V (alignment acc): 76.27 -> 95.76 | Improved |
| add CAVP temporal visual encoder (ablation) | AlignSep · VGGSound-Clean | TA-V (alignment acc): 94.71 -> 96.88 | Improved |
| use concat fusion instead of cross-attention (ablation) | AlignSep · VGGSound-Hard | TA-V (alignment acc): 73.38 -> 95.76 | Improved |
| use concat fusion instead of cross-attention (ablation) | AlignSep · VGGSound-Clean | SA-A (CLAP): 70.13 -> 73.38 | Improved |
| increase denoising steps from 5 to 25 (ablation) | AlignSep · VGGSound-Hard | TA-V (alignment acc): 88.14 -> 95.76 | Improved |
| increase denoising steps from 25 to 100 (ablation) | AlignSep · VGGSound-Hard | TA-V (alignment acc): 95.76 -> 93.22 | No clear effect |
| increase denoising steps from 25 to 100 (ablation) | AlignSep · VGGSound-Hard | inference throughput (FPS): 2.17 -> 0.72 | Hurt |
| use flow matching instead of rectified flow (100 steps) (ablation) | AlignSep · VGGSound-Clean | SA-A (CLAP): 57.36 -> 73.64 | Improved |
| increase reference video frame rate from 0.25 to 4 FPS (ablation) | AlignSep · VGGSound-Hard | alignment accuracy: 0.76 -> 0.95 | Improved |
| increase reference video frame rate (semantic-only baseline) (ablation) | CLIPSep · VGGSound-Hard | alignment accuracy: 0.81 | No clear effect |
Closest papers by meaning
- Seeing Speech and Sound: Distinguishing and Locating Audio Sources in Visual Scenes (CVPR 2025 · similarity 0.88)
- Foley-Flow: Coordinated Video-to-Audio Generation with Masked Audio-Visual Alignment and Dynamic Conditional Flows (CVPR 2025 · similarity 0.87)
- OmniSep: Unified Omni-Modality Sound Separation with Query-Mixup (ICLR 2025 · similarity 0.87)
- Native Audio-Visual Alignment for Generation (NeurIPS 2026 · similarity 0.87)
- Seeing and Hearing: Open-domain Visual-Audio Generation with Diffusion Latent Aligners (CVPR 2024 · similarity 0.87)
- Diverse and Aligned Audio-to-Video Generation via Text-to-Video Model Adaptation (AAAI 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
