Research map / Vision-language and multimodal / Cross-modal learning
SOUPLE: Enhancing Audio-Visual Localization and Segmentation with Learnable Prompt Contexts
Khanh Binh Nguyen, Chae Jung Park
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 |
|---|---|---|---|
| replace fixed prompt 'a photo of a' with instance-conditional learnable context tokens | ACL-SSL · VGG-SS | cIoU: 49.46 -> 53.21 | Improved |
| learnable context tokens instead of fixed prompt | ACL-SSL · VGG-SS open-set (Unheard 110) | cIoU: 41.98 -> 48.40 | Improved |
| learnable context tokens instead of fixed prompt | ACL-SSL · AVSBench S4 | mIoU: 59.76 -> 62.89 | Improved |
| learnable context tokens instead of fixed prompt | ACL-SSL · AVSBench MS3 | mIoU: 41.08 -> 38.96 | Hurt |
| increase context tokens from 4 to 8 or 16 (ablation) | SOUPLE · VGG-SS | cIoU: 53.21 -> 52.01 (8), 51.08 (16) | Hurt |
| put [VA] at end of context instead of start (ablation) | SOUPLE · VGG-SS | cIoU: 49.91 -> 53.21 | Improved |
| fuse audio with visual features before Meta-net (ablation) | SOUPLE · VGG-SS | cIoU: 53.21 -> 49.93 | Hurt |
| ensemble prompts over [VA] positions (ablation) | SOUPLE · VGG-SS | cIoU: 53.21 -> 33.63 | Hurt |
| train for 50 epochs instead of 20 (ablation) | SOUPLE · VGG-SS | cIoU: 53.21 -> 54.76 | Improved |
| learnable context tokens instead of fixed prompt | ACL-SSL · Extended VGG-SS | AP: 40.79 -> 44.77 | Improved |
Closest papers by meaning
- Prompting Segmentation with Sound Is Generalizable Audio-Visual Source Localizer (AAAI 2024 · similarity 0.92)
- Robust Audio-Visual Segmentation via Audio-Guided Visual Convergent Alignment (CVPR 2025 · similarity 0.89)
- Object-aware Sound Source Localization via Audio-Visual Scene Understanding (CVPR 2025 · similarity 0.88)
- Separating the "Chirp" from the "Chat": Self-supervised Visual Grounding of Sound and Language (CVPR 2024 · similarity 0.88)
- From Vision to Audio and Beyond: A Unified Model for Audio-Visual Representation and Generation (ICML 2024 · similarity 0.87)
- Improving Sound Source Localization with Joint Slot Attention on Image and Audio (CVPR 2025 · 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.
