AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

TAViS: Text-bridged Audio-Visual Segmentation with Foundation Models

Ziyang Luo, Nian Liu, Xuguang Yang, Salman Khan, Rao Muhammad Anwer, Hisham Cholakkal, Fahad Shahbaz Khan, Junwei Han

ICCV 2025 · Vision-language and multimodal › Cross-modal learning · approach: Audio-visual source localization · 6 citations (OpenAlex)

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 changedWhereMeasuredEffect
add ImageBind-guided query decomposition (ablation)TAViS · MS3mIoU (M_J): 58.5 -> 68.2Improved
add ImageBind-guided query decomposition (ablation)TAViS · S4mIoU (M_J): 79.6 -> 84.8Improved
add text-bridged alignment supervision (ablation)TAViS · MS3mIoU (M_J): 64.9 -> 68.2Improved
add text-bridged hybrid prompting (ablation)TAViS · MS3mIoU (M_J): 65.1 -> 68.2Improved
add audio-to-text loss (ablation)TAViS · MS3mIoU (M_J): 65.2 -> 68.2Improved
add image-to-text loss (ablation)TAViS · MS3mIoU (M_J): 64.4 -> 68.2Improved
add direct audio-image alignment loss L_a2i (ablation)TAViS · MS3mIoU (M_J): 68.2 -> 66.5Hurt
use only L_a2i instead of text-bridged losses (ablation)TAViS · MS3mIoU (M_J): 68.2 -> 66.5Hurt
add pseudo-text sparse prompt p^t (ablation)TAViS · MS3mIoU (M_J): 63.4 -> 68.2Improved
use ImageBind text encoder instead of MLP prediction of pseudo-text (ablation)TAViS · MS3mIoU (M_J): 63.6 -> 68.2Improved
add image cls dense prompt t_v (ablation)TAViS · MS3mIoU (M_J): 63.6 -> 68.2Improved
increase image size from 224 to 1024TAViS · MS3mIoU (M_J): 68.2 -> 71.2Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy