AI Paper Atlas

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

Audio—Visual Segmentation

Jinxing Zhou, Jianyuan Wang, Jiayi Zhang, Weixuan Sun, Jing Zhang, Stan Birchfield, Dan Guo, Lingpeng Kong, Meng Wang, Yiran Zhong

ECCV 2022 · Vision-language and multimodal › Cross-modal learning · approach: Audio-visual source localization · 168 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 TPAVI module (ablation)AVS (PVT-v2) · AVSBench S4M_J: .778 -> .787Improved
add TPAVI module (ablation)AVS (ResNet50) · AVSBench S4M_J: .701 -> .728Improved
add TPAVI module (ablation)AVS (PVT-v2) · AVSBench MS3M_J: .482 -> .531Improved
add TPAVI instead of simple A+V feature addition (ablation)AVS (PVT-v2) · AVSBench S4M_J: .777 -> .787Improved
add L_AVM-AV loss to BCE (ablation)AVS (ResNet50) · AVSBench MS3M_J: .466 -> .479Improved
add L_AVM-AV loss to BCE (ablation)AVS (PVT-v2) · AVSBench MS3M_F: .626 -> .645Improved
add L_AVM-VV loss to BCE (ablation)AVS (ResNet50) · AVSBench MS3M_J: .466 -> .467No clear effect
apply TPAVI at all four stages instead of only stage 4 (ablation)AVS (PVT-v2) · AVSBench MS3M_J: 0.490 to 0.540Improved
apply TPAVI at stage 4 only instead of stage 3 only (ablation)AVS (PVT-v2) · AVSBench S4M_J: .780 -> .777Hurt
pretrain on Single-source subset (ablation)AVS (PVT-v2) with TPAVI · AVSBench MS3M_J: 0.540 to 0.573Improved
pretrain on Single-source subset (ablation)AVS (ResNet50) with TPAVI · AVSBench MS3M_J: 0.479 to 0.543Improved
pretrain on Single-source subset (ablation)AVS (PVT-v2) without TPAVI · AVSBench MS3M_J: 0.482 to 0.506Improved

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