AI Paper Atlas

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

Dual Attention Matching for Audio-Visual Event Localization

Yu Wu, Linchao Zhu, Yan Yan, Yi Yang

ICCV 2019 · Vision-language and multimodal › Cross-modal learning · approach: Audio-visual source localization · 221 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 dual attention matching (DAM) to the localization model (ablation)supervised AV event localization framework · AVE supervised event localizationaccuracy: 70.7 -> 74.5Improved
use cross-modal matching instead of self-matching (ablation)DAM · AVE cross-modality localizationaverage accuracy: 29.2 -> 47.8Improved
use cross-modal matching instead of self-matching (ablation)DAM · AVE supervised event localizationaccuracy: 74.2 -> 74.5Improved
add self-matching (global feature of same modality) over no matching (ablation)supervised localization framework · AVE supervised event localizationaccuracy: 70.7 -> 74.2Improved
use self-attention embedding instead of RNN (ablation)DAM · AVE cross-modality localizationaverage accuracy: 44.9 -> 47.8Improved
use self-attention embedding instead of averaged pooling (ablation)DAM · AVE cross-modality localizationaverage accuracy: 46.1 -> 47.8Improved
use self-attention embedding instead of max pooling (ablation)DAM · AVE cross-modality localizationaverage accuracy: 46.0 -> 47.8Improved
use self-attention embedding instead of LSTM (ablation)DAM · AVE cross-modality localizationaverage accuracy: 45.8 -> 47.8Improved
use self-attention embedding instead of GRU (ablation)DAM · AVE cross-modality localizationaverage accuracy: 46.5 -> 47.8Improved
use self-attention embedding instead of bidirectional LSTM (ablation)DAM · AVE cross-modality localizationaverage accuracy: 46.2 -> 47.8Improved
use DAM with non-parametric pooling embeddings versus prior local-only methodDAM with pooling · AVE cross-modality localizationaverage accuracy: 40.2 -> 46.1Improved
balance event relevance loss with category loss at lambda 0.5 versus category loss only (ablation)supervised localization framework · AVE supervised event localizationaccuracy: 70.7 -> 74.5Improved

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