AI Paper Atlas

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

Audio-Visual Event Localization in Unconstrained Videos

Yapeng Tian, Jing Shi, Bochen Li, Zhiyao Duan, Chenliang Xu

ECCV 2018 · Vision-language and multimodal › Cross-modal learning · approach: Self-supervised audio-visual representation learning · 541 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 audio features to visual-only model (ablation)visual LSTM model · AVE event localizationaccuracy (%): 55.3 -> 71.4Improved
add visual features to audio-only model (ablation)audio LSTM model · AVE event localizationaccuracy (%): 59.5 -> 71.4Improved
add audio-guided visual attention to visual model (ablation)V model · AVE event localizationaccuracy (%): 55.3 -> 58.6Improved
add audio-guided visual attention to audio-visual model (ablation)A+V model · AVE event localizationaccuracy (%): 71.4 -> 72.7Improved
add audio-guided visual attention to weakly supervised audio-visual model (ablation)W-A+V · AVE event localization (weakly supervised)accuracy (%): 63.7 -> 66.7Improved
use fusion after LSTM (late) instead of before (early) (ablation)Concat fusion · AVE event localizationaccuracy (%): 61.0 -> 72.7Improved
use late fusion instead of decision fusion (ablation)DMRN · AVE event localizationaccuracy (%): 70.4 -> 73.1Improved
use DMRN instead of Concat in late fusionfusion network · AVE event localizationaccuracy (%): 72.7 -> 73.1Improved
use DMRN instead of MRN in late fusionfusion network · AVE event localizationaccuracy (%): 70.8 -> 73.1Improved
use DMRFE ensemble instead of DMRNDMRN · AVE event localizationaccuracy (%): 73.1 -> 73.3Improved
stack multiple residual blocks in fusion network (ablation)MRN and DMRN · AVE event localizationaccuracy (%)No clear effect
add C3D temporal visual features (ablation)visual features · AVE event localizationaccuracy (%)No clear effect

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