AI Paper Atlas

Research map / Video / Video-language understanding

Learning Audio-Video Modalities from Image Captions

Arsha Nagrani, Paul Hongsuck Seo, Bryan Seybold, Anja Hauth, Santiago Manen, Chen Sun, Cordelia Schmid

ECCV 2022 · Video › Video-language understanding · approach: Text-video retrieval · 82 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
pretrain on VideoCC3M instead of HowTo100MRGB retrieval model · MSR-VTT finetunedR@1: 33.1 -> 35.0Improved
pretrain on VideoCC3M instead of HowTo100MRGB retrieval model · MSR-VTT zero-shotR@1: 8.6 -> 18.9Improved
add audio to VideoCC3M-pretrained model (ablation)retrieval model · MSR-VTT finetunedR@1: 35.0 -> 35.8Improved
add audio to VideoCC3M-pretrained model (ablation)retrieval model · MSR-VTT zero-shotR@1: 18.9 -> 19.4Improved
add VideoCC3M pretraining to RGB modelretrieval model · MSR-VTT finetunedR@1: 30.2 -> 35.0Improved
initialise video encoder from K400 instead of ImageNet21K (ablation)RGB video encoder · MSR-VTTR@1: 30.2 -> 30.2No clear effect
initialise video encoder from ImageNet21K instead of scratch (ablation)RGB video encoder · MSR-VTTR@1: 9.4 -> 30.2Improved
add audio stream to K400-initialised video model (ablation)retrieval model · MSR-VTTR@1: 30.2 -> 32.3Improved
increase mined clip length t from 3s to 10s (ablation)VideoCC3M mining · MSR-VTT zero-shotR@1: 16.4 -> 18.9Improved
increase frame stride from 2 to 14 (longer clip coverage) (ablation)RGB retrieval model · MSR-VTTR@1: 24.1 -> 27.3Improved
initialise audio encoder from VGGSound instead of ImageNet21K (ablation)audio retrieval model · AudioCapsR@1: 30.2 -> 32.0Improved
pretrain audio model on VideoCC3Maudio retrieval model · AudioCaps (audio only)R@1: 32.0 -> 35.5Improved

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