AI Paper Atlas

Research map / Video / Video-language understanding

Frozen in Time: A Joint Video and Image Encoder for End-to-End Retrieval

Max Bain, Arsha Nagrani, Gül Varol, Andrew Zisserman

ICCV 2021 · Video › Video-language understanding · approach: Text-video retrieval · 961 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 large-scale pretraining on CC3M+WebVid2M versus no pretraining (ablation)space-time transformer · MSR-VTT 1K-AR@1: 5.6 -> 27.3Improved
pretrain on WebVid2M instead of HowTo-17M subset (ablation)space-time transformer · MSR-VTT 1K-AR@1: 24.1 -> 26.0Improved
add image data (CC3M) to WebVid2M video pretraining (ablation)space-time transformer · MSR-VTT 1K-AR@1: 26.0 -> 27.3Improved
temporal expansion at pretraining rather than at finetuning (ablation)space-time transformer · MSR-VTT 1K-AR@1: 24.9 -> 26.0Improved
add 1 to 4 frame curriculum at pretraining (ablation)space-time transformer · MSR-VTT 1K-AR@1: 26.0 -> 26.6Improved
add 1 to 4 to 8 frame curriculum instead of 8 frames from start (ablation)space-time transformer · MSR-VTT 1K-AR@1: 25.4 -> 27.4Improved
use 4 frames instead of a single frame (ablation)space-time transformer · MSR-VTT 1K-AR@1: 18.8 -> 24.9Improved
replace ResNet-101 with space-time transformer backbone (ablation)video encoder · MSR-VTT testR@1: 11.5 -> 26.8Improved
use distilbert instead of t5-base text backbone (ablation)text encoder · MSR-VTT testR@1: 24.0 -> 26.8Improved
use bert-base instead of distilbert text backbone (ablation)text encoder · MSR-VTT testR@1: 26.8 -> 27.5No clear effect
use the modified space-time attention block instead of divided block (ablation)space-time attention · MSR-VTT 1K-A (zero-shot)R@1: 13.0 -> 14.6Improved
choice of temporal embedding expansion (zero-pad vs bilinear) (ablation)temporal expansion 4 to 8 frames · MSR-VTT 1K-AR@1: 30.7 vs 28.3No 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