AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Image Retrieval on Real-Life Images With Pre-Trained Vision-and-Language Models

Zheyuan Liu, Cristian Rodriguez-Opazo, Damien Teney, Stephen Gould

ICCV 2021 · Vision-language and multimodal › Human and embodied modeling · approach: Composed and text-based image retrieval · 273 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 OSCAR vision-and-language pre-trained initialization (ablation)CIRPLANT · CIRR testRecall@1: 15.18 -> 19.55Improved
add OSCAR vision-and-language pre-trained initialization (ablation)CIRPLANT · CIRR testRecall@5: 43.36 -> 52.55Improved
add OSCAR vision-and-language pre-trained initialization (ablation)CIRPLANT · CIRR testRecall_Subset@1: 33.81 -> 39.20Improved
add OSCAR vision-and-language pre-trained initialization (ablation)CIRPLANT · Fashion-IQ(R@10+R@50)/2: 25.17 -> 30.20Improved
use transformer composition (CIRPLANT without init) instead of TIRG gatingTIRG · CIRR testRecall_Subset@1: 22.67 -> 33.81Improved
use transformer composition (CIRPLANT without init) instead of TIRG gatingTIRG · CIRR testRecall@1: 14.61 -> 15.18No clear effect
add BERT pre-trained word representations (ablation)MAAF · CIRR test(R@5+R_Subset@1)/2: 27.04 -> 27.57No clear effect
remove output of text tokens in final pooling (MAAF-IT) (ablation)MAAF · CIRR test(R@5+R_Subset@1)/2: 27.04 -> 27.02No clear effect
replace final resolution-wise pooling with average pooling (MAAF-RP) (ablation)MAAF · CIRR test(R@5+R_Subset@1)/2: 27.04 -> 27.37No clear effect
inject text at last convolution layer instead of last FC layer (LastConv) (ablation)TIRG · CIRR test(R@5+R_Subset@1)/2: 35.52 -> 29.75Hurt
compose image and text with TIRG instead of using image onlyImage-only baseline · CIRR testRecall@5: 48.46 -> 48.37No clear effect
use text only instead of image-text composition (Recall_Subset)Image+Text Concatenation · CIRR testRecall_Subset@1: 23.74 -> 39.69Mixed

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