AI Paper Atlas

Research map / Video / Video-language understanding

VLANet: Video-Language Alignment Network for Weakly-Supervised Video Moment Retrieval

Minuk Ma, Sunjae Yoon, Junyeong Kim, Youngjoon Lee, Sunghun Kang, Chang D. Yoo

ECCV 2020 · Video › Video-language understanding · approach: Dense video captioning · 91 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 Surrogate Proposal Selection (ablation)VLANet · Charades-STAR@1 IoU=0.5: 25.30 -> 31.82Improved
add Surrogate Proposal Selection (ablation)VLANet · Charades-STAR@5 IoU=0.5: 64.57 -> 82.82Improved
add cascaded cross-modal attention (ablation)VLANet · Charades-STAR@1 IoU=0.5: 30.08 -> 31.82Improved
add cross-modal attention (ablation)VLANet · Charades-STAR@1 IoU=0.7: 13.23 -> 14.19Improved
add self-attention (ablation)VLANet · Charades-STAR@1: 30.81 -> 31.82 (IoU=0.5); 15.38 -> 14.19 (IoU=0.7)Mixed
use stride 8 instead of stride 4 for proposal groups (ablation)VLANet · Charades-STAR@5 IoU=0.5: 63.17 -> 82.82Improved
use stride 8 instead of stride 6 for proposal groups (ablation)VLANet · Charades-STAR@5 IoU=0.5: 74.91 -> 82.82Improved
use stride 4 instead of stride 8 for proposal groups (ablation)VLANet · Charades-STAR@1 IoU=0.5: 31.82 -> 31.53No clear effect
use three scales instead of two (stride 6) (ablation)VLANet · Charades-STAR@5 IoU=0.5: 73.06 -> 74.91Improved
use VLANet instead of SCN (weakly supervised)weakly supervised VMR · Charades-STAR@1 IoU=0.5: 23.58 -> 31.83Improved
use VLANet instead of SCN on high-IoU recallweakly supervised VMR · Charades-STAR@5 IoU=0.7: 38.87 -> 33.09Hurt
use VLANet instead of TGAweakly supervised VMR · DiDeMoR@5: 39.74 -> 65.68Improved

Later papers that compared with this method

1 later accepted paper reports VLANet as a baseline in their results tables; 0 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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