AI Paper Atlas

Research map / Video / Video-language understanding

Natural Language Video Localization: A Revisit in Span-based Question Answering Framework

Hao Zhang, Aixin Sun, Wei Jing, Liangli Zhen, Joey Tianyi Zhou, Rick Siow Mong Goh

TPAMI 2021 · Video › Video-language understanding · approach: Video question answering · 93 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 query-guided highlighting (VSLNet over VSLBase) (ablation)VSLBase · Charades-STA (no fine-tune)Rank@1 IoU=0.7: 24.14 -> 30.19Improved
add query-guided highlighting (ablation)VSLBase · Charades-STA (no fine-tune)mIoU: 42.11 -> 45.15Improved
add query-guided highlighting (ablation)VSLBase · Charades-STA (fine-tuned)mIoU: 47.15 -> 50.02Improved
fine-tune the 3D ConvNet on CharadesVSLNet · Charades-STARank@1 IoU=0.7: 30.19 -> 35.22Improved
replace CAT with context-query attention (keep CMF encoder) (ablation)VSLBase · Charades-STARank@1 IoU=0.7: 27.07 -> 30.16Improved
replace BiLSTM with CMF encoder (keep CQA) (ablation)VSLBase · Charades-STARank@1 IoU=0.7: 28.55 -> 30.16Improved
add multi-scale split-and-concat (VSLNet-L-Pm) over VSLNetVSLNet · TACoS_tan (n=600)mIoU: 31.92 -> 36.61Improved
add multi-scale split-and-concat (VSLNet-L-Pm) over VSLNetVSLNet · TACoS_org (n=600)mIoU: 23.96 -> 26.40Improved
use multi-scale rather than single-scale split (ablation)VSLNet-L · TACoS_org (n=600)mIoU: 25.12 -> 26.40Improved
use multi-scale rather than single-scale split (ablation)VSLNet-L · TACoS_tan (n=600)mIoU: 33.71 -> 36.61Improved
add split-and-concat to videos shorter than 2 minutesVSLNet · TACoS_tan (n=600) 0-2 minmIoU: 46.73 -> 44.32Hurt
add split-and-concat on videos longer than 12 minVSLNet · TACoS_org (n=600) >12 minmIoU: 8.58 -> 11.90Improved

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