AI Paper Atlas

Research map / Video / Video-language understanding

A Large-Scale Comprehensive Dataset and Copy-Overlap Aware Evaluation Protocol for Segment-Level Video Copy Detection

Sifeng He, Xudong Yang, Chen Jiang, Gang Liang, Wei Zhang, Tan Pan, Qing Wang, Furong Xu, Chunguang Li, JinXiong Liu, Hui Xu, Kaiming Huang and 4 more

CVPR 2022 · Video › Video-language understanding · approach: Long video understanding · 21 citations (OpenAlex)

Check 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
Use DINO features with SPD rather than the other tested feature-alignment combinationsSegment-level video copy localization on VCSL positive pairs · VCSL positive pairs (macro over query sets)F-score: 87.40Improved
Use DINO features with TN rather than other tested combinationsVideo copy detection on VCSL positive and negative pairs · VCSL testRecall: 77.61Improved
Use ViSiL features with TNVideo copy detection on VCSL positive and negative pairs · VCSL testF-score: 61.46Improved
Train SPD on VCSL instead of VCDB (ablation)DINO + SPD · VCSL positive pairs (macro over query sets)F-score: 85.69 -> 87.40Improved
Use SPD for segment localizationVideo copy detection on VCSL · VCSL test (positive and negative pairs)F-score: 60.70; FRR 0.2726; FAR 0.1732Mixed
Use different frame feature extractorsSegment-level video copy detection benchmark · VCSLF-score: 48.04; 51.26; 47.86; 49.96 (with HV)Mixed

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