Research map / Video / Video-language understanding
Single-Stage Visual Query Localization in Egocentric Videos
Hanwen Jiang, Santhosh Kumar Ramakrishnan, Kristen Grauman
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 changed | Where | Measured | Effect |
|---|---|---|---|
| lower input resolution 448 -> 224 (ablation) | VQLoC · Ego4D VQ2D val | tAP25: 0.29 -> 0.22 | Hurt |
| smaller backbone ViT-b14 -> ViT-s14 (ablation) | VQLoC · Ego4D VQ2D val | tAP25: 0.29 -> 0.28 | No clear effect |
| replace spatial transformer with conv fusion (ablation) | VQLoC · Ego4D VQ2D val | stAP25: 0.19 -> 0.13 | Hurt |
| use 3 spatial TX layers instead of 1 (ablation) | VQLoC · Ego4D VQ2D val | stAP25: 0.19 -> 0.20 | No clear effect |
| use global temporal window instead of local window 5 (ablation) | VQLoC · Ego4D VQ2D val | tAP25: 0.29 -> 0.08 | Hurt |
| local window size 9 instead of 5 (ablation) | VQLoC · Ego4D VQ2D val | tAP25: 0.29 -> 0.27 | Hurt |
| local window size 7 instead of 5 (ablation) | VQLoC · Ego4D VQ2D val | rec%: 38.58 -> 36.97 | Hurt |
| remove anchors, keep BCE loss (ablation) | VQLoC · Ego4D VQ2D val | tAP25: 0.29 -> 0.17 | Hurt |
| anchors with BCE loss instead of focal loss (ablation) | VQLoC · Ego4D VQ2D val | tAP25: 0.29 -> 0.12 | Hurt |
| hard negative mining (BCE+HNM) instead of focal loss (ablation) | VQLoC · Ego4D VQ2D val | stAP25: 0.19 -> 0.22 | Improved |
| single-stage VQLoC vs CocoFormer, medium objects | VQLoC · Ego4D VQ2D val medium objects | tAP25: 0.206 -> 0.213 | Improved |
| single-stage VQLoC vs CocoFormer, large objects | VQLoC · Ego4D VQ2D val large objects | tAP25: 0.338 -> 0.454 | Improved |
Closest papers by meaning
- PRVQL: Progressive Knowledge-guided Refinement for Robust Egocentric Visual Query Localization (ICCV 2025 · similarity 0.94)
- EgoLoc: Revisiting 3D Object Localization from Egocentric Videos with Visual Queries (ICCV 2023 · similarity 0.94)
- RELOCATE: A Simple Training-Free Baseline for Visual Query Localization Using Region-Based Representations (CVPR 2025 · similarity 0.91)
- EgoThinker: Unveiling Egocentric Reasoning with Spatio-Temporal CoT (NeurIPS 2025 · similarity 0.87)
- QDETRv: Query-Guided DETR for One-Shot Object Localization in Videos (AAAI 2024 · similarity 0.86)
- GPVK-VL: Geometry-Preserving Virtual Keyframes for Visual Localization under Large Viewpoint Changes (CVPR 2025 · similarity 0.86)
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.
