Research map / Image generation / Autoregressive and latent generation
A Benchmark and Asymmetrical-Similarity Learning for Practical Image Copy Detection
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Add ASL to the strong baseline | Strong baseline · NDEC | μAP: +5.22% | Helped |
| Add ASL to the strong baseline | Strong baseline · DISC21 | μAP: -0.97% | Hurt |
| Add ASL post-processing | D²LV · NDEC | μAP | Helped |
| Add crop-to-copy norm-ratio learning (ablation) | Simple baseline · NDEC | precision: 45.30% to 62.98% | Helped |
| Add crop-to-copy norm-ratio learning (ablation) | Simple baseline · NDEC | false positive matches: 2,740 to 1,233 | Helped |
| Treat annotated hard-negative pairs as positive pairs for metric learning (ablation) | ASL using hard negative pairs · NDEC | μAP | Helped |
| Replace the norm-ratio item with canonical triplet loss (ablation) | ASL · NDEC | μAP | Hurt |
| Combine crop-to-copy learning with hard-negative-pair learning (ablation) | ASL · NDEC | precision: +10.86% | Helped |
| Combine crop-to-copy learning with hard-negative-pair learning (ablation) | ASL · NDEC | precision: +2.97% | Helped |
| Increase the share of hard-negative queries | Simple baseline · NDEC | μAP: 72.42% to 47.00% (-25.42%) | Hurt |
Closest papers by meaning
- Image Copy Detection for Diffusion Models (NeurIPS 2024 · similarity 0.85)
- Instance-Level Composed Image Retrieval (NeurIPS 2025 · similarity 0.84)
- Asymmetric Mutual Alignment for Unsupervised Zero-Shot Sketch-Based Image Retrieval (AAAI 2024 · similarity 0.83)
- $\mathbb{X}$-Sample Contrastive Loss: Improving Contrastive Learning with Sample Similarity Graphs (ICLR 2025 · similarity 0.83)
- On Train-Test Class Overlap and Detection for Image Retrieval (CVPR 2024 · similarity 0.82)
- QuRe: Query-Relevant Retrieval through Hard Negative Sampling in Composed Image Retrieval (ICML 2025 · similarity 0.82)
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.