mlsanityData to 7 Oct 2026

Tracing Copied Pixels and Regularizing Patch Affinity in Copy Detection

arXiv 2602.17484 ↗First posted 19 Feb 2026cs.CV cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

Image Copy Detection (ICD) aims to identify manipulated content between image pairs through robust feature representation learning. While self-supervised learning (SSL) has advanced ICD systems, existing view-level contrastive methods struggle with sophisticated edits due to insufficient fine-grained correspondence learning. We address this limitation by exploiting the inherent geometric traceability in edited content through two key innovations. First, we propose PixTrace - a pixel coordinate tracking module that maintains explicit spatial mappings across editing transformations. Second, we introduce CopyNCE, a geometrically-guided contrastive loss that regularizes patch affinity using overlap ratios derived from PixTrace's verified mappings. Our method bridges pixel-level traceability with patch-level similarity learning, suppressing supervision noise in SSL training. Extensive experiments demonstrate not only state-of-the-art performance (88.7% uAP / 83.9% RP90 for matcher, 72.6% uAP / 68.4% RP90 for descriptor on DISC21 dataset) but also better interpretability over existing methods.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 51 of its 63 references that are in the map. First posted on arXiv 19 Feb 2026.

Builds on

51 of 63 references.

References · 51

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  3. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  4. 2018Representation Learning with Contrastive Predictive Codingcited by 7,662
  5. 2019Momentum Contrast for Unsupervised Visual Representation Learningcited by 7,482
  6. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,893
  7. 2015Fast R-CNNcited by 2,930
  8. 2019SuperGlue: Learning Feature Matching with Graph Neural Networkscited by 1,437
  9. 2017SuperPoint: Self-Supervised Interest Point Detection and Descriptioncited by 1,313
  10. 2021LoFTR: Detector-Free Local Feature Matching with Transformerscited by 974
  11. 2023LightGlue: Local Feature Matching at Light Speedcited by 583
  12. 2020Dense Contrastive Learning for Self-Supervised Visual Pre-Trainingcited by 512
38 more
  1. 2018Two at Once: Enhancing Learning and Generalization Capacities via IBN-Netcited by 361
  2. 2021Unsupervised Semantic Segmentation by Contrasting Object Mask Proposalscited by 190
  3. 2022A Self-Supervised Descriptor for Image Copy Detectioncited by 175
  4. 2022ASpanFormer: Detector-Free Image Matching with Adaptive Span Transformercited by 170
  5. 2021Efficient Visual Pretraining with Contrastive Detectioncited by 150
  6. 2021Efficient Self-supervised Vision Transformers for Representation Learningcited by 149
  7. 2016Deep Image Homography Estimationcited by 135
  8. 2022VICRegL: Self-Supervised Learning of Local Visual Featurescited by 118
  9. 2018Spreading vectors for similarity searchcited by 114
  10. 2020Learning Feature Descriptors using Camera Pose Supervisioncited by 104
  11. 2021Instance Localization for Self-supervised Detection Pretrainingcited by 99
  12. 2021Region Similarity Representation Learningcited by 96
  13. 2022Self-Supervised Learning of Object Parts for Semantic Segmentationcited by 90
  14. 2022Self-Supervised Visual Representation Learning with Semantic Groupingcited by 70
  15. 2021Spatially Consistent Representation Learningcited by 64
  16. 2022Patch-level Representation Learning for Self-supervised Vision Transformerscited by 60
  17. 2024OmniGlue: Generalizable Feature Matching with Foundation Model Guidancecited by 55
  18. 2020Learning and aggregating deep local descriptors for instance-level recognitioncited by 54
  19. 2021The 2021 Image Similarity Dataset and Challengecited by 44
  20. 2022Point-Level Region Contrast for Object Detection Pre-Trainingcited by 41
  21. 2021Revitalizing CNN Attentions via Transformers in Self-Supervised Visual Representation Learningcited by 34
  22. 2022Learning Where to Learn in Cross-View Self-Supervised Learningcited by 31
  23. 2021Exploring Set Similarity for Dense Self-supervised Representation Learningcited by 29
  24. 2022CP2: Copy-Paste Contrastive Pretraining for Semantic Segmentationcited by 26
  25. 2021Dense Semantic Contrast for Self-Supervised Visual Representation Learningcited by 24
  26. 2021Contrastive Learning with Large Memory Bank and Negative Embedding Subtraction for Accurate Copy Detectioncited by 23
  27. 2022TransVCL: Attention-enhanced Video Copy Localization Network with Flexible Supervisioncited by 15
  28. 2021Bag of Tricks and A Strong baseline for Image Copy Detectioncited by 14
  29. 2021Video Similarity and Alignment Learning on Partial Video Copy Detectioncited by 12
  30. 2023Self-Supervised Video Similarity Learningcited by 12
  31. 2022Results and findings of the 2021 Image Similarity Challengecited by 10
  32. 2022A Large-scale Comprehensive Dataset and Copy-overlap Aware Evaluation Protocol for Segment-level Video Copy Detectioncited by 10
  33. 2022A Benchmark and Asymmetrical-Similarity Learning for Practical Image Copy Detectioncited by 9
  34. 2023The 2023 Video Similarity Dataset and Challengecited by 9
  35. 2021D2LV: A Data-Driven and Local-Verification Approach for Image Copy Detectioncited by 7
  36. 2021Producing augmentation-invariant embeddings from real-life imagerycited by 6
  37. 2024AnyPattern: Towards In-context Image Copy Detectioncited by 5
  38. 20213rd Place: A Global and Local Dual Retrieval Solution to Facebook AI Image Similarity Challengecited by 4

50 most cited of 51 All 51 (JSON)

Built on by

Not cited yet.