{"slug": "representation-and-reference-selection-in-training-free-synthetic-image", "title": "Representation and Reference Selection in Training-Free Synthetic Image Attribution", "summary": "Researchers at arXiv studied training-free synthetic image attribution, finding that attribution accuracy peaks at intermediate representation levels in CLIP and DINOv2, and that semantically constrained reference selection improves performance, especially with limited references.", "body_md": "arXiv:2607.12052v1 Announce Type: new\nAbstract: Synthetic image attribution aims at identifying the generator responsible for a given AI-generated image. Training-free reference-based attribution methods are easily scalable, since newly emerging generators can be incorporated by adding source-specific references rather than retraining a task-specific classifier. Their performance depends on two coupled factors: the representation space used for comparison and the way source-specific references are constructed. However, the interaction between these two factors remains largely unexplored. In this paper, we provide a controlled analysis of this interaction using references and off-the-shelf pretrained representations. We study representations extracted from different layers of CLIP and DINOv2, along with three reference selection methods with varying semantic constraints: arbitrary, semantically aligned, and resynthesis-based references. Our results show that attribution accuracy consistently peaks at intermediate representation levels, indicating that source-discriminative cues are more accessible before strong semantic abstraction dominates. We further show that intermediate representations are not completely semantically neutral, making reference selection critical: semantically constrained references reduce query-reference mismatch and improve attribution, especially under limited reference budgets. Resynthesis is most useful in low-reference regimes, while semantically aligned references provide a better accuracy-cost trade-off when a moderate-sized reference pool is available. Our findings show that training-free reference-based attribution should be understood as the interaction between where images are compared, how the reference set is constructed, and how many references are available.", "url": "https://wpnews.pro/news/representation-and-reference-selection-in-training-free-synthetic-image", "canonical_source": "https://arxiv.org/abs/2607.12052", "published_at": "2026-07-15 04:00:00+00:00", "updated_at": "2026-07-15 04:01:21.192841+00:00", "lang": "en", "topics": ["artificial-intelligence", "computer-vision", "ai-research"], "entities": ["arXiv", "CLIP", "DINOv2"], "alternates": {"html": "https://wpnews.pro/news/representation-and-reference-selection-in-training-free-synthetic-image", "markdown": "https://wpnews.pro/news/representation-and-reference-selection-in-training-free-synthetic-image.md", "text": "https://wpnews.pro/news/representation-and-reference-selection-in-training-free-synthetic-image.txt", "jsonld": "https://wpnews.pro/news/representation-and-reference-selection-in-training-free-synthetic-image.jsonld"}}