Style or Signature? Artist-Disjoint Evaluation of Style Classification in Frozen Vision Embeddings

• URL: http://arxiv.org/abs/2608.14435v1
• Date: Fri, 14 Aug 2026 16:18:11 GMT
• Title: Style or Signature? Artist-Disjoint Evaluation of Style Classification in Frozen Vision Embeddings
• Abstract summary: We ask whether accuracy reflects an understanding of style or the recognition of individual artists.<n>We re-evaluate style classification under an artist-disjoint protocol.<n> Impressionism and Cubism barely move, while Surrealism falls twenty points.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Frozen image embeddings from models such as CLIP are increasingly used to classify paintings by art-historical style, with high reported accuracy. We ask whether this accuracy reflects an understanding of style or the recognition of individual artists. Standard evaluation uses random splits in which works by the same artist appear on both sides, so a classifier can succeed by recognising the painter rather than the movement. We re-evaluate style classification under an artist-disjoint protocol, holding out every artist in turn so that no work is ever classified using other works by its own painter. On a balanced dataset of 320 paintings across four twentieth-century movements, 5-NN style accuracy falls from 0.87 to 0.77 under this protocol, and the drop is sharply uneven. Impressionism and Cubism barely move, while Surrealism falls twenty points. The pattern holds across four image encoders, including a vision-only self-supervised model, which places the effect in visual structure rather than language. Where an encoder captures genuine shared form, individual artists are barely recognisable yet style is robust, while Surrealism shows the opposite. We argue that artist-disjoint evaluation is necessary to measure stylistic understanding in frozen embeddings.

Related papers

• Does AI See like Art Historians? Interpreting How Vision Language Models Recognize Artistic Style [60.562013130566726]
  We characterize the mechanisms underlying VLMs' ability to predict artistic style.<n>We employ a latent-space decomposition approach to identify concepts that drive art style prediction.<n>In cases where an irrelevant concept was used to successfully predict style, art historians identified possible reasons for its success.
  arXiv  Detail & Related papers  (2026-03-11T17:49:45Z)
• Identifying Prompted Artist Names from Generated Images [59.34482128911978]
  A common and controversial use of text-to-image models is to generate pictures by explicitly naming artists.<n>We introduce a benchmark for prompted-artist recognition.<n>The dataset contains 1.95M images covering 110 artists.
  arXiv  Detail & Related papers  (2025-07-24T17:59:44Z)
• ArtistAuditor: Auditing Artist Style Pirate in Text-to-Image Generation Models [61.55816738318699]
  We propose a novel method for data-use auditing in the text-to-image generation model.<n>ArtistAuditor employs a style extractor to obtain the multi-granularity style representations and treats artworks as samplings of an artist's style.<n>The experimental results on six combinations of models and datasets show that ArtistAuditor can achieve high AUC values.
  arXiv  Detail & Related papers  (2025-04-17T16:15:38Z)
• Learning Artistic Signatures: Symmetry Discovery and Style Transfer [8.288443063900825]
  There is no undisputed definition of artistic style.<n>Style should be thought of as a set of global symmetries that dictate the arrangement of local textures.<n>We show that by considering both local and global features, using both Lie generators and traditional measures of texture, we can quantitatively capture the stylistic similarity between artists better than with either set of features alone.
  arXiv  Detail & Related papers  (2024-12-05T18:56:23Z)
• FedStyle: Style-Based Federated Learning Crowdsourcing Framework for Art Commissions [3.1676484382068315]
  FedStyle is a style-based federated learning crowdsourcing framework. It allows artists to train local style models and share model parameters rather than artworks for collaboration. It addresses extreme data heterogeneity by having artists learn their abstract style representations and align with the server.
  arXiv  Detail & Related papers  (2024-04-25T04:53:43Z)
• Rethinking Artistic Copyright Infringements in the Era of Text-to-Image Generative Models [47.19481598385283]
  ArtSavant is a tool to determine the unique style of an artist by comparing it to a reference dataset of works from WikiArt. We then perform a large-scale empirical study to provide quantitative insight on the prevalence of artistic style copying across 3 popular text-to-image generative models.
  arXiv  Detail & Related papers  (2024-04-11T17:59:43Z)
• Learning to Evaluate the Artness of AI-generated Images [64.48229009396186]
  ArtScore is a metric designed to evaluate the degree to which an image resembles authentic artworks by artists. We employ pre-trained models for photo and artwork generation, resulting in a series of mixed models. This dataset is then employed to train a neural network that learns to estimate quantized artness levels of arbitrary images.
  arXiv  Detail & Related papers  (2023-05-08T17:58:27Z)
• Glaze: Protecting Artists from Style Mimicry by Text-to-Image Models [38.92567577109414]
  Glaze is a tool that enables artists to apply "style cloaks" to their art before sharing online.<n>These cloaks apply barely perceptible perturbations to images, and when used as training data, mislead generative models that try to mimic a specific artist.
  arXiv  Detail & Related papers  (2023-02-08T17:45:23Z)
• Inversion-Based Style Transfer with Diffusion Models [78.93863016223858]
  Previous arbitrary example-guided artistic image generation methods often fail to control shape changes or convey elements. We propose an inversion-based style transfer method (InST), which can efficiently and accurately learn the key information of an image.
  arXiv  Detail & Related papers  (2022-11-23T18:44:25Z)
• Quantifying Confounding Bias in Generative Art: A Case Study [3.198144010381572]
  We propose a simple metric to quantify confounding bias due to the lack of modeling the influence of art movements in learning artists' styles. The proposed metric is more effective than state-of-the-art outlier detection method in understanding the influence of art movements in artworks.
  arXiv  Detail & Related papers  (2021-02-23T21:59:30Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.