EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Vision Transformer attention alignment with human visual perception in aesthetic object evaluation

Miguel Carrasco, César González-Martín iD, José Aranda, Luis Oliveros iD

PLOS One · 03.04.2026 · 21 (4) · e0344006

DOI: 10.1371/journal.pone.0344006 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS One, online erschienen am 03.04.2026. Verfasst von Miguel Carrasco und 3 weiteren Autor:innen. Beteiligte Einrichtung: School of Computer Science and Telecommunications, University fo Diego Portales, Santiago, Chile. Hauptthemen laut MEDLINE: Attention, Visual Perception, Esthetics. Die Arbeit nennt 81 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: University of Córdoba research group Lenguajes, Reuse Reduce Recycle AI-based platform for automated and scalable Maker culture in Circular economy.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Visual attention mechanisms play a crucial role in human perception and aesthetic evaluation. Recent advances in Vision Transformers (ViTs) have demonstrated remarkable capabilities in computer vision tasks, yet their alignment with human visual attention patterns remains underexplored, particularly in aesthetic contexts. This study investigates the correlation between human visual attention and ViT attention mechanisms when evaluating handcrafted objects. We conducted an eye-tracking experiment with 30 participants (9 female, 21 male, mean age 24.6 years) who viewed 20 artisanal objects comprising basketry bags and ginger jars. Using a Pupil Labs eye-tracker, we recorded gaze patterns and generated heatmaps representing human visual attention. Simultaneously, we analyzed the same objects using a pre-trained ViT model with DINO (Self-DIstillation with NO Labels), extracting attention maps from each of the 12 attention heads. We compared human and ViT attention distributions using four complementary metrics—Kullback-Leibler divergence, Structural Similarity Index (SSIM), Pearson’s Correlation Coefficient (CC), and Similarity (SIM)—across varying Gaussian parameters ( σ = 0.1 − 4.0 ), yielding 1,152,000 distance evaluations. Additionally, we performed Areas of Interest (AOI) analysis to quantify ViT attention concentration within object regions. Statistical analysis revealed optimal correlation at σ = 2.4 ± 0.03 , with attention head #12 showing the strongest alignment with human visual patterns across all metrics. Significant differences were found between attention heads, with heads #7 and #9 demonstrating the greatest divergence from human attention ( p ≤ 0.05 ), Tukey HSD test). AOI analysis confirmed that all ViT heads concentrated attention significantly more within object regions than background areas ( p ≤ 0.0001 ), with heads #12, #1, and #3 achieving lift values of +30 to +40 percentage points. Results indicate that while ViTs exhibit more global attention patterns compared to human focal attention, certain attention heads can approximate human visual behavior, particularly for specific object features like buckles in basketry items. These findings suggest potential applications of ViT attention mechanisms in product design and aesthetic evaluation, while highlighting fundamental differences in attention strategies between human perception and current AI models.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

AttentionVisual PerceptionEstheticsHumansFemaleMaleAdultYoung AdultEye-Tracking Technology

Autor:innen und Institutionen

  • School of Computer Science and Telecommunications, University fo Diego Portales, Santiago, Chile.
  • Faculty of Education and Psychology, University of Cordoba, Cordoba, Spain.
  • School of Computer Science and Telecommunications, University of Adolfo Ibáñez, Santiago, Chile.

ORCID-Profile: César González-Martín, Luis Oliveros

Förderung

  • University of Córdoba research group Lenguajes
  • Reuse Reduce Recycle AI-based platform for automated and scalable Maker culture in Circular economy

Zitieren

APA

Carrasco, M., González-Martín, C., Aranda, J., & Oliveros, L. (2026). Vision Transformer attention alignment with human visual perception in aesthetic object evaluation. PLOS One, 21(4), e0344006. https://doi.org/10.1371/journal.pone.0344006

Vancouver

Carrasco M, González-Martín C, Aranda J, Oliveros L. Vision Transformer attention alignment with human visual perception in aesthetic object evaluation. PLOS One. 2026;21(4):e0344006. doi:10.1371/journal.pone.0344006

BibTeX
@article{Carrasco2026,
  author = {Carrasco, Miguel and González-Martín, César and Aranda, José and Oliveros, Luis},
  title = {Vision Transformer attention alignment with human visual perception in aesthetic object evaluation},
  journal = {PLOS One},
  year = {2026},
  volume = {21},
  number = {4},
  pages = {e0344006},
  doi = {10.1371/journal.pone.0344006}
}
Kostenlos starten →