Exposing Blindspots: Cultural Bias Evaluation in Generative Image Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915869327622144 |
|---|---|
| author | Seo, Huichan Choi, Sieun Hong, Minki Zhou, Yi Kim, Junseo Ismaila, Lukman Etori, Naome Agarwal, Mehul Liu, Zhixuan Kim, Jihie Oh, Jean |
| author_facet | Seo, Huichan Choi, Sieun Hong, Minki Zhou, Yi Kim, Junseo Ismaila, Lukman Etori, Naome Agarwal, Mehul Liu, Zhixuan Kim, Jihie Oh, Jean |
| contents | Generative image models produce striking visuals yet often misrepresent culture. Prior work has examined cultural bias mainly in text-to-image (T2I) systems, leaving image-to-image (I2I) editors underexplored. We bridge this gap with a unified evaluation across six countries, an 8-category/36-subcategory schema, and era-aware prompts, auditing both T2I generation and I2I editing under a standardized protocol that yields comparable diagnostics. Using open models with fixed settings, we derive cross-country, cross-era, and cross-category evaluations. Our framework combines standard automatic metrics, a culture-aware retrieval-augmented VQA, and expert human judgments collected from native reviewers. To enable reproducibility, we release the complete image corpus, prompts, and configurations. Our study reveals three findings: (1) under country-agnostic prompts, models default to Global-North, modern-leaning depictions that flatten cross-country distinctions; (2) iterative I2I editing erodes cultural fidelity even when conventional metrics remain flat or improve; and (3) I2I models apply superficial cues (palette shifts, generic props) rather than era-consistent, context-aware changes, often retaining source identity for Global-South targets. These results highlight that culture-sensitive edits remain unreliable in current systems. By releasing standardized data, prompts, and human evaluation protocols, we provide a reproducible, culture-centered benchmark for diagnosing and tracking cultural bias in generative image models. Project page: https://seochan99.github.io/ECB |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_20042 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Exposing Blindspots: Cultural Bias Evaluation in Generative Image Models Seo, Huichan Choi, Sieun Hong, Minki Zhou, Yi Kim, Junseo Ismaila, Lukman Etori, Naome Agarwal, Mehul Liu, Zhixuan Kim, Jihie Oh, Jean Computer Vision and Pattern Recognition I.2.10; I.2.6; I.4.9 Generative image models produce striking visuals yet often misrepresent culture. Prior work has examined cultural bias mainly in text-to-image (T2I) systems, leaving image-to-image (I2I) editors underexplored. We bridge this gap with a unified evaluation across six countries, an 8-category/36-subcategory schema, and era-aware prompts, auditing both T2I generation and I2I editing under a standardized protocol that yields comparable diagnostics. Using open models with fixed settings, we derive cross-country, cross-era, and cross-category evaluations. Our framework combines standard automatic metrics, a culture-aware retrieval-augmented VQA, and expert human judgments collected from native reviewers. To enable reproducibility, we release the complete image corpus, prompts, and configurations. Our study reveals three findings: (1) under country-agnostic prompts, models default to Global-North, modern-leaning depictions that flatten cross-country distinctions; (2) iterative I2I editing erodes cultural fidelity even when conventional metrics remain flat or improve; and (3) I2I models apply superficial cues (palette shifts, generic props) rather than era-consistent, context-aware changes, often retaining source identity for Global-South targets. These results highlight that culture-sensitive edits remain unreliable in current systems. By releasing standardized data, prompts, and human evaluation protocols, we provide a reproducible, culture-centered benchmark for diagnosing and tracking cultural bias in generative image models. Project page: https://seochan99.github.io/ECB |
| title | Exposing Blindspots: Cultural Bias Evaluation in Generative Image Models |
| topic | Computer Vision and Pattern Recognition I.2.10; I.2.6; I.4.9 |
| url | https://arxiv.org/abs/2510.20042 |