AI Ethics

OUR WORK

FHIBE: A Fair Reflection

Sony AI’s Fair Human-Centric Image Benchmark (FHIBE) is the first publicly available globally diverse, consensually-collected fairness evaluation dataset for a wide variety of human-centric computer vision tasks.

OUR WORK

Beyond Skin Tone: A Multidimensional Measure of Apparent Skin Color

As part of this project, our researchers strive to measure apparent skin color in computer vision, beyond a unidimensional scale on skintone.

OUR WORK

Ethical Considerations for Responsible Data Curation

Human-centric computer vision (HCCV) data curation practices often neglect privacy and bias concerns. Our research focuses on proactive, domain-specific recommendations, covering purpose, privacy and consent, and diversity, for curating HCCV evaluation datasets, addressing privacy and bias concerns.

Fair Human-Centric Image Dataset for Ethical AI Benchmarking

Learn how FHIBE can be used responsibly as a fairness evaluation dataset for many human-centric computer vision tasks, including pose estimation, person segmentation, face detection and verification, and visual question answering.