Table 3 Dataset comparison by geographical region and income level
From: Fair human-centric image dataset for ethical AI benchmarking
FHIBE (%) | CCv1 (%) | CCv2 (%) | FACET (%) | COCO (%) | MIAP (%) | |
|---|---|---|---|---|---|---|
Africa | 44.7 | 0.0 | 0.0 | 2.8 | 3.0 | 1.7 |
Asia and Oceania | 40.6 | 0.0 | 49.8 | 36.2 | 11.4 | 14.3 |
Europe | 4.4 | 0.0 | 0.0 | 49.8 | 34.2 | 36.2 |
Latin America and Caribbean | 4.2 | 0.0 | 42.5 | 3.5 | 3.1 | 5.0 |
North America | 6.0 | 100.0 | 7.7 | 7.7 | 48.3 | 42.8 |
High-income economies | 11.5 | 100.0 | 7.7 | 54.0 | 89.1 | 87.5 |
Upper-middle-income economies | 14.5 | 0.0 | 50.5 | 45.0 | 10.5 | 12.0 |
Lower-middle-income economies | 71.5 | 0.0 | 41.8 | |||
Low-income economies | 0.3 | 0.0 | 0.0 | 0.9 | 0.4 | 0.5 |
