13,143 face images annotated with 73 distinct facial attributes and identity labels across 5,749 unique individuals. The data provides continuous values for traits such as 'Middle Aged', 'Male', and 'Smiling' to support semantic face recognition.
Use Cases
- Train regression models to predict facial feature intensity using the 73 continuous attribute scores
- Develop face verification systems that utilize attribute-based descriptions to match identities
- Perform exploratory data analysis on the correlation between specific facial attributes like 'Male' and 'Brown Hair'
- Evaluate the impact of environmental factors like 'Flash' or 'Harsh Lighting' on the accuracy of face recognition algorithms
Strengths
- 13,143 images of faces with corresponding attribute labels
- 73 continuous attribute scores for features like 'Big Nose', 'Bushy Eyebrows', and 'Attractive'
- Identity labels for 5,749 unique individuals
- 250x250 pixel JPEG images captured from unconstrained web news sources