Home BlogDataset Evaluating Computer Vision Models

Evaluating Computer Vision Models

by Morgan Stevens
by
Code

Meta has created a dataset to improve fairness in computer vision models. The dataset contains 32,000 images featuring 50,000 people as well as annotations describing their perceived demographic attributes, physical characteristics, and roles. It also has 69,000 annotations that identify people, hair, or clothing in the images. Researchers can use the dataset to evaluate computer vision models for fairness and robustness across demographic categories.

Get the data.

Image credit: Flickr user Christiaan Colen

You may also like

Show Buttons
Hide Buttons