RAF-DB, short for Real-world Affective Faces, is a large-scale facial expression dataset. The dataset consists of 29,672 diverse facial images, with 40 labelers annotating the images with basic or compound expressions.

In addition, each image also includes 5 precise landmark locations, 37 automatic landmark locations, bounding box, race, age range and gender attribute annotations.

The faces in the images in this dataset vary greatly in age, gender and ethnicity, head pose, lighting conditions, occlusions (such as glasses, facial hair or self-occlusion), and post-processing operations (such as various filters and special effects).