RMFD, short for Real-World Masked Face Dataset, is a dataset on masked face recognition. Based on this data, researchers can design algorithms for detecting and recognizing faces obscured by masks, aiding in community lockdowns, upgrading facial recognition gates and access control/attendance devices, and more.

The dataset includes:

• Real-world masked face recognition dataset (RMFRD): 5,000 masked face images of 525 people and 90,000 normal face images.
• Simulated masked face recognition datasets (SMFRD): 500,000 simulated masked face images of 10,000 people.
• The Real-world Masked Face Verification Dataset (RMFVD) consists of 4,015 face images from 426 individuals, which were further organized into 7,178 sample pairs (masked faces/normal faces), including 3,589 pairs of samples with the same identity and 3,589 pairs of face sample pairs with different identities.