Two Panasonic divisions and a Singapore university have developed a way to train face biometrics algorithms that they say improves the performance of facial recognition for demographic groups represented by less training data.
Their method involves generating diverse data partitions iteratively in an unsupervised fashion, according to the abstract. The data partitions act as a self-annotation feature to deconfound the model through Invariant Feature Regularization. The researchers tested the method against the Masked Face and found that error rates were reduced across four racial groups and images of females.
