Using Augmented Face Images to Improve Facial Recognition Tasks
arXiv:2205.06873v1h-index: 25
Originality Synthesis-oriented
AI Analysis
This addresses the issue of bias in facial recognition systems for underrepresented groups, but it is incremental as it builds on existing GAN-based augmentation methods.
The paper tackles the problem of underrepresented attributes in facial recognition by using GAN-augmented images to complement these attributes during training, resulting in improved inference quality for those specific attributes.
We present a framework that uses GAN-augmented images to complement certain specific attributes, usually underrepresented, for machine learning model training. This allows us to improve inference quality over those attributes for the facial recognition tasks.