Researchers at the University of Michigan developed a deep learning based-system that performs real-time facial recognition and verifies the photo against the corresponding passport and government-issued IDs. The method has the potential to help law enforcement prevent fraud, could serve as an alternative payment method, and could also keep known criminals from entering a sensitive or secured area.
“Numerous activities in our daily life, including transactions, access to services and transportation, require us to verify who we are by showing our ID documents containing face images, e.g., passports and driver licenses,” the researchers stated in their research paper. “An automatic system for matching ID document photos to live face images in real time with high accuracy would speed up the verification process and remove the burden on human operators.”
Using NVIDIA GeForce GTX 1080 Ti GPUs with the cuDNN accelerated TensorFlow deep learning framework, the team trained their neural network on over 30,000 images of ID cards and headshots. The team then validated their results on a separate dataset and achieved a 92% accuracy level.
The researchers said that only a few studies have been conducted on this topic and all of them are over five years old. The current methods also struggle to match ID card photos with the corresponding images.
“Face recognition technology has made tremendous strides in the past five years, mainly due to the availability of large-scale face training data and deep neural network models for face recognition,” the researchers stated.
The researchers will continue to test their system on photos from different devices, such as selfies from mobile phones, and photos from stationary cameras in different lighting conditions.