Prosthetic Face Recognition using a Siamese Neural Network Approach
Udit Kr. Chakraborty, Atharva Bendre, Shuvam Ganguli, Rinzing Namgyal · 2023
Neural Network based approaches have been extensively used for facial recognition. The success rate being high, such approaches have also gained popularity. In most facial recognition tasks, the identify is associated with the golden ratio. However, for made up faces, especially with prosthetic makeup, the golden ratio approach does not work. This paper proposes a network based on Siamese Neural Model that can identify faces even on use of prosthetic makeup. The network is tuned to make use of the powerful discriminative features to make a generalization of the predictive power of the network. The model’s ability to recognize faces, even when prosthetics have been used, has been confirmed through experiments with reasonable success under constrained laboratory conditions.