Face Identification based on Informative Knowledge Distillation using Morphological Model

Abstract

This manuscript deals with face detection and recognition with the help of a morphological-based Informative knowledge distillation model. It suggests a face prototype compression technique via discriminatory knowledge distillation, that can thoroughly study human potential with two simultaneous tasks, age estimation as well as face verification, for facial pictures captured at varying ages with greatly reduced information. The detailed and rigorous experimental study helps to understand the involvement of many variables such as gender, age gap, race, and age group. We propose Informative Knowledge Distillation Using Morphological Model (IKDMM) which can draw from current face patterns the most discriminatory facial attributes, which may be utilized to oversee the training process for face models with low resolution; and complete extensive tests to demonstrate that the compressed pattern is able to obtain an extreme high identification rate with comparative precision with the latest high-resolution face patterns.