Determining Safe Areas for Injecting Cosmetic Fluids into the Face using Deep Learning Algorithms
The application of deep learning algorithms to recognize faces after cosmetic procedures is a growing area of research. It has been a challenging area of research for the past decade as a result of an exponential rise in both invasive and non-invasive procedures. Facial recognition is important when faces change with age. However, it has remained a puzzling challenge to computer vision for decades. Due to the sudden increase in the acceptance and implementation of deep learning models in computer vision, the features extracted by deep learning models can be used for facial recognition. In this paper, a method with minimum computational requirements and the least time complexity is presented to identify the faces altered by cosmetic fluid injections to know whether the person has injected the area or not and to recognize him. In this study, we present a technique to recognize faces and classify them to know if a person has undergone plastic surgery based on a deep learning algorithm: Conv Net (Conv Net) with Mobilenet_v2. Data were collected from the HDA- Plastic Surgery dataset containing samples for five facial regions for training. The results show that the shown technology enhances the accuracy of facial recognition and whether the person has been injected or not, with an accuracy of 99.8%.
Keywords - Deep Learning, CNN, Mobilenet_V2, Cosmetic Surgery, Image Classification.