PlasticGAN: Holistic generative adversarial network on face plastic and aesthetic surgery

Area of Science:

• Computer Vision
• Artificial Intelligence
• Medical Imaging

Background:

• Generative Adversarial Networks (GANs) have advanced face-related applications.
• Post-surgery facial changes present unique challenges for image generation and recognition.

Purpose of the Study:

• Introduce PlasticGAN, a framework for generating and reconstructing post-surgery faces.
• Aid surgeons and researchers in clinical decision-making through pre-treatment visualizations.
• Simulate aging patterns to explore tentative treatment alternatives.

Main Methods:

• Developed PlasticGAN, a novel framework based on Generative Adversarial Networks.
• Utilized face recognition systems to evaluate individuals with and without post-surgery facial alterations.
• Incorporated current trends like forensic, security, and COVID-related mask scenarios.

Main Results:

• Plastic surgery-based synthetic cross-age face recognition (PSBSCAFR) is identified as a significant research challenge.
• State-of-the-art face recognition systems demonstrate performance degradation when faced with post-surgery altered faces.
• The study highlights the impact of surgical modifications on biometric identification accuracy.

Conclusions:

• PlasticGAN offers a promising tool for visualizing and understanding the effects of plastic surgery.
• The findings underscore the need for robust face recognition systems capable of handling surgical facial changes.
• This research opens new avenues for exploring face recognition in forensic, security, and medical contexts.