A Deep Dive into Style-Based Age Manipulation: Revisiting SAM for Lifespan Face Editing

Swathi Gowroju; Devineni Gireesh Kumar; S A Md Noorulla Baig; K. Moulya Reddy; P. Devika; D. Sai Chand

doi:10.5935/jetia.v12i59.3690

Authors

Swathi Gowroju Associate Professor, Department of CSE(AI&ML), Sreyas Institute of Engineering and Technology, Hyderabad, 500068, Telangana, India
Devineni Gireesh Kumar Associate Professor, Department of CSE(AI&ML), Sreyas Institute of Engineering and Technology, Hyderabad, 500068, Telangana, India
S A Md Noorulla Baig Associate Professor, Department of CSE (AI), Vemu Institute Of Technology, P Kothakota, Chittoor - 517 112, AP, India.
K. Moulya Reddy UG Student, Department of CSE(AI&ML), Sreyas Institute of Engineering and Technology, Hyderabad, 500068, Telangana, India
P. Devika UG Student, Department of CSE(AI&ML), Sreyas Institute of Engineering and Technology, Hyderabad, 500068, Telangana, India
D. Sai Chand UG Student, Department of CSE(AI&ML), Sreyas Institute of Engineering and Technology, Hyderabad, 500068, Telangana, India

DOI:

https://doi.org/10.5935/jetia.v12i59.3690

Abstract

Change of facial age aims to alter the age of a person's photograph realistically but without loss of identity, a very important aspect for forensic work, entertainment, and all digital media. The Style-Based Age Manipulation (SAM) model is based on the expressive latent space of StyleGAN2 to provide photorealistic, seamless age progression and regression over the human life cycle (5 years to 100 years). SAM uses a pSp encoder to embed real images into the