Exploring deep learning methods for short-term skin texture simulation

• Main Author: Chen, Ziyu
• Other Authors: Alex Chichung Kot
• Format: Thesis-Master by Coursework
• Language: English
• Published: Nanyang Technological University 2022
• Subjects: Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
• Online Access: https://hdl.handle.net/10356/158925
• Institution: Nanyang Technological University

The skin plays a fundamental and significant role in our survival and body health. Also, our skin condition affects our mental health. Given the importance of skincare, it has attracted more and more interest in recent years and one of its significant parts is skin simulation, especially facial skin simulation. However, traditional methods require complicated modeling and lack robustness, and deep learning methods related to short-term facial skin texture simulation are yet to be explored. Hence it’s necessary to explore possible approaches to short-term facial skin texture simulation. In this dissertation, we explore several different deep learning methods, examine the viability of these approaches and give analysis. Moreover, we propose a possible approach: adopt U-Net to conduct pore segmentation and apply it to generator training.