3D human face portraits generation with mobile devices

• Main Author: Dong, Jijun
• Other Authors: Alex Chichung Kot
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Engineering
• Online Access: https://hdl.handle.net/10356/181643
• Institution: Nanyang Technological University

This report presents the overview and implementation of techniques for generating 3D human face portraits using mobile devices. The primary focus is on two models: the Hierarchical Representation Network (HRN) and the Tensorial Radiance Fields (TensoRF). These advanced models aim to enhance the quality of 3D face reconstruction. The HRN model applies hierarchical representations to capture intricate facial features with better precision. TensoRF utilises neural radiance fields to enhance the reconstruction process and achieve a 3D face reconstruction model. Multiple experiments and optimisations will be conducted on both models to improve their accuracy, realism, and performance. Through comparative analysis, insight into further advancements in 3D face reconstruction will be offered.