Model based 3D Face Reconstruction: A Survey

PhD Qualifying Examination


Title: "Model based 3D Face Reconstruction: A Survey"

by

Mr. Jiaxiang SHANG


Abstract:

It takes a long standing endeavor by the computer graphics and vision 
groups in structure a framework for 3D face reconstructing, analyzing 
human faces based on image info and depth sensor. This topic gets rapid 
advancement in the previous years, that prompted novel and incredible 
algorithms which obtain great outcomes and solid result.

Motivated by this rapid progress, this survey summarizes recent trends in 
3D face reconstruction and its applications, include global optimize the 
method and deep learning network. We focus our discussion on methods where 
the central task is to recover a three dimensional model of the human face 
and the scene properties (camera pose and illumination).

For this purpose, the 3D Morphable Model (3DMM) is commonly used. 3DMMs 
have been widely used for face analysis because the properties of 3DMM 
provide an ideal representation that is immune to intra-personal 
variations such as smoothing, symmetric and legal. Given a single facial 
input image or with depth map, a 3DMM can recover 3D face (shape and 
texture) via a fitting process.

We provide an overview of the underlying concepts of 3D face 
reconstruction pipeline, and we discuss common assumptions and 
simplifications that make these algorithms practical. In addition, we 
extensively apply the facial landmark detection(face-alignment) as a 
pre-process, which is a crucial step helps us find the correspondences 
between input data and 3DMM, and then we discuss the optimization 
techniques that are employed to recover dense photo-geometric 3D face 
models and pose from RGB or RGB-D data. Finally, we discuss a variety of 
use cases for the reviewed algorithms in the context of rigid 
registration, nor-rigid registration, as well as image and depth data 
fitting.


Date:			Tuesday, 27 August 2019

Time:                  	2:00pm - 4:00pm

Venue:                  Room 4472
                         Lifts 25/26

Committee Members:	Prof. Long Quan (Supervisor)
 			Prof. Chiew-Lan Tai (Chairperson)
 			Prof. Huamin Qu
 			Dr. Pedro Sander


**** ALL are Welcome ****