英语原文共 9 页，剩余内容已隐藏，支付完成后下载完整资料
说明：
封面之后放英文原文；英文原文之后放译文。
原文：
Date:2019
A Brief Survey: 3D Face Reconstruction
Tianhan Gao1(B) and Hui An2(B)
Engineering Research Center of Scurity Technology of Complex Network System,
Ministry of Education; Liaoning Research Center of Safety Engineering Technology in Industrial Control, NorthEastern University, Shenyang, China gaoth@mail.neu.edu.cn
NorthEastern University, Shenyang, China 1871069@stu.neu.edu.cn
Abstract. 3D reconstruction technology is an important branch in the field of computer vision. Due to the large proportion of face information when people browsing pictures and the rapid development of virtual reality applications, face reconstruction technology has attracted more attention from researchers over many years. In terms of the special structure of human face, 3D faces reconstruction is quite different from ordinary object 3D reconstruction. This paper summaries the approaches of obtaining face data and the expression of 3d face. The reconstruction effect of each method is analyzed, including the advantages and disadvantages, the issues that can be improved, as well as the prospect of the future work.
Keywords: Computer vision · 3D face reconstruction · Expression of 3D face
Introduction
With the rapid development of computer vision and virtual reality technology, a branch of it the face reconstruction technology has made great progress, and its performance in the film and television industry is the most significant. The film and face effects we have seen are based on three-dimensional facial expression capture and construction. The mold technology, high-precision modeling can clearly express the wrinkles, pores and light transmission effect of the face, and its fineness can be compared with the real face. The face reconstruction technology in the game industry has also developed rapidly, and the virtual characters generated by the 3D face reconstruction technology are widely used. Face is an important symbol of human identity. In theory, each face should only correspond to a specific identity, so the research of face reconstruction has its important value. Face reconstruction has an important position in identity recognition, film production, animation games, social applications, and medical effects preview. Currently, the methods for obtaining face data are different, and are divided into active acquisition and passive acquisition. At the same time, depending on the different forms of the acquired data, there are different reconstruction ideas. Sogou and Xinhua News Agency jointly launched the worldrsquo;s first AI anchor. The emergence of lsquo;virtual anchorrsquo; is the result of the combination of face reconstruction technology and digital animation. This successful face reconstruction technology in the industry has caused a great sensation that shows the influence of 3D face reconstruction technology.
3D Face Reconstruct Method
Abstract. Nowadays, the main 3d face reconstruction technologies in the industry are divided into three categories: (1) active methods, which mainly include methods based on structured light technology, laser scanning, and Time-of-flight (ToF). (2) Passive method, which mainly includes Multi-view Stereo, shape from shading, data-driven method; (3) method based on a single image.
Active Method
High-quality equipment can extract more accurate information. In some professional fields, it is necessary to restore the face model with high precision which requires us to use some equipment actively to extract more accurate face information and obtain accurate Information through high-quality equipment to meet the high precision requirements of the industry.
Structured Light
The basic principle of 3D structured light [8] is to project a specific coding pattern structured light onto the object which to be measured, and then collect it by a photoelectric sensor such as a special infrared camera. This kind of light with a certain structure will acquire different image phase information according to the depth of different object area, and then calculate the depth information through the arithmetic unit to obtain the three-dimensional structure of the object. Zhang et al. [19]used a structured light system with two projectors to project grayscale stripes onto the human face, obtain depth data by calculation, and obtain geometric features and texture features and establish a correspondence between facial geometric features by matching a common facial template with the acquired depth data. In real life, Apple mobile phones and Xiaomi mobile phones are also based on structured light-based face recognition technology. They project a specific structure of light through the front projector of the mobile phone, as shown in Fig.1, and then capture these with an infrared camera. The light spot uses the reflected signal to construct a 3D face that is contrasted with the previously entered information. The only difference is that the projected structured light are different. Well-known Kinect also use structured light technology.
Fig.1. Structured light pattern of Apple (left), Structured light pattern of Xiaomi
(right)
Laser Scanning
The basic principle of 3D structured light [1] is to project a specific coding pattern structured light onto the objectLidar based on triangulation algorithm is mostly non-real-time laser thunder, its principle is as follows: When the laser is emitted from the laser
剩余内容已隐藏，支付完成后下载完整资料
资料编号：[253915]，资料为PDF文档或Word文档，PDF文档可免费转换为Word
以上是毕业论文外文翻译，课题毕业论文、任务书、文献综述、开题报告、程序设计、图纸设计等资料可联系客服协助查找。
