A Comprehensive Survey on Face Image Analysis

A Comprehensive Survey on Face Image Analysis

Yu-Jin Zhang (Department of Electronic Engineering, Tsinghua University, China)
Copyright: © 2015 |Pages: 10
DOI: 10.4018/978-1-4666-5888-2.ch047

Chapter Preview



Using image analysis techniques to study face images has started some 30 years ago, one of the earliest work is dated to the 1980s (Baron, 1981). However, though automatic face recognition has been actively studied for over three decades as a means of human identification, it is only in recent the face image analysis has experienced significant growth and progress with the mature of image techniques.

The progress of research in face image analysis is reflected by the large number of papers published in these years. Take “face image analysis” as a key word and search it in the column of “Subject/Title/Abstract” in EI Compendex (http://www.ei.org) provides the results (the total number is more than 10000) shown in Figure 1. It shows the numbers of publication for each three years (accumulating the numbers over each of three years makes the curve smooth), which clearly illustrates the fast increasing in the last years. A jump from 2000-02 to 2003-05 is also obviously noticeable.

Figure 1.

Illustration of the numbers of publications with “facial image analysis”


Key Terms in this Chapter

Shape From Contour (SFC): A scene recovering technique with only one image. It recovers the surface orientation of objects with the contour information projected on the image.

Stereo Vision: A type of techniques for computing depth information based on images captured with two or more cameras. The main idea is to observe a scene from two or more viewpoints and to use the disparity to refer the position, structure and relation of objects in scene.

Image Engineering (IE): An integrated discipline/subject comprising the study of all the different branches of image and video techniques. As a general term for all image techniques, it could be considered as a broad subject encompassing mathematics, physics, biology, physiology, psychology, electrical engineering, computer science, automation, etc. Its advances are also closely related to the development of telecommunications, biomedical engineering, remote sensing, document processing, industrial applications, etc.

Structure From Motion (SFM): Recovers the information on the relationship among objects in space and/or the object structure of scene by solving the optical flow equations, which characterizes the motion information.

Active Appearance Model (AAM): A statistical model for matching shapes or image appearances. The model can be built by a training stage with registrated images. The application of this model in face tracking and matching could be helped by an optimization process with a least square criterion.

Shape From Shading (SFS): Recovers the shape information of a 3-D scene and objects from a single image. The image shading caused by the object surface illumination in space is used to reconstruct the surface orientation.

Facial Expression: Corresponding to face skin motions from one position to another position (from one state to another state) and conveying the emotional meaning of the actor to other observers. Facial expression classification is one of the important task of face image analysis.

Photometric Stereo: An important group of techniques for recovering surface orientation. The technique requires a set of images, which are taken from the same view angles but with different lighting conditions. This technique is easy to implement, but requires the control of the lighting in the application environment.

Image Analysis (IA): One of three layers of image engineering, which is concerned with the extraction of information (by meaningful measurements with descriptive parameters) from an image (especially from interesting objects).

Complete Chapter List

Search this Book: