Background learning for robust face recognition with PCA in the presence of clutter
TL;DR: A new method within the framework of principal component analysis (PCA) to robustly recognize faces in the presence of clutter by learning the distribution of background patterns and it is shown how this can be done for a given test image.
Abstract: We propose a new method within the framework of principal component analysis (PCA) to robustly recognize faces in the presence of clutter. The traditional eigenface recognition (EFR) method, which is based on PCA, works quite well when the input test patterns are faces. However, when confronted with the more general task of recognizing faces appearing against a background, the performance of the EFR method can be quite poor. It may miss faces completely or may wrongly associate many of the background image patterns to faces in the training set. In order to improve performance in the presence of background, we argue in favor of learning the distribution of background patterns and show how this can be done for a given test image. An eigenbackground space is constructed corresponding to the given test image and this space in conjunction with the eigenface space is used to impart robustness. A suitable classifier is derived to distinguish nonface patterns from faces. When tested on images depicting face recognition in real situations against cluttered background, the performance of the proposed method is quite good with fewer false alarms.
Citations
More filters
TL;DR: This paper has considered the performance of about twenty five different subspace algorithms on data taken from four standard face and object databases namely ORL, Yale, FERET and the COIL-20 object database.
38 citations
TL;DR: A positive correlation was found between learning improvement and attention, indicating that video-capture facial-recognition technology can be used to provide timely learning assistance and appropriate stimulation to enhance the educational benefits of e-learning.
Abstract: Recognition of students' facial expressions can be used to understand their level of attention. In a traditional classroom setting, teachers guide the classes and continuously monitor and engage the students to evaluate their understanding and progress. Given the current popularity of e-learning environments, it has become important to assess the degree of attention during the online learning process. In this study, we used interactive video-capture facial-recognition technology to automatically detect the facial expressions of students as a means of analyzing their attention state during the e-learning process. Participants were divided into three different learning-strategy groups for a course on computer networks. An attention-detection feedback module evaluated participants' attention span during the learning sessions and initiated a response to redirect the participants' attention when they became distracted. The three groups of participants showed significant differences in their course achievement;...
30 citations
29 citations
16 Sep 2018
TL;DR: The results reveal that the BNT thin films can crystallize well into a single phase of bismuth layered perovskite structure at 650°C and 700°C, however, a higher annealing temperature results in larger grains and better properties, which should be employed in various sensor applications.
Abstract: The Hong Kong Polytechnic University Hu Tian Tian i Abstract With the trends in environmental protection, there are growing interest and demand in developing lead-free ferroelectric materials for replacing the currently used lead-based materials in various piezoelectric and pyroelectric applications. In the present work, bismuth titanate-based ferroelectric materials Bi3.5-x/3Nd0.5Ti3-xNbxO12 (BNTN), both in the forms of ceramic and film, with good piezoelectric and pyroelectric properties have been successfully fabricated and have been shown to be potential candidates for high-temperature sensor applications. BNTN ceramics were prepared using a conventional mixed-oxide technique. The ceramics were sintered well into a single phase of bismuth layered perovskite structure at 1000°C for 4 hours. The effects of the niobium (Nb) dopant on the electrical, ferroelectric, piezoelectric and pyroelectric properties have then been investigated and discussed. No significant effect of the Nb dopant on the Curie temperatures (Tc) is observed; all the BNTN ceramics exhibit similar high values of Tc (~ 540°C). Concluding from the observations on the temperature and frequency dependences of the dielectric loss, the Nb dopant can effectively reduce the oxygen vacancies in the BNTN ceramics, while the oxygen vacancies are responsible for the high conductivity of the bismuth tianate-based ceramics. Our results also reveal that after the doping with Nb, the dielectric constant ε, coercive field Ec and leakage current J of the Bi3.5Nd0.5Ti3O12 (BNT) ceramic decrease, while its remanent polarization Pr, piezoelectric (charge) coefficient d33 and pyroelectric coefficient p increase. At the optimum doping level (~ 6 mol%), the ceramic exhibits the largest Abstract The Hong Kong Polytechnic UniversityThe Hong Kong Polytechnic University Hu Tian Tian ii Pr (15 μC/cm), largest d33 (22 μC/N), largest p (129 μC/mK), and the lowest J (3×10 A/cm). Since the ceramics also have a relatively low value of ε (~ 99), their piezoelectric (voltage) coefficient and figure of merit for pyroelectricity are large and comparable to those of a lead-based ferroelectric ceramic. Together with the high Curie temperature, the BNTN ceramics therefore should be good candidates for various high-temperature sensing applications. For MEMS application studies, Bi3.5Nd0.5Ti3O12 (BNT) thin films of thickness 900 nm have been successfully fabricated on platinized silicon substrates using a sol-gel method. The dielectric, ferroelectric as well as the piezoelectric and pyroelectric properties of the films were investigated. Our results reveal that the BNT thin films can crystallize well into a single phase of bismuth layered perovskite structure at 650°C and 700°C. However, a higher annealing temperature results in larger grains and better properties. For the BNT thin film annealed at 700°C, the observed ε, tanδ, Pr, p and e31,f are 132, 0.034, 22 μC/cm, 148 μC/mK and 1.45 C/m, respectively. Similar to the cases for ceramics, owing to the low ε, the piezoelectric (voltage) coefficient and figure of merit for pyroelectricity of the films are large and comparable to those of a lead-based ferroelectric thin film. Therefore, the BNT thin films could be employed in various sensor applications. Acknowledgment The Hong Kong Polytechnic University Hu Tian Tian iii Acknowledgments Without the great help of many kind people, this work would not have been possible. First of all, I am deeply indebted to my supervisor Dr. K. W. Kwok and the department head Prof. H.L.W. Chan for giving me the great opportunity to work in their excellent group, for giving me the numerous constructive suggestions, for giving me the countless encouragement, and for being patience during the course of completing my master degree. Then, I am also grateful to the colleagues, past and present, Mr. Rodney C. W. Tsang, and Ms. M. K. Cheung for their friendship, for their assistant in experimental and technical supports, and for their valuable discussions. Last, but not least, I appreciate the financial support from the Hong Kong Polytechnic University and the technical support provided by the Centre of Smart Materials of the Hong Kong Polytechnic University. Table of
29 citations
Cites methods from "Background learning for robust face..."
...The different applications have developed different techniques, such as face detection [11-15], face recognition [1, 2, 5, 16, 17], face tracking [18-20], facial expression recognition [5, 21-25], gender determination [26-28], age classification [29, 30], aging simulation [29, 31], face synthesizing [32-34] and 3D face analysis [35-38]....
[...]
21 Oct 2008
TL;DR: This paper addresses the problem of fully automated image segmentation in the context of ear biometrics using a low computational-cost appearance-based features and learning based Bayesian classifier in order to determine whether the segmentation outcome is proper or improper segment.
Abstract: Fully automated image segmentation is an essential step for designing automated identification systems. In this paper, we address the problem of fully automated image segmentation in the context of ear biometrics. Our segmentation approach achieves more than 90% accuracy based on three different sets of 3750 facial images for 376 persons. We also present an approach for the automated evaluation of the quality of segmented images. Our approach is based on low computational-cost appearance-based features and learning based Bayesian classifier in order to determine whether the segmentation outcome is proper or improper segment. Experimental results for evaluating the segmentation outcomes of ear images indicate the benefits of the proposed scheme.
22 citations
Cites background from "Background learning for robust face..."
...Special case 2: as mentioned in [21], when we choose the same number of Eigen subspace of each sub-class i w , ρ ρ ρ = = = 2 1 , and the same prior probabilities , we can write (8) as: )) ( min arg( 2 x i j j ε > < =...
[...]
References
More filters
14,948 citations
TL;DR: A near-real-time computer system that can locate and track a subject's head, and then recognize the person by comparing characteristics of the face to those of known individuals, and that is easy to implement using a neural network architecture.
Abstract: We have developed a near-real-time computer system that can locate and track a subject's head, and then recognize the person by comparing characteristics of the face to those of known individuals. The computational approach taken in this system is motivated by both physiology and information theory, as well as by the practical requirements of near-real-time performance and accuracy. Our approach treats the face recognition problem as an intrinsically two-dimensional (2-D) recognition problem rather than requiring recovery of three-dimensional geometry, taking advantage of the fact that faces are normally upright and thus may be described by a small set of 2-D characteristic views. The system functions by projecting face images onto a feature space that spans the significant variations among known face images. The significant features are known as "eigenfaces," because they are the eigenvectors (principal components) of the set of faces; they do not necessarily correspond to features such as eyes, ears, and noses. The projection operation characterizes an individual face by a weighted sum of the eigenface features, and so to recognize a particular face it is necessary only to compare these weights to those of known individuals. Some particular advantages of our approach are that it provides for the ability to learn and later recognize new faces in an unsupervised manner, and that it is easy to implement using a neural network architecture.
14,562 citations
Book•
01 Jan 1973
TL;DR: In this article, a unified, comprehensive and up-to-date treatment of both statistical and descriptive methods for pattern recognition is provided, including Bayesian decision theory, supervised and unsupervised learning, nonparametric techniques, discriminant analysis, clustering, preprosessing of pictorial data, spatial filtering, shape description techniques, perspective transformations, projective invariants, linguistic procedures, and artificial intelligence techniques for scene analysis.
Abstract: Provides a unified, comprehensive and up-to-date treatment of both statistical and descriptive methods for pattern recognition. The topics treated include Bayesian decision theory, supervised and unsupervised learning, nonparametric techniques, discriminant analysis, clustering, preprosessing of pictorial data, spatial filtering, shape description techniques, perspective transformations, projective invariants, linguistic procedures, and artificial intelligence techniques for scene analysis.
13,647 citations
TL;DR: A face recognition algorithm which is insensitive to large variation in lighting direction and facial expression is developed, based on Fisher's linear discriminant and produces well separated classes in a low-dimensional subspace, even under severe variations in lighting and facial expressions.
Abstract: We develop a face recognition algorithm which is insensitive to large variation in lighting direction and facial expression. Taking a pattern classification approach, we consider each pixel in an image as a coordinate in a high-dimensional space. We take advantage of the observation that the images of a particular face, under varying illumination but fixed pose, lie in a 3D linear subspace of the high dimensional image space-if the face is a Lambertian surface without shadowing. However, since faces are not truly Lambertian surfaces and do indeed produce self-shadowing, images will deviate from this linear subspace. Rather than explicitly modeling this deviation, we linearly project the image into a subspace in a manner which discounts those regions of the face with large deviation. Our projection method is based on Fisher's linear discriminant and produces well separated classes in a low-dimensional subspace, even under severe variation in lighting and facial expressions. The eigenface technique, another method based on linearly projecting the image space to a low dimensional subspace, has similar computational requirements. Yet, extensive experimental results demonstrate that the proposed "Fisherface" method has error rates that are lower than those of the eigenface technique for tests on the Harvard and Yale face databases.
11,674 citations
Book•
01 Jan 1972
TL;DR: This completely revised second edition presents an introduction to statistical pattern recognition, which is appropriate as a text for introductory courses in pattern recognition and as a reference book for workers in the field.
Abstract: This completely revised second edition presents an introduction to statistical pattern recognition Pattern recognition in general covers a wide range of problems: it is applied to engineering problems, such as character readers and wave form analysis as well as to brain modeling in biology and psychology Statistical decision and estimation, which are the main subjects of this book, are regarded as fundamental to the study of pattern recognition This book is appropriate as a text for introductory courses in pattern recognition and as a reference book for workers in the field Each chapter contains computer projects as well as exercises
10,526 citations