Showing papers by "Stan Z. Li published in 2004"

FloatBoost learning and statistical face detection

Abstract: A novel learning procedure, called FloatBoost, is proposed for learning a boosted classifier for achieving the minimum error rate. FloatBoost learning uses a backtrack mechanism after each iteration of AdaBoost learning to minimize the error rate directly, rather than minimizing an exponential function of the margin as in the traditional AdaBoost algorithms. A second contribution of the paper is a novel statistical model for learning best weak classifiers using a stagewise approximation of the posterior probability. These novel techniques lead to a classifier which requires fewer weak classifiers than AdaBoost yet achieves lower error rates in both training and testing, as demonstrated by extensive experiments. Applied to face detection, the FloatBoost learning method, together with a proposed detector pyramid architecture, leads to the first real-time multiview face detection system reported.

Face recognition under varying lighting conditions using self quotient image

Abstract: In this paper, we introduce the concept of self-quotient image (SQI) for robust face recognition under varying lighting conditions. It is based on the quotient image method (Amnon Shashua et al., 2001) (T. Riklin-Raviv et al., 1999) to achieve lighting invariance. However, the SQI has three advantages: (1) it needs only one face image for extraction of intrinsic lighting invariant property of a face while removing extrinsic factor corresponding to the lighting; (2) no alignment is needed; (3) it works in shadow regions. The theoretical analysis on conditions where the algorithm is applicable and a non-iterative filtering algorithm for computing SQI are presented. Experiment results demonstrate the effectiveness of our method for robust face recognition under varying lighting conditions.

Boosting local binary pattern (LBP)-Based face recognition

Abstract: This paper presents a novel approach for face recognition by boosting statistical local features based classifiers The face image is scanned with a scalable sub-window from which the Local Binary Pattern (LBP) histograms [14] are obtained to describe the local features of a face image The multi-class problem of face recognition is transformed into a two-class one by classifying every two face images as intra-personal or extra-personal ones [9] The Chi square distance between corresponding Local Binary Pattern histograms of two face images is used as discriminative feature for intra/extra-personal classification We use AdaBoost algorithm to learn a similarity of every face image pairs The proposed method was tested on the FERET FA/FB image sets and yielded an exciting recognition rate of 979%

Generalized quotient image

Abstract: We present a unified framework for modeling intrinsic properties of face images for recognition. It is based on the quotient image (QI) concept, in particular on the existing works of QI, spherical harmonic, image ratio and retinex. Under this framework, we generalize these previous works into two new algorithms: (1) non-point light quotient image (NPL-QI) extends QI to deal with non-point light sources by modeling non-point light directions using spherical harmonic bases; (2) self-quotient image (S-QI) extends QI to perform illumination subtraction without the need for alignment and no shadow assumption. Experimental results show that our algorithms can significantly improve the performance of face recognition under varying illumination conditions.

Face recognition using Ada-Boosted Gabor features

Abstract: Face representation based on Gabor features has attracted much attention and achieved great success in face recognition area for the advantages of the Gabor features. However, Gabor features currently adopted by most systems are redundant and too high dimensional. In this paper, we propose a face recognition method using AdaBoosted Gabor features, which are not only low dimensional but also discriminant. The main contribution of the paper lies in two points: (1) AdaBoost is successfully applied to face recognition by introducing the intra-face and extra-face difference space in the Gabor feature space; (2) an appropriate re-sampling scheme is adopted to deal with the imbalance between the amount of the positive samples and that of the negative samples. By using the proposed method, only hundreds of Gabor features are selected. Experiments on FERET database have shown that these hundreds of Gabor features are enough to achieve good performance comparable to that of methods using the complete set of Gabor features.

Self quotient image for face recognition

Abstract: The reliability of facial recognition techniques is often affected by the variation of illumination, such as shadows and illumination direction changes. In this paper, we present a novel framework, called the self-quotient image, for the elimination of the lighting effect in the image. Although this method has a similar invariant form to the quotient image by Shashua etc. (2001), it does not need the alignment and bootstrap images. Our method combines the image processing technique of edge-preserved filtering with the Retinex applications of by Jobson, et al., (1997) and Gross and Brajovie (2003). We have analyzed this algorithm with a 3D imaging model and formulated the conditions where illumination-invariant and -variant properties can be realized, respectively. A fast anisotropic filter is also presented. The experiment results show that our method is effective in removing the effect of illumination for robust face recognition.

Null space approach of fisher discriminant analysis for face recognition

Abstract: The null space of the within-class scatter matrix is found to express most discriminative information for the small sample size problem (SSSP). The null space-based LDA takes full advantage of the null space while the other methods remove the null space. It proves to be optimal in performance. From the theoretical analysis, we present the NLDA algorithm and the most suitable situation for NLDA. Our method is simpler than all other null space approaches, it saves the computational cost and maintains the performance simultaneously. Furthermore, kernel technique is incorporated into discriminant analysis in the null space. Firstly, all samples are mapped to the kernel space through a better kernel function, called Cosine kernel, which is proposed to increase the discriminating capability of the original polynomial kernel function. Secondly, a truncated NLDA is employed. The novel approach only requires one eigenvalue analysis and is also applicable to the large sample size problem. Experiments are carried out on different face data sets to demonstrate the effectiveness of the proposed methods.

Boosting Local Feature Based Classifiers for Face Recognition

Abstract: In this paper, we present a method for face recognition using boosted Gabor feature based classifiers. Weak classifiers are constructed based on both magnitude and phase features derived from Gabor filters [Quadrature-phase simple-cell pairs are ap-propriately described in complex analytic from]. The multi-class problem is transformed into a two-class one of intra- and extra-class classification using intra-personal and extra-personal difference images, as in [Beyond euclidean eigenspaces:bayesian matching for visian recognition]. A cascade of strong classifiers are learned using bootstrapped negative examples, similar to the way in face detection framework [Robust real time object detection]. The combination of classifiers based on two different types of features produces better results than using either type. Experiments on FERET database show good results comparable to the best one reported in literature [The FERET evaluation methodology for face-recognition algorithms].

Nearest manifold approach for face recognition

Abstract: Faces under varying illumination, pose and non-rigid deformation are empirically thought of as a highly nonlinear manifold in the observation space. How to discover intrinsic low-dimensional manifold is important to characterize meaningful face distributions and classify them using a simpler, such as linear or Gaussian based, classifier. In this paper, we present a manifold learning algorithm (MLA) for learning a mapping from highly-dimensional manifold into the intrinsic low-dimensional linear manifold. We also propose the nearest manifold (NM) criterion for the classification and present an algorithm for computing the distance from the sample to be classified to the nearest face manifolds in light of local linearity of manifold. Based on these works, face recognition is achieved with the combination of MLA and NM. Experiments on several face databases show that the advantages of our proposed combinational approach.

Real-time and accurate segmentation of moving objects in dynamic scene

Abstract: Fast and accurate segmentation of moving objects in video sequences is a basic task in many computer vision and video analysis applications. It has a critical impact on the performance of object tracking and classification and activity analysis. This paper presents effective methods for solving this problem. Firstly, a fast and efficient algorithm is presented for background update to handle various sources of scene changes, including ghosts, left objects, camera shaking, and abrupt illumination changes. This is done by analyzing properties of object motion in image pixels and temporal frames, and combining both levels of constraints. Moreover, the algorithm does not need training sequence. Secondly, a real-time and accurate moving object segmentation algorithm is presented for moving object localization. Here, a novel filtering method is presented based on multiple scale and fast connected blob extraction. An intelligent video surveillance system is developed to test the performance of the algorithms. Experiments are performed using long video sequences under different conditions indoor and outdoor. The results show that the proposed algorithm is effective and efficient in real-time and accurate background update and moving object segmentation.

Null space-based kernel Fisher discriminant analysis for face recognition

Abstract: The null space-based LDA takes full advantage of the null space while the other methods remove the null space. It proves to be optimal in performance. From the theoretical analysis, we present the NLDA algorithm and the most suitable situation for NLDA. Our method is simpler than all other null space approaches, it saves the computational cost and maintains the performance simultaneously. Furthermore, kernel technique is incorporated into our null space method. Firstly, all samples are mapped to the kernel space through an efficient kernel function, called cosine kernel, which have been demonstrated to increase the discriminating capability of the original polynomial kernel function. Secondly, a truncated NLDA is employed. The novel approachh only requires one eigenvalu analysis and is also applicable to the large sample size problem. Experiments are carried out on different face data sets to demonstrate the effectiveness of the proposed method.

Real-time face detection using boosting in hierarchical feature spaces

Abstract: Boosting-based methods have recently led to the state-of-the-art-face detection systems. In these systems, weak classifiers to be boosted are based on simple, local, Haar-like features. However, it can be empirically observed that in later stages of the boosting process, the non-face examples collected by bootstrapping become very similar to the face examples, and the classification error of Haar-like feature based weak classifiers is thus very close to 50%. As a result, the performance of a face detector cannot be further improved. This paper proposed a solution to this problem, introducing a face detection method based on boosting in hierarchical feature spaces (both local and global). We argue that global features, like those derived from principal component analysis, can be advantageously used in the later stages of boosting, when local features do not provide any further benefit. We show that weak classifiers learned in hierarchical feature spaces are better boosted. Our methodology leads to a face detection system that achieves higher performance than a current state-of-the-art system, at a comparable speed.

Face alignment under variable illumination

Abstract: This paper presents an approach to face alignment under variable illumination, an obstacle largely ignored in previous 2D alignment work. To account for illumination variation, our method employs two forms of relatively lighting-invariant information. Edge phase congruency is adopted to coarsely localize facial features, since prominent feature edges can be robustly located in the presence of shading and shadows. To accurately deal with features with less pronounced edges, final alignment is then computed from intrinsic gray-level information recovered using a proposed form of local intensity normalization. With this approach, our face alignment system works efficiently and effectively under a wide range of illumination conditions, as evidenced by extensive experimentation.

Shape evaluation for weighted active shape models

Abstract: Active Shape Models (ASM) is a powerful statistical tool for face alignment. However, no evaluation is performed on the final results. Nevertheless, the shape evaluation information is very useful for the search and the final results. In this paper, a shape evaluation method and a new search algorithm, called weighted ASM, are proposed. The shape evaluation is based on the local appearance model of ASM to determine how well the searching shape match models derived from the training set. It is used to guide the search procedure to get more accurate results. The weighted-ASM also uses this evaluation information to project the searching shape into the solution shape space in a weighted way. Compared with ASM’s orthogonal projection, the weighted projection can drag the search out of local minima to be more accurate and more robust. Experiments have been done to show the ability of this method to align shapes.

Subspace analysis and optimization for AAM based face alignment

Abstract: Active appearance models (AAM) is very powerful for extracting objects, e.g. faces, from images. It is composed of two parts: the AAM subspace model and the AAM search. While these two parts are closely correlated, existing efforts treated them separately and had not considered how to optimize them overall. In this paper, an approach is proposed to optimize the subspace model while considering the search procedure. We first perform a subspace error analysis, and then to minimize the AAM error we propose an approach which optimizes the subspace model according to the search procedure. For the subspace error analysis, we decomposed the subspace error into two parts, which are introduced by the subspace model and the search procedure respectively. This decomposition shows that the optimal results of AAM can be achieved only by optimizing both of them jointly rather than separately. Furthermore, based on this error decomposition, we develop a method to find the optimal subspace model according to the search procedure by considering both the two decomposed errors. Experimental results demonstrate that our method can find the optimal AAM subspace model rapidly and improve the performance of AAM significantly.

Multiple layer based background maintenance in complex environment

Abstract: A fast and efficient multiple layer background maintenance model is built to conserve the original and the current background separately. Fusing the properties of object motion in image pixels and the changes between the input video and the multiple background layers, this method could handle various sources of scene changes, including ghosts, abandon objects and illumination changes. An intelligent video surveillance system is developed to test the performance of the algorithm. Experiments are performed using long video sequences under different conditions indoor and outdoor. The results show that the proposed algorithm is effective and efficient in real-time and accurate background maintenance in complex environment.

Statistical learning of evaluation function for ASM/AAM image alignment

Abstract: Alignment between the input and target objects has great impact on the performance of image analysis and recognition system, such as those for medical image and face recognition. Active Shape Models (ASM) [1] and Active Appearance Models (AAM) [2, 3] provide an important framework for this task. However, an effective method for the evaluation of ASM/AAM alignment results has been lacking. Without an alignment quality evaluation mechanism, a bad alignment cannot be identified and this can drop system performance.

Nearest intra-class space classifier for face recognition

Abstract: We propose a novel classification method, called nearest intra-class space (NICS), for face recognition. In our method, the distribution of face patterns of each person is represented by the intra-class space to capture all intra-class variations. Then, a regular principal subspace is derived from each intra-class space using principal component analysis. The classification is based on the nearest weighted distance, combining distance-from-subspace and distance-in-subspace, between the query face and each intra-class subspace. Experimental results show that the NICS classifier outperforms other classifiers in terms of recognition performance.

Learning with Cascade for Classification of Non-Convex Manifolds

Abstract: Images of a visual object, such as human face, reside in a complicated manifold in the high dimensional image space, when the object is subject to variations in pose, illumination, and other factors. Viola and Jones have successfully tackled difficult nonlinear classification problem for face detection using AdaBoost learning. Moreover, their simple-to-complex cascade of classifiers structure makes the learning and classification even more effective. While training with cascade has been used effectively in many works [4, 5, 6, 7, 2, 3, 8, 9, 10], an understanding of the role of the cascade strategy is still lacking. In this paper, we analyze the problem of classifying non-convex manifolds using AdaBoost learning with and without using cascade. We explain that the divide-and-conquer strategy in cascade learning has a great contribution on learning a complex classifier for non-convex manifolds. We prove that AdaBoost learning with cascade is effective when a complete or over-complete set of features (or weak classifiers) is available. Experiments with both synthesized and real data demonstrate that AdaBoost learning with cascade leads to improved convergence and accuracy.

Face recognition: technical challenges and research directions

Abstract: Face recognition performance has improved significantly since the first automatic face recognition system developed by Kanade Face detection, facial feature extraction, and recognition can now be performed in “realtime” for images captured under favorable, constrained situations Although progress in face recognition has been encouraging, the task has also turned out to be a difficult endeavor, especially for unconstrained tasks where viewpoint, illumination, expression, occlusion, accessories, and so on vary considerably In this talk, I will analyze challenges from the viewpoint of face manifolds and points out possible research directions towards highly accurate face recognition I will show that the challenges come from high nonconvexity of face manifolds, in the image space, under variations in lighting, pose and so on; unfortunately, there have been no good methods from theories of pattern recognition for solving such difficult problems, especially when the size of training data is small However, there are two directions to look at towards possible solutions: One is to construct a “good” feature space in which the face manifolds become less complex i.e., less nonlinear and nonconvex than those in other spaces This includes two levels of processing: (1) normalize face images geometrically and photometrically, such as using morphing and histogram equalization; and (2) extract features in the normalized images which are stable with respect to the said variations, such as based on Gabor wavelets The second strategy is to construct classification engines able to solve less, although still, nonlinear problems in the feature space, and to generalize better A successful algorithm usually combines both strategies Still another direction is on system design, including sensor hardware, to make the pattern recognition problems thereafter less challenging.

Evaluation of face alignment solutions using statistical learning

Abstract: We propose a statistical learning approach for constructing an evaluation function for face alignment. A nonlinear classification function is learned from a set of positive (good alignment) and negative (bad alignment) training examples to effectively distinguish between qualified and un-qualified alignment results. The AdaBoost learning algorithm is used, where weak classifiers are constructed based on edge features and combined into a strong classifier. Several strong classifiers are learned in stages using bootstrap samples during the training. The evaluation function thus learned gives a quantitative confidence and the good-bad classification is achieved by comparing the confidence with a learned optimal threshold. We point out the importance of using cascade strategy in the stagewise learning of strong classifiers. The divide-and-conquer strategy not only dramatically increases the speed of classification, but also makes the training easier and the good-bad classification more effective. Experimental results demonstrate that the classification function learned using the proposed approach provides semantically more meaningful scoring than the reconstruction error used in AAM for classification between qualified and un-qualified face alignment.

Advances in biometric person authentication : 5th Chinese Conference on Biometric Recognition, SINOBIOMETRICS 2004, Guangzhou, China, December 13-14, 2004 : proceedings

Abstract: Biometrics.- Biometrics: When Identity Matters.- Face Recognition: Technical Challenges and Research Directions.- Fingerprints: Recognition, Performance Evaluation and Synthetic Generation.- Recognising Persons by Their Iris Patterns.- Multiple Classifier Fusion for Biometric Authentication.- Performance Evaluation in 1 : 1 Biometric Engines.- Best Performing Biometric Engines.- Discussions on Some Problems in Face Recognition.- Improving Fingerprint Recognition Performance Based on Feature Fusion and Adaptive Registration Pattern.- Iris Recognition Based on Non-local Comparisons.- Palmprint Authentication Technologies, Systems and Applications.- Face Recognition.- Novel Face Detection Method Based on Gabor Features.- Optimal Shape Space and Searching in ASM Based Face Alignment.- Gabor Wavelet-Based Eyes and Mouth Detection Algorithm.- An Entropy-Based Diversity Measure for Classifier Combining and Its Application to Face Classifier Ensemble Thinning.- Estimating the Visual Direction with Two-Circle Algorithm.- Multiple Face Contour Detection Using Adaptive Flows.- Pose Normalization Using Generic 3D Face Model as a Priori for Pose-Insensitive Face Recognition.- Gabor-Based Kernel Fisher Discriminant Analysis for Pose Discrimination.- Robust Pose Estimation of Face Using Genetic Algorithm.- Facial Pose Estimation Based on the Mongolian Race's Feature Characteristic from a Monocular Image.- Boosting Local Binary Pattern (LBP)-Based Face Recognition.- Gabor Features Based Method Using HDR (G-HDR) for Multiview Face Recognition.- Face Recognition Under Varying Lighting Based on Derivates of Log Image.- A Fast Method of Lighting Estimate Using Multi-linear Algebra.- Face Recognition Using More than One Still Image: What Is More?.- Video-Based Face Recognition Using a Metric of Average Euclidean Distance.- 3D Face Recognition Based on G-H Shape Variation.- 3D Face Recognition Based on Geometrical Measurement.- 3D Face Recognition Using Eigen-Spectrum on the Flattened Facial Surface.- Building a 3D Morphable Face Model by Using Thin Plate Splines for Face Reconstruction.- 3D Surface Reconstruction Based on One Non-symmetric Face Image.- Recent Advances in Subspace Analysis for Face Recognition.- Component-Based Cascade Linear Discriminant Analysis for Face Recognition.- Unified Locally Linear Embedding and Linear Discriminant Analysis Algorithm (ULLELDA) for Face Recognition.- On Dimensionality Reduction for Client Specific Discriminant Analysis with Application to Face Verification.- The Solution Space for Fisher Discriminant Analysis and the Uniqueness Under Constraints.- A Novel One-Parameter Regularized Linear Discriminant Analysis for Solving Small Sample Size Problem in Face Recognition.- Fast Calculation for Fisher Criteria in Small Sample Size Problem.- Vision-Based Face Understanding Technologies and Their Applications.- International Standardization on Face Recognition Technology.- System Design and Assessment Methodology for Face Recognition Algorithms.- Baseline Evaluations on the CAS-PEAL-R1 Face Database.- An Efficient Compression and Reconstruction Method of Face Image for Low Rate Net.- How Can We Reconstruct Facial Image from Partially Occluded or Low-Resolution One?.- A Matrix-Oriented Method for Appearance-Based Data Compression - An Idea from Group Representation Theory.- Fingerprint Recognition.- An Adaptive Fingerprint Post-processing Algorithm Based on Mathematical Morphology.- Fingerprint Image Segmentation by Energy of Gaussian-Hermite Moments.- Robust Ridge Following in Fingerprints.- A New Approach for Fingerprint Minutiae Extraction.- A Top-Down Fingerprint Image Enhancement Method Based on Fourier Analysis.- Fingerprint Templates Combination.- Skeletonization of Fingerprint Based-on Modulus Minima of Wavelet Transform.- Transformation-Variants Estimation Using Similarity Relative Histogram Grouping Model.- A Study of Minutiae Matching Algorithm Based on Orientation Validation.- Cascading a Couple of Registration Methods for a High Accurate Fingerprint Verification System.- A Hierarchical Fingerprint Matching Method Based on Rotation Invariant Features.- Phase-Correlation Based Registration of Swipe Fingerprints.- An Improved Method for Singularity Detection of Fingerprint Images.- Fingerprint Classifier Using Embedded Hidden Markov Models.- A Robust Pseudoridges Extraction Algorithm for Fingerprints.- Iris Recognition.- Iris Image Capture System Design for Personal Identification.- An Iris Segmentation Procedure for Iris Recognition.- Zernike Moment Invariants Based Iris Recognition.- Two-Dimensional Projection and Crossing for Iris Optimal Localization.- Speaker Recognition.- Improvement of Speaker Identification by Combining Prosodic Features with Acoustic Features.- Bimodal Speaker Identification Using Dynamic Bayesian Network.- A Novel Pitch Period Detection Algorithm Based on Hilbert-Huang Transform.- Noisy Speech Pitch Detection Based on Mathematical Morphology and Weighted MACF.- Glottal Information Based Spectral Recuperation in Multi-channel Speaker Recognition.- Speaker Modeling Technique Based on Regression Class for Speaker Identification with Sparse Training.- Other Biometrics.- Some Issues Pertaining to Adaptive Multimodal Biometric Authentication.- Protecting Biometric Data for Personal Identification.- Digital Curvelet Transform for Palmprint Recognition.- On-line Writer Verification Using Force Features of Basic Strokes.- A Novel Force Sensitive Tablet for Handwriting Information Acquisition.- Shape and Structural Feature Based Ear Recognition.- LLE Based Gait Analysis and Recognition.- Personal Identification Using Knuckleprint.- AAM Based Matching of Hand Appearance for User Verification.

Optimal shape space and searching in ASM based face alignment

Abstract: The Active Shape Models (ASM) is composed of two parts: the ASM shape model and the ASM search The standard ASM, with the shape variance components all discarded and searching in image subspace and shape subspace independently, has blind searching and unstable search result In this paper, we propose a novel idea, called Optimal Shape Subspace, for optimizing ASM search It is constructed by both main shape and shape variance information It allows the reconstructed shape to vary more than that reconstructed in the standard ASM shape space, hence is more expressive in representing shapes in real life A cost function is developed, based on a careful study on the search process especially regarding relations between the ASM shape model and the ASM search An Optimal Searching method using the feedback provided by the evaluation cost can significantly improve the performance of ASM alignment This is demonstrated by experimental results.