Face recognition using spectral features

Face recognition is a challenging task in computer vision and pattern recognition. It is well-known that obtaining a low-dimensional feature representation with enhanced discriminatory power is of paramount importance to face recognition. Moreover, recent research has shown that the face images reside on a possibly nonlinear manifold. Thus, how to effectively exploit the hidden structure is a key problem that significantly affects the recognition results. In this paper, we propose a new unsupervised nonlinear feature extraction method called spectral feature analysis (SFA). The main advantages of SFA over traditional feature extraction methods are: (1) SFA does not suffer from the small-sample-size problem; (2) SFA can extract discriminatory information from the data, and we show that linear discriminant analysis can be subsumed under the SFA framework; (3) SFA can effectively discover the nonlinear structure hidden in the data. These appealing properties make SFA very suitable for face recognition tasks. Experimental results on three benchmark face databases illustrate the superiority of SFA over traditional methods.     

Adaptive correlation filters for pattern recognition

Adaptive correlation filters based on synthetic discriminant functions (SDFs) for reliable pattern recognition are proposed. A given value of discrimination capability can be achieved by adapting a SDF filter to the input scene. This can be done by iterative training. Computer simulation results obtained with the proposed filters are compared with those of various correlation filters in terms of recognition performance. The text was submitted by the authors in English. Vitaly Kober obtained his MS degree in Applied Mathematics from the Air-Space University of Samara (Russia) in 1984 and his PhD degree in 1992 and Doctor of Sciences degree in 2004 in Image Processing from the Institute of Information Transmission Problems, Russian Academy of Sciences. He is now a titular researcher at the Centro de Investigatión Cientifica y de Educatión Superior de Ensenada (Cicese), Mexico. His research interests include signal and image processing and pattern recognition. Mikhail Mozerov received his MS degree in Physics from Moscow State University in 1982 and his PhD degree in Image Processing from the Institute of Information Transmission Problems, Russian Academy of Sciences, in 1995. He is with the Laboratory of Digital Optics of the Institute of Information Transmission Problems, Russian Academy of Sciences. His research interests include signal and image processing, pattern recognition, and digital holography. Iosif A. Ovseyevich graduated from the Moscow Electrotechnical Institute of Telecommunications. He received his candidate’s degree in 1953 and doctoral degree in Information Theory in 1972. At present he is Emeritus Professor at the Institute of Information Transmission Problems, Russian Academy of Sciences. His research interests include information theory, signal processing, and expert systems. He is a Member of the IEEE and Popov Radio Society.     

Pattern Recognition Theory in Nonlinear Signal Processing

A body of research has developed within the context of nonlinear signal and image processing that deals with the automatic, statistical design of digital window-based filters. Based on pairs of ideal and observed signals, a filter is designed in an effort to minimize the error between the ideal and filtered signals. The goodness of an optimal filter depends on the relation between the ideal and observed signals, but the goodness of a designed filter also depends on the amount of sample data from which it is designed. In order to lessen the design cost, a filter is often chosen from a given class of filters, thereby constraining the optimization and increasing the error of the optimal filter. To a great extent, the problem of filter design concerns striking the correct balance between the degree of constraint and the design cost. From a different perspective and in a different context, the problem of constraint versus sample size has been a major focus of study within the theory of pattern recognition. This paper discusses the design problem for nonlinear signal processing, shows how the issue naturally transitions into pattern recognition, and then provides a review of salient related pattern-recognition theory. In particular, it discusses classification rules, constrained classification, the Vapnik-Chervonenkis theory, and implications of that theory for morphological classifiers and neural networks. The paper closes by discussing some design approaches developed for nonlinear signal processing, and how the nature of these naturally lead to a decomposition of the error of a designed filter into a sum of the following components: the Bayes error of the unconstrained optimal filter, the cost of constraint, the cost of reducing complexity by compressing the original signal distribution, the design cost, and the contribution of prior knowledge to a decrease in the error. The main purpose of the paper is to present fundamental principles of pattern recognition theory within the framework of active research in nonlinear signal processing.     

Model based odia numeral recognition using fuzzy aggregated features

In this paper, an efficient scheme for recognition of handwritten Odia numerals using hidden markov model （HMM） has been proposed. Three different feature vectors for each of the numeral is generated through a polygonal approximation of object contour. Subsequently, aggregated feature vector for each numeral is derived from these three primary feature vectors using a fuzzy inference system. The final feature vector is divided into three levels and interpreted as three different states for HMM. Ten different three-state ergodic hidden markov models （HMMs） are thus constructed corresponding to ten numeral classes and parameters are calculated from these models. For the recognition of a probe numeral, its log-likelihood against these models are computed to decide its class label. The proposed scheme is implemented on a dataset of 2500 handwritten samples and a recognition accuracy of 96.3% has been achieved. The scheme is compared with other competent schemes.     

2-D invariant object recognition using distributed associativememory

Complex-log conformal mapping is combined with a distributed associative memory to create a system that recognizes objects regardless of changes in rotation or scale. Information recalled from the memorized database is used to classify an object, reconstruct the memorized version of the object, and estimate the magnitude of changes in scale or rotation. The system response is resistant to moderate amounts of noise and occlusion. Several experiments using real gray-scale images are presented to show the feasibility of the approach     

Learning optimized features for hierarchical models of invariant object recognition

There is an ongoing debate over the capabilities of hierarchical neural feedforward architectures for performing real-world invariant object recognition. Although a variety of hierarchical models exists, appropriate supervised and unsupervised learning methods are still an issue of intense research. We propose a feedforward model for recognition that shares components like weight sharing, pooling stages, and competitive nonlinearities with earlier approaches but focuses on new methods for learning optimal feature-detecting cells in intermediate stages of the hierarchical network. We show that principles of sparse coding, which were previously mostly applied to the initial feature detection stages, can also be employed to obtain optimized intermediate complex features. We suggest a new approach to optimize the learning of sparse features under the constraints of a weight-sharing or convolutional architecture that uses pooling operations to achieve gradual invariance in the feature hierarchy. The approach explicitly enforces symmetry constraints like translation invariance on the feature set. This leads to a dimension reduction in the search space of optimal features and allows determining more efficiently the basis representatives, which achieve a sparse decomposition of the input. We analyze the quality of the learned feature representation by investigating the recognition performance of the resulting hierarchical network on object and face databases. We show that a hierarchy with features learned on a single object data set can also be applied to face recognition without parameter changes and is competitive with other recent machine learning recognition approaches. To investigate the effect of the interplay between sparse coding and processing nonlinearities, we also consider alternative feedforward pooling nonlinearities such as presynaptic maximum selection and sum-of-squares integration. The comparison shows that a combination of strong competitive nonlinearities with sparse coding offers the best recognition performance in the difficult scenario of segmentation-free recognition in cluttered surround. We demonstrate that for both learning and recognition, a precise segmentation of the objects is not necessary.     

Three-dimensional object recognition using laser range sensor

Image information provided by cameras is strongly affected by environmental influence of an object’s circumjacent and circumference. In order to reduce environmental influence, a system which was integrated distance information provided from a laser range sensor (LRS) and image information provided by a camera was developed, and consisted of an object extraction section and a recognition processing section. In this paper the effectiveness of the system was inspected by performing an object extraction experiment using the combined integrated distance information and image information. From these results, this system could remove a background and a floor surface by using the distance information, and simpler object extraction was enabled. This work was presented in part at the 13th International Symposium on Artificial Life and Robotics, Oita, Japan, January 31–February 2, 2008     

Subband entropy-based features for clothing invariant human gait recognition

This paper presents a wavelet-based feature extraction method for human gait recognition. The selection of features with most discriminative information is the key to improve recognition performance. The frequency domain representation of the gait image is obtained by using fast Fourier transforms. Next, a discrete wavelet transform is applied to the obtained spectrum. With single-level wavelet decomposition, four coefficients are generated. The sum of the entropy of these four wavelet coefficients is computed yielding the wavelet Entropy Image (wEnI) which is used here as the potential feature for human gait recognition. A template matching-based approach is used as the classification. The performance of the proposed wEnI feature is evaluated using whole-based and part-based methods. The experimental results show that the wEnI feature performs better compared to state-of-the-art gait features in common use.     

Robust 3D face recognition based on resolution invariant features

A novel resolution invariant local feature based method is proposed for 3D face recognition. Scale space extrema on shape index images and texture images are detected and matched, through which resolution and noise insensitive face matching is achieved without complex preprocessing and normalization. An outlier removal strategy is designed to eliminate incorrect matching points while keeping relevant ones. Six different scale invariant similarity measures are proposed and fused at the score level, which increases the robustness against expression variations. Systematical experiments are conducted on the FRGC v2.0 database, achieving in the neutral vs. all experiment a verification rate of 90.7% with un-normalized similarity scores, and 96.3% with normalized similarity scores at False Acceptance Rate (FAR) of 0.1%, and 96.2% rank-1 identification rate, which are comparable to the state of the art, and promising considering the significantly reduced preprocessing requirement.     

A higher-order neural network for distortion invariant pattern recognition

Recently, it is shown that a single layer, higher-order neural network is effective for scale, rotation and shift invariance and in the training process it requires only one example for one category and a very small number of iterations. However, there are problems that scale invariance doesn't hold precisely and it is not so effective for distortion of unknown patterns. In this paper we present an idea to realize the scale invariance precisely and suggest a method that is available to distorted patterns. The experimental results are presented to show the feasibility of our approach.     

Deformation invariant visual object recognition: Experiments with a self-organising neural architecture

A self-organising neural network architecture for grey-scale visual object rcognition is presented. The network is composed of three processing layers with an architecture designed to give deformation tolerance. The processing layers involve feature extraction, sub-pattern detection and classification. Training is generally performed on-line in an unsupervised manner, classes being created when objects are presented that cannot be classified. The results given show the effect of the two discrimination parameters when the network is applied to two very different sets of images, namely hand written numerals and hand gestures images. The sensitivity of the network to the parameters that govern the size of detectable patterns and the areas over which they are detected is also tested. The robustness of the network to the order of image presentation is also demonstrated. The results show that parameter choice is not critical and heuristically chosen parameters provide near optimum performance.     

View invariant action recognition using weighted fundamental ratios

In this paper, we fully investigate the concept of fundamental ratios, demonstrate their application and significance in view-invariant action recognition, and explore the importance of different body parts in action recognition. A moving plane observed by a fixed camera induces a fundamental matrix F between two frames, where the ratios among the elements in the upper left 2 × 2 submatrix are herein referred to as the fundamental ratios. We show that fundamental ratios are invariant to camera internal parameters and orientation, and hence can be used to identify similar motions of line segments from varying viewpoints. By representing the human body as a set of points, we decompose a body posture into a set of line segments. The similarity between two actions is therefore measured by the motion of line segments and hence by their associated fundamental ratios. We further investigate to what extent a body part plays a role in recognition of different actions and propose a generic method of assigning weights to different body points. Experiments are performed on three categories of data: the controlled CMU MoCap dataset, the partially controlled IXMAS data, and the more challenging uncontrolled UCF-CIL dataset collected on the internet. Extensive experiments are reported on testing (i) view-invariance, (ii) robustness to noisy localization of body points, (iii) effect of assigning different weights to different body points, (iv) effect of partial occlusion on recognition accuracy, and (v) determining how soon our method recognizes an action correctly from the starting point of the query video.     

Efficient 3D object recognition using foveated point clouds

Recent hardware technologies have enabled acquisition of 3D point clouds from real world scenes in real time. A variety of interactive applications with the 3D world can be developed on top of this new technological scenario. However, a main problem that still remains is that most processing techniques for such 3D point clouds are computationally intensive, requiring optimized approaches to handle such images, especially when real time performance is required. As a possible solution, we propose the use of a 3D moving fovea based on a multiresolution technique that processes parts of the acquired scene using multiple levels of resolution. Such approach can be used to identify objects in point clouds with efficient timing. Experiments show that the use of the moving fovea shows a seven fold performance gain in processing time while keeping 91.6% of true recognition rate in comparison with state-of-the-art 3D object recognition methods.     

Visual object recognition using probabilistic kernel subspace similarity

Probabilistic subspace similarity-based face matching is an efficient face recognition algorithm proposed by Moghaddam et al. It makes one basic assumption: the intra-class face image set spans a linear space. However, there are yet no rational geometric interpretations of the similarity under that assumption. This paper investigates two subjects. First, we present one interpretation of the intra-class linear subspace assumption from the perspective of manifold analysis, and thus discover the geometric nature of the similarity. Second, we also note that the linear subspace assumption does not hold in some cases, and generalize it to nonlinear cases by introducing kernel tricks. The proposed model is named probabilistic kernel subspace similarity (PKSS). Experiments on synthetic data and real visual object recognition tasks show that PKSS can achieve promising performance, and outperform many other current popular object recognition algorithms.     

A unifying framework for invariant pattern recognition

We introduce a group-theoretic model of invariant pattern recognition, the Group Representation Network. We show that many standard invariance techniques can be viewed as GRNs, including the DFT power spectrum, higher order neural network and fast translation-invariant transform.     

基于角度统计特征的步态识别

提取角度统计特征,为步态识别提出了一种新途径.用统计的方法,等角度间隔地计算归一化步态轮廓图像各像素点至质心距离的均值与方差,并用其构造步态识别的特征向量.与提取步态轮廓图像边界特征的方法相比,该方法具有算法简单、运算速度快、无需建立复杂的数学模型等优点.以Matlab7.5为平台,以中科院自动化研究所提供的CASIA数据库为样本进行了大量实验,实验结果表明:该步态识别方法具有较好的识别性能.