This paper introduces a dual-task recognition system that uses Wi-Fi channel state information (CSI) for human Activity Recognition (AR) and Indoor Localization (IL), effectively addressing the limitations of single-task recognition approaches. The proposed system leverages complementary information derived from dual-stream signal characteristics by capitalizing on the relationship between the two tasks. Moreover, it incorporates a technique to mitigate sampling irregularities, effectively reducing data misrepresentation. Extensive experimental evaluations on two datasets substantiate the efficacy of the proposed method in managing dual-task recognition scenarios, consistently outperforming specialized single-task recognition solutions. These results highlight its potential as a versatile solution for Wi-Fi CSI-based recognition applications, including IoT monitoring, surveillance, and healthcare.

In this paper, we construct a network for human activity recognition based on the tokenized Wi-Fi signals on an attention mechanism. After standardizing the signals, the WiFi channel state information is utilized as a set of time-series data, acknowledging its inherent temporal structure. Motivated by the Transformer’s ability to model temporal dependencies, the construction is enriched with a frequency-based tokenization scheme. This unique construction is adept at managing noise and sensitivity intrinsic to Wi-Fi signals, effectively mitigating the challenges in Wi-Fi-based human activity recognition. Our experimental evaluations validated the effectiveness of the proposed structure.

This study introduces an innovative Human Activity Recognition system using Wi-Fi Channel State Information (CSI). Our modified GRU input enhances representation capability by integrating the past input information. The system employs three data augmentation techniques— Adding Gaussian Noise, Data Shifting, and CutMix— to expand the dataset and introduce variability for overfitting handling. We conducted three key experiments to validate our model's performance: optimizing hyperparameters, performing ablation study to assess each technique's impact, and comparing our model with state-of-the-art models.11The implemented codes are available in https://github.com/GRUwithAugmentedData

In this paper, we propose a fusion network for human activity recognition based on the Wi-Fi Channel State Information (CSI) signals. The system employs a Bidirectional Long Short-Term Memory (BiLSTM) layer to extract action features from CSI data blocks and then trains them using the Kernel Ridge Regression (KRR). The trained block responses are subsequently fused based on the sum-rule to form the final decision. In contrast to deep learning, this process is computationally efficient because there is no need to train the BiLSTM. The proposed method has been tested on two publicly available databases to validate the accuracy performance.

In this paper, we propose to extract the intersection points of the palmprint and the palm-vein lines from multi-spectral images and use them as reliable features for identity verification. Essentially, by utilizing a sum of cardinal directional image difference operation, the palmprint and palm-vein line features are respectively extracted from palm images of the Blue channel and the NIR channel of image spectrums based on simple matrix projection. Subsequently, the intersection locations of the two biometric line features are extracted and utilized to compute a set of keypoint descriptors. After calculating the match scores based on the extracted keypoint descriptors, a score level fusion of the matching results obtained from the Blue channel and the NIR channel is adopted to enhance the verification performance. The proposed method has been experimented on a public domain multispectral palm database where encouraging results in terms of verification accuracy have been obtained.

Class-incremental learning (CIL) learns a classification model with training data of different classes arising progressively. Existing CIL either suffers from serious accuracy loss due to catastrophic forgetting, or invades data privacy by revisiting used exemplars. Inspired by learning of linear problems, we propose an analytic class-incremental learning (ACIL) with absolute memorization of past knowledge while avoiding breaching of data privacy (i.e., without storing historical data). The absolute memorization is demonstrated in the sense that the CIL using ACIL given present data would give identical results to that from its joint-learning counterpart that consumes both present and historical samples. This equality is theoretically validated. The data privacy is ensured by showing that no historical data are involved during the learning process. Empirical validations demonstrate ACIL's competitive accuracy performance with near-identical results for various incremental task settings (e.g., 5-50 phases). This also allows ACIL to outperform the state-of-the-art methods for large-phase scenarios (e.g., 25 and 50 phases).

Training neural networks with the Moore–Penrose (MP) inverse has recently gained attention in view of its noniterative training nature. However, a significant drawback of learning based on the MP inverse is that the computational memory consumption grows along with the size of a dataset. In this article, based on the partitioning of the MP inverse, we propose a blockwise recursive MP inverse formulation (BRMP) for network learning with low-memory property while preserving its training effectiveness. The BRMP is an equivalent formulation to its batchwise counterpart since neither approximation nor assumption is made in the derivation process. Our further exploration of this recursive method leads to a switching structure among three different scenarios. This structure also reveals that the well-known recursive least squares method is a special case of our proposed technique. Subsequently, we apply BRMP to the training of radial basis function networks as well as multilayer perceptrons. The experimental validation covers both regression and classification tasks.

This paper proposes a kernel projection (KP) neural network that analytically determines its network parameters. The proposed network is composed of cascaded modules of 2-layer sub-networks. A technique which encodes the label information into each module has been introduced to enable a locally supervised learning. Such a supervised learning in the 2-layer module begins with a kernel projection in the first layer and determines its parameters analytically via solving a least squares problem in the second layer. We show that the analytic nature of the proposed network allows a learning process significantly faster than that of the traditional backpropagation method as it only needs to visit the dataset once. Experiments of classification tasks on various datasets are carried out, showing comparable or better results compared with several competing methods.

Pedestrian detection in the embedded system, such as video surveillance equipment, usually involves low-resolution pedestrian samples and requires a low computational cost. Many pedestrian detectors rely on a large feature pool and suffer in their efficiency and performance for real-time monitoring. In this paper, a set of light-weight features is proposed to enhance the pedestrian detection performance when a small-medium scale of training data with low-resolution images is available. To address this issue, a difference matrix projection (DMP) is developed to compute aggregated multi-oriented pixel differences using global matrix operations. Both the pixel differences and aggregation are computed using global matrix projection to avoid the laborious iterative operations. We tested our method on the INRIA, Daimler Chrysler classification (Daimler-CB), NICTA, and Caltech Pedestrian datasets. The experiments on these benchmark data sets show encouraging results in terms of detection performance, particularly for image datasets with low-resolution pedestrians.

In this paper, we propose a recursive learning formulation for online heterogeneous face recognition (HFR). The main task is to compare between images which are acquired from different sensing spectrums for identity recognition. Using an extreme learning machine, the proposed recursive formulation seeks a direct optimization to the classification error goal where the solution converges exactly to the batch mode solution. Due to the nonlinear nature of the classification error objective function, formulation of a recursive solution that converges is an important and nontrivial task. Based on this recursive formulation, an online HFR system is designed. The system is evaluated using two challenging heterogeneous face databases with images captured under visible, near infrared and infrared spectrums. The proposed system shows promising performance which is comparable with that of competing state-of-the-arts.

In this paper, we propose a system for identity verification based on the gesture signals of handwritten signature captured by the Wi-Fi CSI wave packets at different positions using transfer learning. Essentially, a ConvNet is first pretrained using the Wi-Fi signature signals collected from one position. Subsequently, the pretrained feature extractor is transferred to recognize signals collected from another position via a rapid retraining process. We utilize the kernel and the range space projection learning when we retrain the transferred model. Our experimental results on an in-house Wi-Fi handwritten signature signal dataset show that the signature signals from the new position can be effectively classified without needing to retrain the model from scratch.

Recently, a learning method based on the kernel and the range space projections has been introduced. This method has been applied to learn the multilayer network analytically with interpretable relationships among the weight matrices. However, the learning method carries a high-memory demand during training. In this study, a low-memory formulation is proposed to address this issue of high-memory demand. The developed method is inspired by a recursive implementation of the Moore-Penrose inverse and is shown to be mathematically equivalent to the original batch learning. Next, we further improved our proposed low-memory formulation to annul the potential divergence caused by rounding errors. The regression and classification behaviors of the proposed learning method are demonstrated using both synthetic and benchmark datasets. Our experiments confirm that the proposed formulation consumes significantly lower memory.

Pattern recognition algorithms have recently been pursued for automatic analysis of delayered IC images, i.e. the detection of circuit components. Wide experimentation on the existing training-based approaches are hampered by heavy data labeling, expensive model training, or long processing time. In this paper, we propose a global template projection and matching (GTPM) method that requires no training and a minimal amount of data labeling for circuit component detection. Our proposed GTPM method achieves a higher or comparable accuracy as the reported approaches while being more computationally efficient.

In this study, the authors propose an efficient extension of the standard empirical mode decomposition (EMD) for complex‐valued univariate signal decomposition. The key idea of the extension is to convert a complex‐valued univariate signal into a longer real‐valued signal by augmenting the real part with the flipped imaginary part, and then to decompose it into intrinsic mode functions (IMFs) using the EMD once only. The bivariate IMFs are then retrieved from the obtained IMFs. Their empirical results on synthetic data show that the proposed method significantly outperforms the traditional bivariate EMD (BEMD) method in terms of computational efficiency while producing a comparable extraction error. Moreover, the proposed method shows better micro‐Doppler signature analysis performance on physically measured continuous‐wave radar data than that of the BEMD.

In this article, a novel approach to learning a complex function which can be written as the system of linear equations is introduced. This learning is grounded on the observation that solving the system of linear equations by a manipulation in the kernel and the range space boils down to an estimation based on the least squares error approximation. The learning approach is applied to learn a deep feedforward network with full weight connections. The numerical experiments on network learning of synthetic and benchmark data not only show feasibility of the proposed learning approach but also provide insights regarding the curve fitting mechanism.

In this paper, we first investigate into six popular entropies extracted from a set of intrinsic mode functions (IMFs) as a feature pattern for radar-based mini-size unmanned aerial vehicles (mini-UAV) classification. The six entropies include Shannon entropy, spectral entropy, log energy entropy, approximate entropy, fuzzy entropy and permutation entropy. Via an empirical comparison among the six entropies on real measurement radar data, the first three are selected as the representative due to their high efficiency and accuracy. To enhance the classification accuracy, the three selected entropies are then extracted from eight different sets of IMFs obtained by signal downsampling, and then fused at feature level. The nonlinear support vector machine classifier is adopted to predict the class label of unseen test radar signals. Our empirical results on a set of real-world continuous wave radar data show that the proposed method outperforms the state-of-the-art method in terms of the mini-UAV classification accuracy.

In this letter, we propose an empirical-mode decomposition (EMD)-based method for automatic multicategory mini-unmanned aerial vehicle (UAV) classification. The radar echo signal is first decomposed into a set of oscillating waveforms by EMD. Then, eight statistical and geometrical features are extracted from the oscillating waveforms to capture the phenomenon of blade flashes. After feature normalization and fusion, a nonlinear support vector machine is trained for target class-label prediction. Our empirical results on real measurement of radar signals show encouraging mini-UAV classification accuracy performance.

Gabor magnitude is known to be among the most discriminative representations for face images due to its space- frequency co-localization property. However, such property causes adverse effects even when the images are acquired under moderate head pose variations. To address this pose sensitivity issue and other moderate imaging variations, we propose an analytic Gabor feedforward network which can absorb such moderate changes. Essentially, the network works directly on the raw face images and produces directionally projected Gabor magnitude features at the hidden layer. Subsequently, several sets of magnitude features obtained from various orientations and scales are fused at the output layer for final classification decision. The network model is analytically trained using a single sample per identity. The obtained solution is globally optimal with respect to the classification total error rate. Our empirical experiments conducted on five face data sets (six subsets) from the public domain show encouraging results in terms of identification accuracy and computational efficiency.

This paper proposes a method based on the bag-of-words (BoW) and the softmax regression for microscopic image classification. Essentially, the locality-constrained linear coding (LLC) is adopted for local feature encoding. Compared with the traditionally adopted vector quantization (VQ) in the BoW framework, the LLC encodes local structures of microscopic images with lower quantization errors and generates a sparse image representation. This enables the use of linear classifiers with low computational complexity. A softmax regression classifier is then adopted to address the multi-categorical classification task where the confidence of categorical prediction is quantified by posterior probabilities. Compared with other linear classifiers (such as the linear SVM) which only assign labels to images, such probabilistic outputs provide extra quantitative information to analyze misclassified images. Our experiments on the 2D-Hela and the PAP smear data sets show significant performance improvement of the proposed method comparing with competing methods using different features and classifiers under the BoW framework.

The principal component analysis network (PCANet) is asserted as a parsimonious stacking-based convolutional neural networks (CNNs) instance for generic object recognition including face. However, to be regarded a CNN resemblance, PCANet lacks a nonlinearity in between two successive convolutional layers. The multilayer PCANet (by neglecting the nonlinearity pre-requisite) is also deemed far-fetched for the network depth beyond two, due to feature dimensionality explosion. We thus devise a PCANet alternative, dubbed PCANet+ in this letter, to untangle these constraints. To be more precise, conforming to the CNN essentials, PCANet+ conveys a mean-pooling unit manipulating each feature map. On top of that, we streamline the PCANet topology to permit a deep construction with an expanded PCA filter ensemble. We scrutinize the PCANet+ performance using face recognition technology and other two faces in the wild datasets, namely, labeled faces in the wild and YouTube faces. The experimental results reveal that the PCANet+ descriptor prevails over its predecessor and other stacking-based descriptors in face identification and verification, serving a baseline for ConvNets.

Based on the knowledge that input data distribution is important for learning, a data density-dependent quantization scheme (DQS) is proposed for sparse input data representation. The usefulness of the representation scheme is demonstrated by using it as a data preprocessing unit attached to the well-known least squares support vector machine (LS-SVM) for application on big data sets. Essentially, the proposed DQS adopts a single shrinkage threshold to obtain a simple quantization scheme, which adapts its outputs to input data density. With this quantization scheme, a large data set is quantized to a small subset where considerable sample size reduction is generally obtained. In particular, the sample size reduction can save significant computational cost when using the quantized subset for feature approximation via the Nyström method. Based on the quantized subset, the approximated features are incorporated into LS-SVM to develop a data density-dependent quantized LS-SVM (DQLS-SVM), where an analytic solution is obtained in the primal solution space. The developed DQLS-SVM is evaluated on synthetic and benchmark data with particular emphasis on large data sets. Extensive experimental results show that the learning machine incorporating DQS attains not only high computational efficiency but also good generalization performance.

Independent component analysis (ICA) by an information measure has seen wide applications in engineering. Different from traditional probability density function based information measures, a probability survival distribution based Cauchy-Schwartz information measure for multiple variables is proposed in this paper. Empirical estimation of survival distribution is parameter-free which is inherited by the estimation of the new information measure. This measure is proved to be a valid statistical independence measure and is adopted as an objective function to develop an ICA algorithm which is validated by an experiment. This work shows promising potential regarding the use of survival distribution based information measure for ICA.

In this paper, a single hidden-layer feedforward fusion network is proposed for face identity verification. Essentially, the feature extraction, matching score calculation and fusion algorithm design steps are integrated and absorbed into a hidden layer of the model. Each hidden node works on the raw face image directly and produces an Euclidean distance based match score within the network. These scores are then incorporated with output weights to produce a fused score at the final stage. Our experimental study conducted using three face databases shows that the proposed model consistently outperforms competing methods.

A weighted least squares scheme based on an empirical survival error potential function is proposed in this paper. The empirical survival error potential function provides an error compensation scheme for noise distributions far from being Gaussian. This error compensation procedure is efficiently implemented via a weighted least squares formulation where an analytical solution form is obtained. The performance of the developed scheme is extensively tested on 16 benchmark data sets where the results show promising potential of the proposed empirical survival error distribution compensation scheme for binary pattern classification.

This paper proposes a novel solution for compressive polynomial regression learning. The solution comes in primal and dual closed-forms similar to that of ridge regression. Essentially, the proposed solution stretches the covariance computation by a power term thereby compresses or amplifies the estimation. Our experiments on both synthetic data and real-world data show effectiveness of the proposed method for compressive learning.

Gabor-based region covariance matrix (GRCM) is an emerging face feature descriptor, which has been shown promising for face recognition. The GRCM lies on Tensor manifold is inherently non-Euclidean, hence a disconnect exists between GRCM descriptor and vector-based classifiers, such as collaborative representation-based classifier (CRC). CRC is a strong alternative to sparse representation-based classifier yet enjoys high efficiency. In this paper, we bridge GRCM and CRC with kernel learning method. We investigate several geodesic distances on Tensor manifold that satisfy the Mercer's condition for kernel CRC construction as well as for speedy computation. Apart from that, we also devise two strategies to jointly combine the regionalized GRCMs with Tensor kernel CRC. Extensive experiments on the ORL and FERET datasets are conducted to verify the efficacy of the proposed method.

Biometric discretization is a key component in biometric cryptographic key generation. It converts an extracted biometric feature vector into a binary string via typical steps such as segmentation of each feature element into a number of labeled intervals, mapping of each interval-captured feature element onto a binary space, and concatenation of the resulted binary output of all feature elements into a binary string. Currently, the detection rate optimized bit allocation (DROBA) scheme is one of the most effective biometric discretization schemes in terms of its capability to assign binary bits dynamically to user-specific features with respect to their discriminability. However, we learn that DROBA suffers from potential discriminative feature misdetection and underdiscretization in its bit allocation process. This paper highlights such drawbacks and improves upon DROBA based on a novel two-stage algorithm: 1) a dynamic search method to efficiently recapture such misdetected features and to optimize the bit allocation of underdiscretized features and 2) a genuine interval concealment technique to alleviate crucial information leakage resulted from the dynamic search. Improvements in classification accuracy on two popular face data sets vindicate the feasibility of our approach compared with DROBA.

In this paper, we propose a user authentication system based on hand-gesture signature without the need of any handheld device. The system uses a depth image sensor to locate the fingertip and palm mass-center from detected hand region for trajectory processing. Apart from the positional information, the velocity and acceleration information are included as input features for trajectory matching. The system verification performance is evaluated in terms of equal error rate. In addition, an investigation of fusion at feature level is conducted for possible performance enhancement. Our empirical results show the potential of the proposed bare-hand in-air signature system.

This work proposes a structured random projection via feature weighting for cancelable identity verification. Essentially, projected facial features are weighted based on their discrimination capability prior to a matching process. In order to conceal the face identity, an averaging over several templates with different transformations is performed. Finally, several cancelable templates extracted from partial face images are fused at score level via a total error rate minimization. Our empirical experiments on two experimental scenarios using AR, FERET' and Sheffield databases show that the proposed method consistently outperforms competing state-of-the-art unsupervised methods in terms of verification accuracy.

Abstract This paper presents a two‐stage procedure to combine multiple face traits for identity authentication. At the first stage, a high dimensional random projection is applied to the raw visual and infrared face images to extract useful information relevant to each identity. This is followed by a dimension reduction using eigenfeature regularization and extraction (ERE). At the second stage, the scores from two verification systems based on each face modality are fused by an error minimization algorithm. This error minimization algorithm directly optimizes the verification accuracy by adjusting the parameters of a polynomial classifier. Two data sets consisting of visual and infrared face images have been used for experimentation. Our empirical observation shows encouraging results regarding the effectiveness of the proposed method. Copyright © 2011 John Wiley & Sons, Ltd.

This paper provides an editorial statement for the special issue on biometric security in mobile computing environment.

This paper presents a projection framework for biométrie scores fusion. Essentially, the framework consists of a projection stage and a learning stage. Apart from investigating into several relatively new projection models for biométrie fusion, the projection stage attempts to unify these models into a single parametric structure. Three learning methods are investigated in conjunction with six projection models for their impacts on verification accuracy expressed in terms of equal error rate. An extensive experiment of these model and learning combinations on 32 fusion data sets are performed in the evaluation.

Face recognition has many useful applications spanning surveillance, law enforcement, information security, smart card and entertainment technologies. Very recently, a learning based face recognition system is also seen to be applied to web platform combining face recognition and web service. However, many existing methods which focused on recognition accuracy cannot cope with the new social network platform because the adopted static learning approach is not adaptive to daily updated photographs among the massive number of users. In this paper, we discuss the difference between a stand-alone based system and a social network based system and propose a new collaborative face recognition framework where a redundant tagging can be avoided via sharing the identification information for efficient update under the social network platform. Our Experiments (including a web stress test) using a public database show that the proposed method records a better accuracy than that of the state-of-the-art classifier SVM adopting a polynomial kernel and has fast execution time for both training and testing.

This paper presents a score level fusion of visual and infrared face image verification systems. A high dimensional random projection is first applied to the raw visual and infrared face images to extract useful information relevant to each identity. This is followed by a dimension reduction using eigenfeature regularization and extraction. The resultant templates are then compared for decision scores generation. Finally the scores from the visual and infrared face image verification systems are fused by an error rate minimization formulation. Our empirical observation shows encouraging results regarding the effectiveness of the fusion.

In this paper, we propose a CIECAM02-based color image enhancement method which is particularly robust to scenes with bright surrounding. The proposed method detects color edges using a distance metric based on the characteristics of a human visual system (HVS). The CIECAM02 is inherently strong considering both HVS and surrounding conditions. A major problem of HVS is that the dark region appears darker under a bright surrounding condition, leading to masking of details within the dark region. This phenomenon causes deterioration of edges which are among the most important and sensitive components for the HVS. To overcome this problem, we propose to weight the deteriorated edges at bright scenes. Adaptively, we estimate the surrounding image by the CIECAM02, and then use a vector gradient edge detector with a newly proposed distance metric to perform the weighting. The proposed method is seen to enhance edges without introducing unwanted color artifacts. We subjectively confirm the performance with clearly enhanced images.

This letter presents a minimum classification error learning formulation for a single-layer feedforward network (SLFN). By approximating the nonlinear counting step function using a quadratic function, the classification error rate is shown to be deterministically solvable. Essentially the derived solution is related to an existing weighted least-squares method with class-specific weights set according to the size of data set. By considering the class-specific weights as adjustable parameters, the learning formulation extends the classification robustness of the SLFN without sacrificing its intrinsic advantage of being a closed-form algorithm. While the method is applicable to other linear formulations, our empirical results indicate SLFN's effectiveness on classification generalization.

This paper presents a deterministic solution to an approximated classification-error based objective function. In the formulation, we propose a quadratic approximation as the function for achieving smooth error counting. The solution is subsequently found to be related to the weighted least-squares whereby a robust tuning process can be incorporated. The tuning traverses between the least-squares estimate and the approximated total-error-rate estimate to cater for various situations of unbalanced attribute distributions. By adopting a linear parametric classifier model, the proposed classification-error based learning formulation is empirically shown to be superior to that using the original least-squares-error cost function. Finally, it will be seen that the performance of the proposed formulation is comparable to other classification-error based and state-of-the-art classifiers without sacrificing the computational simplicity.

A real-time vision system operating at an outdoor swimming pool is presented in this paper. The system is designed to automatically recognize different swimming activities and to detect occurrence of early drowning incidents. We have named this system the Drowning Early Warning System (DEWS). One key challenge we faced in the problem is the relatively high level of noise in the steps of foreground detection and behavior recognition. Therefore, a set of methods in the fields of background subtraction, denoising, data fusion and blob splitting are proposed, which have been motivated by characteristics of aquatic background and crowded scenario at the pool. In the step to detect an early drowning incident, visual indicators of distress and drowning are incorporated through a set of foreground descriptors. A module comprising data fusion and hidden Markov modeling is developed to learn unique traits of different swimming behaviors, in particular, those early drowning events. The experiment of this work reports realistic on-site evaluations performed. Examples of interesting behaviors, i.e., distress, drowning, treading and numerous swimming styles, are simulated and collected. Experimental results show that we have established a prototype system which is robust and beyond the stage of proof-of-concept.

To replace compromised biometric templates, cancelable biometrics has recently been introduced. The concept is to transform a biometric signal or feature into a new one for enrollment and matching. For making cancelable fingerprint templates, previous approaches used either the relative position of a minutia to a core point or the absolute position of a minutia in a given fingerprint image. Thus, a query fingerprint is required to be accurately aligned to the enrolled fingerprint in order to obtain identically transformed minutiae. In this paper, we propose a new method for making cancelable fingerprint templates that do not require alignment. For each minutia, a rotation and translation invariant value is computed from the orientation information of neighboring local regions around the minutia. The invariant value is used as the input to two changing functions that output two values for the translational and rotational movements of the original minutia, respectively, in the cancelable template. When a template is compromised, it is replaced by a new one generated by different changing functions. Our approach preserves the original geometric relationships (translation and rotation) between the enrolled and query templates after they are transformed. Therefore, the transformed templates can be used to verify a person without requiring alignment of the input fingerprint images. In our experiments, we evaluated the proposed method in terms of two criteria: performance and changeability. When evaluating the performance, we examined how verification accuracy varied as the transformed templates were used for matching. When evaluating the changeability, we measured the dissimilarities between the original and transformed templates, and between two differently transformed templates, which were obtained from the same original fingerprint. The experimental results show that the two criteria mutually affect each other and can be controlled by varying the control parameters of the changing functions.

By exploiting the specialist capabilities of each classifier, a combined classifier may yield results which would not be possible with a single classifier. In this paper, we propose to combine the fingerprint and speaker verification decisions using a functional link network. This is to circumvent the nontrivial trial-and-error and iterative training effort as seen in backpropagation neural networks which cannot guarantee global optimal solutions. In many data fusion applications, as individual classifiers to be combined would have attained a certain level of classification accuracy, the proposed functional link network can be used to combine these classifiers by taking their outputs as the inputs to the network. The proposed network is first applied to a pattern recognition problem to illustrate its approximation capability. The network is then used to combine the fingerprint and speaker verification decisions with much improved receiver operating characteristics performance as compared to several decision fusion methods from the literature.

There are many learning algorithms available in the field of pattern classification and people are still discovering new algorithms that they hope will work better. Any new learning algorithm, beside its theoretical foundation, needs to be justified in many aspects including accuracy and efficiency when applied to real life problems. In this paper, we report the empirical comparison of a recent algorithm RM, its new extensions and three classical classifiers in different aspects including classification accuracy, computational time and storage requirement. The comparison is performed in a standardized way and we believe that this would give a good insight into the algorithm RM and its extension. The experiments also show that nominal attributes do have an impact on the performance of those compared learning algorithms.

The idea of user-specific multimodal biometrics is pioneered by Jain, A. K., et al. (2002) and further exploited by Toh, K-A, et al. (2004) recently. In this paper, we look into several issues pertaining to user-specific multimodal biometric verification. These issues include the small sample size problem, learning of local decision hyperplanes and setting of local thresholds. For small sample size problem, the noise-injection technique and a feature scaling-space technique are considered. For local learning, we adopt a recently proposed reduced polynomial since it has fast single-step computation and accurate estimation. For setting of local decision thresholds, nine baselines are identified. Extensive experiments are performed on a moderate data set and relatively conclusive results are observed.

This paper discusses methods in evaluating fingerprint image quality on a local level. Feature vectors covering directional strength, sinusoidal local ridge/valley pattern, ridge/valley uniformity and core occurrences are first extracted from fingerprint image subblocks. Each subblock is then assigned a quality level through pattern classification. Three different classifiers are employed to compare each of its different effectiveness. Positive results have been obtained based on our database.

A novel method using a reduced multivariate polynomial model has been developed for biometric decision fusion where simplicity and ease of use could be a concern. However, much to our surprise, the reduced model was found to have good classification accuracy for several commonly used data sets from the Web. In this paper, we extend the single output model to a multiple outputs model to handle multiple class problems. The method is particularly suitable for problems with small number of features and large number of examples. Basic component of this polynomial model boils down to construction of new pattern features which are sums of the original features and combination of these new and original features using power and product terms. A linear regularized least-squares predictor is then built using these constructed features. The number of constructed feature terms varies linearly with the order of the polynomial, instead of having a power law in the case of full multivariate polynomials. The method is simple as it amounts to only a few lines of Matlab code. We perform extensive experiments on this reduced model using 42 data sets. Our results compared remarkably well with best reported results of several commonly used algorithms from the literature. Both the classification accuracy and efficiency aspects are reported for this reduced model.

The multivariate polynomial model provides an effective way to describe complex nonlinear input-output relationships since it is tractable for optimization, sensitivity analysis, and prediction of confidence intervals. However, for high-dimensional and high-order problems, multivariate polynomial regression becomes impractical due to its huge number of product terms. This is especially true for the case of a full interaction model. In this paper, we propose a reduced multivariate polynomial model to circumvent the dimensionality problem with some compromise in its approximation capability. In multimodal biometrics and many classifiers fusion applications, as individual classifiers to be combined would have attained a certain level of classification accuracy, this reduced multivariate polynomial model can be used to combine these classifiers in the next level of classification taking their outputs as the inputs to the reduced multivariate polynomial model. The model is first applied to a well-known pattern classification problem to illustrate its classification capability. The reduced multivariate polynomial model is then applied to combine two biometric verification systems with improved receiver operating characteristics performance as compared to an optimal weighing method and a few commonly used classifiers.

In this paper, we treat the problem of combining fingerprint and speech biometric decisions as a classifier fusion problem. By exploiting the specialist capabilities of each classifier, a combined classifier may yield results which would not be possible in a single classifier. The Feedforward Neural Network provides a natural choice for such data fusion as it has been shown to be a universal approximator. However, the training process remains much to be a trial-and-error effort since no learning algorithm can guarantee convergence to optimal solution within finite iterations. In this work, we propose a network model to generate different combinations of the hyperbolic functions to achieve some approximation and classification properties. This is to circumvent the iterative training problem as seen in neural networks learning. In many decision data fusion applications, since individual classifiers or estimators to be combined would have attained a certain level of classification or approximation accuracy, this hyperbolic functions network can be used to combine these classifiers taking their decision outputs as the inputs to the network. The proposed hyperbolic functions network model is first applied to a function approximation problem to illustrate its approximation capability. This is followed by some case studies on pattern classification problems. The model is finally applied to combine the fingerprint and speaker verification decisions which show either better or comparable results with respect to several commonly used methods.

Automatically understanding events happening at a site is the ultimate goal of visual surveillance system. We investigate the challenges faced by automated surveillance systems operating in hostile conditions and demonstrate the developed algorithms via a system that detects water crises within highly dynamic aquatic environments. An efficient segmentation algorithm based on robust block-based background modelling and thresholding-with-hysteresis methodology enables swimmers to be reliably detected amid reflections, ripples, splashes and rapid lighting changes. Partial occlusions are resolved using a Markov Random Field framework that enhances the tracking capability of the system. Visual indicators of water crises are identified based on professional knowledge of water crises detection, based on which a set of swimmer descriptors has been defined. Through seamlessly fusing the extracted swimmer descriptors based on a novel functional link network, the system achieves promising results for water crises detection. The developed algorithms have been incorporated into a live system with robust performance for different hostile environments faced by an outdoor swimming pool.

This paper addresses the issue of training feedforward neural networks by global optimization. The main contributions include characterization of global optimality of a network error function, and formulation of a global descent algorithm to solve the network training problem. A network with a single hidden-layer and a single-output unit is considered. By means of a monotonic transformation, a sufficient condition for global optimality of a network error function is presented. Based on this, a penalty-based algorithm is derived directing the search towards possible regions containing the global minima. Numerical comparison with benchmark problems from the neural network literature shows superiority of the proposed algorithm over some local methods, in terms of the percentage of trials attaining the desired solutions. The algorithm is also shown to be effective for several pattern recognition problems.

The multivariate polynomials model provides an effective way to describe complex nonlinear input-output relationships as it is tractable for optimization, sensitivity analysis, and prediction of confidence intervals. However, for high dimensional and high order problems, multivariate polynomial regression becomes impractical due to its prohibitive number of product terms. This is especially true for the case of a full interaction model. In this paper, we propose a reduced multivariate polynomials model to circumvent the dimensionality problem with some compromise in the approximation capability. When applied to multi-modal biometrics fusion, this mode! is demonstrated to improve the combined classification performance in terms of classification accuracy.

In this paper, we address the false rejection problem due to the small solid state sensor area available for fingerprint image capture. We propose a minutiae data synthesis approach to circumvent this problem. The main advantages of this approach over the existing image mosaicing approach include low memory storage requirements and low computational complexity. Moreover, the possible matching search overhead due to data redundancy can be reduced. Extensive experiments are conducted to determine the best transformation suitable for minutiae alignment. Among the three transformations presented, affine transformation is found to be most suited for minutiae alignment. We demonstrate the idea of synthesis with an example using physical fingerprint images. The proposed synthesis system is also shown to reduce the number of false rejects caused by the use of different fingerprint regions for matching.