Internet of Things (IoT) networks face increasing security threats due to their distributed nature and resource constraints. Although federated learning (FL) has gained prominence as a privacy-preserving framework for distributed IoT environments, current federated principal component analysis (PCA) methods fail to incorporate personalized noise inherent in local data. To address these limitations, we propose an efficient personalized federated PCA (FedEP) method for anomaly detection in IoT networks. The proposed model achieves personalization through introducing local representations with the ℓ1-norm for element-wise sparsity, while maintaining robustness via enforcing local models with the ℓ2,1-norm for row-wise sparsity. To solve this non-convex problem, we develop a manifold optimization algorithm based on the alternating direction method of multipliers (ADMM) with theoretical convergence guarantees. Experimental results confirm that our proposed FedEP outperforms the state-of-the-art FL methods, achieving excellent accuracy in various IoT security scenarios. Our implementation code is accessible at https://github.com/xianchaoxiu/FedEP.

Tensor canonical correlation analysis (TCCA) has garnered significant attention due to its effectiveness in capturing high-order correlations in multi-view learning. However, existing TCCA methods often underemphasize the characterization of individual structures and lack algorithmic convergence guarantees. In order to deal with these challenges, we propose a novel sparse TCCA model called STCCA-L, which integrates sparse regularization of canonical matrices and Laplacian regularization of multi-order graphs into the TCCA framework, thereby effectively exploiting the geometric structure of individual views. To solve this non-convex model, we develop an efficient alternating manifold proximal gradient algorithm based on manifold optimization, which avoids computationally expensive full tensor decomposition and leverages a semi-smooth Newton method for resolving the subproblem. Furthermore, we rigorously prove the convergence of the algorithm and analyze its complexity. Experimental results on eight benchmark datasets demonstrate the superior classification performance of the proposed method. Notably, on the 3Sources dataset, it achieves improvements of at least 4.50% in accuracy and 6.77% in F1 score over competitors. Our code is available at https://github.com/zhudafa/STCCA-L.

Remarkable achievements have been made in utilizing information granulation for improving the accuracy and interpretability of time series forecasting. However, existing information granule-based methods suffer from multiple limitations, including the loss of scale information and the neglect of nonlinear trends and magnitude information, which compromise prediction quality. To address these issues, this article proposes an information granule-based time series prediction method by synergizing multiscale and multitype information granulations. Firstly, an adaptive multiscale sequence generation method is designed to generate the multiscale subsequences, which provide the multiscale views of time series. Then, the dual-mode granulation mechanism integrating magnitude-type and trend-type information granulations is proposed to comprehensively exploit linear trends, non-linear trends, and magnitude information of time series. Thirdly, the cross-scale information granule fusion is designed to model the inherent interrelationships between multiscale information granules, which can promote the collaboration across different-scale information granules. Finally, multitype information granules are synergized to perform prediction, and the prediction process also benefits from the fused multiscale information. Experiments on small, medium, and large size time series datasets demonstrate that the proposed method achieves significant performance compared with the state-of-the-art methods.

Sporadic ambient radio frequency signals can offer opportunistic spectrum and energy sources for backscatter communications, but they also induce unpredictable transmission interruptions in ambient backscatter wireless networks (AmBWNs). Integrating self-carrier-generative active transmissions with backscatter communications could significantly enhance transmission stability but require frequent mode switching to accommodate the ever-changing ambient radio frequency signals. However, this will result in frequent changes in network topology and link capacity, posing significant challenges in solving the network maximum flow problem in hybrid AmBWNs. To address this problem, we design a link-aware online maximum flow (OnMAXFlow) scheme to tackle agile and adaptive flow scheduling and communication mode selection. Specifically, we first employ an online learning framework to dynamically track changes in ambient signal strength and channel states, enabling real-time evaluation of link capacity. We then model the network maximum flow problem as a stochastic multi-armed bandit (MAB) problem and solve it with a Kullback-Leibler upper confidence bound (KL-UCB) algorithm. Our experimental evaluation results reveal that our OnMAXFlow scheme exhibits rapid convergence and superior adaptability against the varying network states, while maintaining spectrum efficiency and latency performance comparable to the Oracle scheme, which always selects the optimal transmission modes and paths.

In order to represent the influences of different semantics on targets and improve the prediction with interpretability ability for multi-dimensional time series, we integrate Axiomatic Fuzzy Set (AFS) and Fuzzy Cognitive Map (FCM) with memory for fuzzy knowledge representation and prediction in this paper. The AFS is used to extract semantics of concepts for fuzzy representation using data distribution. The FCM with memory is trained to model the influence relationships between different semantics of concepts and multiple targets based on multi-dimensional time series data. And a multi- dimensional learning algorithm of AFS-FCM with memory based on gradient descent is developed to investigate the influences of different semantics of concepts on multiple targets. Finally, we validate our model by comparing with other FCMs, intrinsic interpretable models and machine learning methods for prediction of air quality multidimensional time series data, and discuss the performance of AFS-FCM with different transformation functions. The model can not only predict air quality accurately, but also explicitly reveal the specific quantitative relationship of different semantics of meteorology on air quality.

Lesion segmentation on nasal endoscopic images is challenging due to its complex lesion features. Fully-supervised learning methods achieve promising performance with pixel-level annotations but impose a significant annotation burden on experts. Although weakly supervised or semi-supervised methods can reduce the labelling burden, their performance is still limited. Some weakly semi-supervised methods employ a novel annotation strategy that labels weak single-point annotations for the entire training set while providing pixel-level annotations for a small subset of the data. However, the relevant weakly semi-supervised methods only mine the limited information of the point itself, while ignoring its label property and surrounding reliable information. This paper proposes a simple yet efficient weakly semi-supervised method called the Point-Neighborhood Learning (PNL) framework. PNL incorporates the surrounding area of the point, referred to as the point-neighborhood, into the learning process. In PNL, we propose a point-neighborhood supervision loss and a pseudo-label scoring mechanism to explicitly guide the model’s training. Meanwhile, we proposed a more reliable data augmentation scheme. The proposed method obviously improves performance without increasing the parameters of the segmentation neural network. Experimental results indicate that our method consistently achieves better performance compared to SOTA methods. Additional validation on colonoscopic polyp segmentation datasets confirms our method’s generalizability.

Impressive advancements have been achieved in utilizing information granulation for solving long-term time series prediction problems. However, most state-of-the-art methods suffer from limitations due to not only using the single-scale information granulation but also the lack of trend information. As a result, the prediction models are difficult to capture the multiscale temporal dependencies and dynamic behavior of time series. To address these problems, this article proposes a multiscale information granule-based time series forecasting model. First, the trend-based information granulation strategy is proposed to generate trend information granules that can capture dynamic behavior and trend information in an incremental manner. Then, the multiscale fusion mechanism is proposed to form multiscale information granules with diversified information, which fuses local and global information at different scales. Finally, the two-stage prediction mechanism is proposed to capture multiscale temporal dependencies and perform long-term prediction. A series of experiments were conducted on publicly available time series. Comparative analysis shows that the proposed method outperforms existing numeric models and granular models in long-term prediction on regular and large data time series.

To deal with high-dimensional unlabeled datasets in many areas, principal component analysis (PCA) has become a rising technique for unsupervised feature selection (UFS). However, most existing PCA-based methods only consider the structure of datasets by embedding a single sparse regularization or constraint on the transformation matrix. In this paper, we introduce a novel bi-sparse method called BSUFS to improve the performance of UFS. The core idea of BSUFS is to incorporate $\ell _{2,p}$ -norm and $\ell _{q}$ -norm into the classical PCA, which enables our method to select relevant features and filter out irrelevant noises, thereby obtaining discriminative features. Here, the parameters $p$ and $q$ are within the range of [ $0, 1$ ). Therefore, BSUFS not only constructs a unified framework for bi-sparse optimization, but also includes some existing works as special cases. To solve the resulting non-convex model, we propose an efficient proximal alternating minimization (PAM) algorithm using Stiefel manifold optimization and sparse optimization techniques. In addition, the computational complexity analysis is presented. Extensive numerical experiments on synthetic and real-world datasets demonstrate the effectiveness of our proposed BSUFS. The results reveal the advantages of bi-sparse optimization in feature selection and show its potential for other fields in image processing. Our code is available at https://github.com/xianchaoxiu/BSUFS.

It is necessary to establish a spatio-temporal correlation model in the traffic data to predict the state of the transportation system. Existing research has focused on traditional graph neural networks, which use predefined graphs and have shared parameters. But intuitive predefined graphs introduce biases into prediction tasks and the fine-grained spatio-temporal information can not be obtained by the parameter sharing model. In this paper, we consider it is crucial to learn node-specific parameters and adaptive graphs with complete edge information. To show this, we design a model based on graph structure that decouples nodes and edges into two modules. Each module extracts temporal and spatial features simultaneously. The adaptive node optimization module is used to learn the specific parameter patterns of all nodes, and the adaptive edge optimization module aims to mine the interdependencies among different nodes. Then we propose a Decoupled Adaptive Graph Convolution Attention Network for Traffic Forecasting (DAGCAN), which relies on the above two modules to dynamically capture the fine-grained spatio-temporal relationships in traffic data. Experimental results on four public transportation datasets, demonstrate that our model can further improve the accuracy of traffic prediction.

In light of the shortage of acupuncturists and issues such as subjectivity leading to inconsistent efficacy and variability in accuracy and stability, an eight-degree-of-freedom (DOF) dexterous acupuncture robot system is developed in this article. An effective model predictive control (MPC) scheme, integrating needle bending compensation and manipulative techniques, is proposed to ensure the safety, accuracy, and stability of acupuncture manipulative trajectory tracking control. Based on acupuncture’s working scenarios, a dexterous 2-DOF acupuncture mechanism is designed. Then, the kinematic and acupoint models are established. To enhance the controller’s robustness, trajectory tracking constraints are included as soft constraints in the objective function before needle insertion, leading to the development of a trajectory-constrained MPC (TCMPC) model. This model not only improves the tracking accuracy and convergence speed but also ensures the stability of the solution. Additionally, a needle bending correction trajectory control method, considering needle deformation, is derived to perceive and correct the needle bending deformation under complex force environments. Furthermore, an adaptive impedance control method, integrating feedforward-feedback control (FFC), is proposed to simultaneously control force and track manipulative techniques under various contact environments. The uniform ultimate boundedness of the closed-loop system is verified using the Lyapunov theory. Finally, the effectiveness of the proposed methods and prototypes is validated through numerical simulations and practical experiments.

Bilevel optimization has attracted substantial attentions in recent years due to its wide applications in machine learning. However, most existing algorithms either primarily developed under centralized setting or suffer expensive inner-loop updates for hypergradient estimation. What's worse, the projection operator in constrained scenario may demand prohibitively computational cost, which further necessitates efficient projection-free bilevel optimization algorithms over networks. To fill this gap, we propose a novel single-loop distributed Frank-Wolfe algorithm DBO-FW for constrained bilevel optimization problems by simultaneously leveraging a nested approximation technique and a gradient tracking mechanism to locally estimate the global hypergradient. Moreover, we provide the convergence guarantee for the proposed DBO-FW. Numerical results also validate the efficiency of our algorithm.

In this paper, we propose to create animatable avatars for interacting hands with 3D Gaussian Splatting (GS) and single-image inputs. Existing GS-based methods designed for single subjects often yield unsatisfactory results due to limited input views, various hand poses, and occlusions. To address these challenges, we introduce a novel two-stage interaction-aware GS framework that exploits cross-subject hand priors and refines 3D Gaussians in interacting areas. Particularly, to handle hand variations, we disentangle the 3D presentation of hands into optimization-based identity maps and learning-based latent geometric features and neural texture maps. Learning-based features are captured by trained networks to provide reliable priors for poses, shapes, and textures, while optimization-based identity maps enable efficient one-shot fitting of out-of-distribution hands. Furthermore, we devise an interaction-aware attention module and a self-adaptive Gaussian refinement module. These modules enhance image rendering quality in areas with intra- and inter-hand interactions, overcoming the limitations of existing GS-based methods. Our proposed method is validated via extensive experiments on the large-scale InterHand2.6M dataset, and it significantly improves the state-of-the-art performance in image quality. Code and models will be released upon acceptance.

Residual blocks have been widely used in deep learning networks. However, information may be lost in residual blocks due to the relinquishment of information in rectifier linear units (ReLUs). To address this issue, invertible residual networks have been proposed recently but are generally under strict restrictions which limit their applications. In this brief, we investigate the conditions under which a residual block is invertible. A sufficient and necessary condition is presented for the invertibility of residual blocks with one layer of ReLU inside the block. In particular, for widely used residual blocks with convolutions, we show that such residual blocks are invertible under weak conditions if the convolution is implemented with certain zero-padding methods. Inverse algorithms are also proposed, and experiments are conducted to show the effectiveness of the proposed inverse algorithms and prove the correctness of the theoretical results.

The canonical correlation analysis (CCA) has attracted wide attention in fault detection (FD). To improve the detection performance, we propose a new joint sparse constrained CCA (JSCCCA) model that integrates the$\ell _{2,0}$-norm joint sparse constraints into classical CCA. The key idea is that JSCCCA can fully exploit the joint sparse structure to determine the number of extracted variables. We then develop an efficient alternating minimization algorithm using the improved iterative hard thresholding and manifold constrained gradient descent method. More importantly, we establish the convergence guarantee with detailed analysis. Finally, we provide extensive numerical studies on the simulated dataset, the benchmark Tennessee Eastman process, and a practical cylinder-piston process. In some cases, the computing time is reduced by 600 times, and the FD rate is increased by 12.62% compared with classical CCA. The results suggest that the proposed approach is efficient and fast.

Federated learning (FL) is a promising distributed machine learning scheme where multiple clients collaborate by sharing a common learning model while maintaining their private data locally. It can be applied to a lot of applications, e.g., training an automatic driving system by the perception of multiple vehicles. However, some clients may join the training system dynamically, which affects the stability and accuracy of the learning system a lot. Meanwhile, data heterogeneity in the FL system exacerbates the above problem further due to imbalanced data distribution. To solve the above problems, we propose a novel FL framework named FedREM (Retain-Expansion and Matching), which guides clients training models by two mechanisms. They are 1) a Retain-Expansion mechanism that can let clients perform local training and extract data characteristics automatically during the training; 2) a Matching mechanism that can ensure new clients quickly adapt to the global model based on matching their data characteristics and adjusting the model accordingly. Results of extensive experiments verify that our FedREM outperforms various baselines in terms of model accuracy, communication efficiency, and system robustness.

Canonical correlation analysis (CCA) has attracted great interest in multi-view representation. However, most of the CCA methods heavily rely on the matrix structure, which may neglect the prior geometric information in high-order data. To deal with the above issue, we first propose a novel tensor CCA formulation with orthogonality, called TCCA-O, based on the Tucker decomposition to preserve the orthogonality. Then, we incorporate a structured sparse regularization term into the TCCA-O, called TCCA-OS, to improve feature representation. In addition, we develop an efficient alternating direction method of multipliers (ADMM)-based algorithm to solve TCCA-OS and conduct numerical comparisons on four public datasets. The results validate the advantages of the proposed methods in terms of classification accuracy, parameter sensitivity, noise robustness, and model stability. In particular, TCCA-O and TCCA-OS improve the classification accuracy by at least 10.03% and 10.36%, respectively, over the state-of-the-art CCA methods on the Caltech101-7 dataset.

Positioning and posture of the robotic joints and end effector could probably introduce random initial errors. Those errors could exponentially deteriorate with compounded of common noise perturbation to cause the final failure of repetitive tracking control. To better improve the tolerance of those complex errors, a novel reciprocal of the exponential varying-parameter recurrent neural network (RE-VP-RNN) is proposed in this article to consider superimposed noise interference including the initial position deviation and noise perturbation together. Theoretical analysis further proves the convergence of the proposed method. The effectiveness, accuracy, and robustness of the proposed RE-VP-RNN solver are verified by simulation and physical experiments on three representative redundant and hype-redundant manipulators. The proposed model could be widely used in robot control for high-precision machining scenarios such as medical, industry, and aviation.

In this article, we propose an efficient fault diagnosis framework to achieve accurate fault isolation. The core is to introduce the$\ell _{2,0}$-norm sparsity constrained optimization to reduce the variable redundancy and determine the variable number, which is different from the existing sparse variants. In order to illustrate the idea, this article takes principal component analysis (PCA) as an essential step. First, a sparsity-aware PCA is constructed by taking advantage of the$\ell _{2,0}$-norm constrained optimization. Afterward, a two-stage monitoring strategy is designed, including fault detection and fault isolation. Once the fault is detected, the sparsity level is then shrunk to achieve accurate fault isolation. Moreover, an alternating direction method of multipliers-based optimization algorithm is developed with detailed implementation. Finally, the detection improvement and accurate isolation performance are validated by two simulated examples, the Tennessee Eastman benchmark process, and a practical cylinder-piston process.

This paper presents a novel approach to multi-view graph learning that combines weight learning and graph learning in an alternating optimization framework. Multi-view graph learning refers to the problem of constructing a unified affinity graph using heterogeneous sources of data representation, which is a popular technique in many learning systems where no prior knowledge of data distribution is available. Our approach is based on a fusion-and-diffusion strategy, in which multiple affinity graphs are fused together via a weight learning scheme based on the unsupervised graph smoothness and utilised as a consensus prior to the diffusion. We propose a novel multi-view diffusion process that learns a manifold-aware affinity graph by propagating affinities on tensor product graphs, leveraging high-order contextual information to enhance pairwise affinities. In contrast to existing multi-view graph learning approaches, our approach is not limited by the quality of initial graphs or the assumption of a latent common subspace among multiple views. Instead, our approach is able to identify the consistency among views and fuse multiple graphs adaptively. We formulate both weight learning and diffusion-based affinity learning in a unified framework and propose an alternating optimization solver that is guaranteed to converge. The proposed approach is applied to image retrieval and clustering tasks on 16 real-world datasets. Extensive experimental results demonstrate that our approach outperforms state-of-the-art methods for both retrieval and clustering on 13 out of 16 datasets.

With the accelerated popularization of 5G applications, accurate cellular traffic prediction is becoming increasingly important for efficient network management. Currently, the latest algorithms for cellular traffic prediction generally neglect extraction of the shallow features of cellular traffic and the prediction accuracy is hence limited. Therefore, we propose a global-local spatial-temporal transformer network (GLSTTN) that can fully excavate diverse spatial-temporal characteristics of cellular traffic for accurate cellular traffic prediction. Specifically, GLSTTN achieves this goal by constructing two modules: the global spatial-temporal module and the local spatial-temporal module. In the global spatial-temporal module, GLSTTN captures global correlations using stacked spatial-temporal blocks, where each block is composed of one spatial transformer and one temporal transformer. A skip connection is then used in each block to strengthen feature propagation. In the local spatial-temporal module, GLSTTN fully extracts the local spatial-temporal dependencies hidden in globally encoded features using densely connected convolutional neural networks. Extensive experiments demonstrate that GLSTTN achieves more accurate cellular traffic prediction than existing approaches on a real-world cellular traffic dataset.

Generalized canonical correlation analysis (GCCA) that aims to deal with multi-view data has attracted extensive attention in signal processing. To improve the representation performance, this letter proposes a new sparsity constrained GCCA (SCGCCA). Technically, it integrates the$\ell _{2,0}$-norm constrained optimization into GCCA, which has not been investigated in the literature. Compared with the existing$\ell _{2,1}$-norm regularized GCCA, the proposed SCGCCA can not only exploit the similarity information belonging to the same features but also determine the number of extracted features. Although it is a nonconvex minimization problem, an efficient alternating minimization algorithm can be designed. Furthermore, a Newton hard thresholding pursuit technique is developed to accelerate the convergence tremendously. Empirical studies suggest both the effectiveness and efficiency of the proposed SCGCCA comparing with the existing GCCA and its variants. In particular, the speed can be increased by 150 times for the simulated dataset.

Time-series prediction based on information granule in which the algorithm is developed by deriving the relations existing in the granular time series, has achieved excellent success. However, the existing uncertainty in data and the computational demand of the granulation process make it difficult for these methods to accurately and efficiently achieve long-term prediction. In this article, a fuzzy-probabilistic prediction approach with evolving clustering-based granulation is proposed. First, the evolving clustering-based granulation strategy is proposed to transform the original numerical data into information granules. The granulation process is performed in an incremental way and the information granules are represented with the triplets, which can efficiently reduce the computation overhead. Then, the proposed information granule clustering is used to derive the group relations in the information granules. Based on the logical relationships of information granules in the temporal order, the information granule forecasting the integrated fuzzy and probability theory is proposed to deal with uncertainties and perform final long-term prediction. A series of experiments using publicly available time series are conducted, and the comparative analysis demonstrates that the proposed approach can achieve a better performance for regular and Big Data time series than the existing granular and numeric models for long-term prediction.

Most traditional methods for extracting key points from the 3D point cloud are based on the geometric features of points and they pose problems such as low accuracy. In order to solve these problems, this paper proposes a novel approach based on 2D image mapping, making it able to achieve highly accurate localization of key points. Specifically, it works as follows: input images are first selected for Harris corner detection; the three pairs of marker points of the images and the point cloud are then selected to calculate the transformation matrix T between them; next, the image key points are mapped onto the three-dimensional points through the transformation matrix T, for which the extraction of key points is achieved. Experimental results show that the proposed algorithm is able to greatly improve the extraction accuracy of key points in comparison with traditional algorithms.

This article proposes an efficient nonlinear process monitoring method (DCCA-SCO) by integrating canonical correlation analysis (CCA), deep autoencoder neural networks (DAENNs), and sparsity-constrained optimization (SCO). Specifically, DAENNs are first used to learn a nonlinear function automatically, which characterizes intrinsic features of the original process data. Then, the CCA is performed in that low-dimensional representation space to extract the most correlated variables. In addition, the SCO is imposed to reduce the redundancy of the hidden representation. Unlike other deep CCA methods, the DCCA-SCO provides a new nonlinear method that is able to learn a nonlinear mapping with a sparse prior. The validity of the proposed DCCA-SCO is extensively demonstrated on the benchmark Tennessee Eastman (TE) process and the diesel generator process. In particular, compared with the classical CCA, the fault detection rate is increased by 8.00% for the fault IDV(11) in the TE process.

Modern convolutional neural network (CNN)-based object detectors focus on feature configuration during training but often ignore feature optimization during inference. In this article, we propose a new feature optimization approach to enhance features and suppress background noise in both the training and inference stages. We introduce a generic inference-aware feature filtering (IFF) module that can be easily combined with existing detectors, resulting in our iffDetector. Unlike conventional open-loop feature calculation approaches without feedback, the proposed IFF module performs the closed-loop feature optimization by leveraging high-level semantics to enhance the convolutional features. By applying the Fourier transform to analyze our detector, we prove that the IFF module acts as a negative feedback that can theoretically guarantee the stability of the feature learning. IFF can be fused with CNN-based object detectors in a plug-and-play manner with little computational cost overhead. Experiments on the PASCAL VOC and MS COCO datasets demonstrate that our iffDetector consistently outperforms state-of-the-art methods with significant margins.

This article proposes a novel modeling method for the stochastic nonlinear degradation process by using the relevance vector machine (RVM), which can describe the nonlinearity of degradation process more flexibly and accurately. Compared with the existing methods, where degradation processes are modeled as the Wiener process with a nonlinear drift function formulized as the power law or exponential law, this kind of modeling method can characterize degradation processes with more nonlinear behavior. Instead of modeling the drift coefficient of the Wiener process directly, the weighted combination of basis functions is utilized to express the increment of the Wiener process and the parameters are calculated by a sparse Bayesian learning algorithm. Based on the proposed model, a numerical approximation formula for the probability density function (PDF) of the remaining useful life (RUL) is derived. Finally, comparison studies, including a numerical simulation and a practical case, are provided to demonstrate the effectiveness and the accuracy of the proposed methods for RUL estimation.

Long-term time series prediction is a challenging and essential task both in theory and practice. Recently, information granulation is shown to be an appropriate tool for the long-term forecast. Though some models for the long-term prediction problem have been proposed using information granulation recently, there is still a growing need to develop new prediction approaches for time series data based on information granule, which can capture the dynamic trend change with high accuracy. In this article, a long-term prediction approach, based on back-propagation neural network and information granule, is proposed. First, the individual numerical intervals for the time series are obtained by using the principle of justifiable granularity in information granule. Then, an automatic linear trend extraction method is developed to extract the trend change, which is inherited in granules. Finally, a hierarchy of neural network is constructed to carry out prediction by using information granule as input. Experiments using publicly available time series datasets demonstrate that the proposed approach can achieve better performance than the existing models for long-term prediction.

Graph-based semisupervised learning is of great importance in many effective learning systems, particularly in agnostic settings where no parametric information or other prior knowledge about the data distribution is available. It leverages the graph structure to propagate labels from labeled nodes to unlabeled ones. Two separate stages are usually involved: constructing an affinity graph and propagating labels on the graph for transductive inference. It is suboptimal to manage them independently, as the correlation between the affinity graph and the labels would not be fully exploited. In this article, we integrate these two stages into one unified framework by formulating the graph construction as a regularized function estimation problem, similar to label propagation. We then propose an alternating diffusion process to solve them alternately, which allows us to learn the graph and unknown labels in an iterative fashion. With the proposed framework, we can construct a dynamic graph adapted to the given and predicted labels iteratively, resulting in more accurate and robust label propagation performance. Extensive experiments on synthetic data and various real-world data have demonstrated the superiority of the proposed method compared with other state-of-the-art methods.

In this article, a novel interpretable classifier is proposed on the basis of axiomatic fuzzy set (AFS) theory and semantic entropy. AFS theory provides a unified and coherent way to deal with uncertainty of randomness and imprecision of fuzziness in data mining and knowledge discovery, which support many investigations in classification area. However, the existing AFS-based classifiers are weak in obtaining the optimal semantic description. To address this drawback, a new measure, named semantic entropy extended in Shannon's entropy, is developed to evaluate the discriminatory capabilities of semantic descriptions for each category. Moreover, the semantic entropy is utilized to design a classifier in the framework of AFS theory, called axiomatic fuzzy set and semantic entropy (AFSSE), which is capable of achieving sound classification performance and interpretability. Meanwhile, it provides a new framework of classifier design that can adapt more human-oriented recognition mechanisms. Furthermore, an evaluation index is used to prune descriptions to deliver a promising performance. Compared to the previous AFS-based classifiers, the proposed approach offers a semantic entropy to measure the information that is derived from semantic descriptions of data, so that the optimal semantic descriptions of each class can be obtained. For the purpose of illustrating the effectiveness of the classifier, several datasets are utilized to facilitate a comparative analysis of the proposed approach and other state-of-the-art classifiers. The experimental studies demonstrate that the proposed approach can achieve the semantic descriptions of each class and the performance of AFSSE is comparable with the performance of other approaches.

Different facial components carry different amount of information being conveyed for 3D dynamic expression recognition. Hence, identifying facial components that are highly relevant to specific expression changes is crucial for discriminative facial expression recognition. This work aims to learn expression-sensitive features, which are expected to not only yield comparable recognition performance with the state-of-the-art ones, but also can be interpreted by human. Firstly, spatio-temporal features (HOG3D) are extracted from local depth patch-sequences to represent facial expression dynamics. A two-phase feature selection process is then proposed to determine the facial components that can best distinguish the expressions. In order to verify the effectiveness of the resulting facial components, the expression-sensitive features from the corresponding area are fed into a hierarchical classifier for facial expression recognition. The proposed method is evaluated on the BU-4DFE benchmark database, and results show that learned expression-sensitive features can achieve a comparable recognition performance with existing methods. Additionally, the resulting HOG3D features after feature selection can be used to generate semantic interpretation of the expression dynamics.

The key challenge of face recognition is to develop effective feature representations for reducing intra-personal variations while enlarging inter-personal differences. This paper presents a novel non-linear discriminant error criterion which can be used in effective feature learning from raw pixels. Unlike many existing methods which assume the problem to be linear in nature, the proposed method utilizes a novel deep learning (DL) framework which makes no prior assumptions thus exploiting the full potential of learning a highly non-linear transformation. High level representations learnt via the proposed model are highly supervised and can help to boost the performance of subsequent classifiers such as LDA. This study clearly shows the value of using non-linear discriminant error criterion as a tractable objective to guide the learning of useful high level features in various face related problems. The extracted features are learnt from local face regions and the results of the experiments performed on 3 different face image databases demonstrate the superiority and the generalizability of our method compared to existing work, as well as the applicability of the concept onto many different deep learning models of the same nature.

This paper addresses the problem of person-independent 4D facial expression recognition. Unlike the majority of existing works, we propose to extract spatio-temporal features in 4D data (3D expression sequences changing over time) to represent 3D facial expression dynamics sufficiently, rather than extracting features frame-by-frame. First, the proposed method extracts local depth patch-sequences from consecutive expression frames based on the automatically detected facial landmarks. Three dimension discrete cosine transform (3D-DCT) is then applied on these patch-sequences to extract spatio-temporal features for facial expression dynamic representation. Finally, the extracted compact features (3D-DCT coefficients) are fed to nearest-neighbor classifier to finish expression recognition after feature selection and dimension reduction, in which the redundant features are filtered out. Experiments on the benchmark BU-4DFE database show that the proposed method achieves the best average recognition rate 78.8% among the existing automatic approaches, and outperforms the existing techniques in the recognition of those easily confused expressions (anger and sadness) significantly.

Many vision problems deal with high-dimensional data, such as motion segmentation and face clustering. However, these high-dimensional data usually lie in a low-dimensional structure. Sparse representation is a powerful principle for solving a number of clustering problems with high-dimensional data. This principle is motivated from an ideal modeling of data points according to linear algebra theory. However, real data in computer vision are unlikely to follow the ideal model perfectly. In this paper, we exploit the mixed norm regularization for sparse subspace clustering. This regularization term is a convex combination of the ℓ1norm, which promotes sparsity at the individual level and the block norm ℓ2/1which promotes group sparsity. Combining these powerful regularization terms will provide a more accurate modeling, subsequently leading to a better solution for the affinity matrix used in sparse subspace clustering. This could help us achieve better performance on motion segmentation and face clustering problems. This formulation also caters for different types of data corruptions. We derive a provably convergent algorithm based on the alternating direction method of multipliers (ADMM) framework, which is computationally efficient, to solve the formulation. We demonstrate that this formulation outperforms other state-of-arts on both motion segmentation and face clustering.

In this paper, we aim to improve one of the current state-of-the-art models for facial components detection/localization. The objectives are to increase the amount of landmark points detected and improve the landmark extraction accuracy for frontal faces. The model is following Zhu and Ramanan's approach with a tree-structure. The popular AR dataset is chosen as an alternative training dataset as it provides more landmark points requested. Our extension models are compared with Zhu and Ramanan's frontal face models in terms of detection accuracy. We also compare our models with another robust facial components detector called CompASM. Our experiments show that our models can achieve lower error rate on some fiducial points by providing more landmarks, and these accurate fiducial points will provide more accurate features for some applications related to facial shapes. The impact of image colour spaces other than RGB on the proposed detector is also investigated.

We propose a new unsupervised algorithm for the automatic alignment of two manifolds of different datasets with possibly different dimensionalities. Alignment is performed automatically without any assumptions on the correspondences between the two manifolds. The proposed algorithm automatically establishes an initial set of sparse correspondences between the two datasets by matching their underlying manifold structures. Local feature histograms are extracted at each point of the manifolds and matched using a robust algorithm to find the initial correspondences. Based on these sparse correspondences, an embedding space is estimated where the distance between the two manifolds is minimized while maximally retaining the original structure of the manifolds. The problem is formulated as a generalized eigenvalue problem and solved efficiently. Dense correspondences are then established between the two manifolds and the process is iteratively implemented until the two manifolds are correctly aligned consequently revealing their joint structure. We demonstrate the effectiveness of our algorithm on aligning protein structures, facial images of different subjects under pose variations and RGB and Depth data from Kinect. Comparison with an state-of-the-art algorithm shows the superiority of the proposed manifold alignment algorithm in terms of accuracy and computational time.

Facial expressions form a significant part of our nonverbal communications and understanding them is essential for effective human computer interaction. Due to the diversity of facial geometry and expressions, automatic expression recognition is a challenging task. This paper deals with the problem of person-independent facial expression recognition from a single 3D scan. We consider only the 3D shape because facial expressions are mostly encoded in facial geometry deformations rather than textures. Unlike the majority of existing works, our method is fully automatic including the detection of landmarks. We detect the four eye corners and nose tip in real time on the depth image and its gradients using Haar-like features and AdaBoost classifier. From these five points, another 25 heuristic points are defined to extract local depth features for representing facial expressions. The depth features are projected to a lower dimensional linear subspace where feature selection is performed by maximizing their relevance and minimizing their redundancy. The selected features are then used to train a multi-class SVM for the final classification. Experiments on the benchmark BU-3DFE database show that the proposed method outperforms existing automatic techniques, and is comparable even to the approaches using manual landmarks.

In this paper, we consider the problem of robust face recognition using color information. In this context, sparse representation-based algorithms are the state-of-the-art solutions for gray facial images. We will integrate the existing sparse representation-based algorithms with color information and this integration can improve the previous performances significantly. Furthermore, we propose a new performance metric, namely the discriminativeness (DIS) to describe the recognition effectiveness for sparse representation algorithms. We find out that the richer information in color space can be used to increase the DIS, i.e. enhancing the robustness in face recognition. Extensive experiments have been conducted under different conditions, including various feature extractors, random pixel corruptions and occlusions on AR and GT databases, to demonstrate the advantages of using color information in robust face recognition. Detailed analysis is also included for each experiment to explain why and how color improve the robustness of different sparse representation-based methods.

We present an algorithm that uses a low resolution 3D sensor for robust face recognition under challenging conditions. A preprocessing algorithm is proposed which exploits the facial symmetry at the 3D point cloud level to obtain a canonical frontal view, shape and texture, of the faces irrespective of their initial pose. This algorithm also fills holes and smooths the noisy depth data produced by the low resolution sensor. The canonical depth map and texture of a query face are then sparse approximated from separate dictionaries learned from training data. The texture is transformed from the RGB to Discriminant Color Space before sparse coding and the reconstruction errors from the two sparse coding steps are added for individual identities in the dictionary. The query face is assigned the identity with the smallest reconstruction error. Experiments are performed using a publicly available database containing over 5000 facial images (RGB-D) with varying poses, expressions, illumination and disguise, acquired using the Kinect sensor. Recognition rates are 96.7% for the RGB-D data and 88.7% for the noisy depth data alone. Our results justify the feasibility of low resolution 3D sensors for robust face recognition.

Subwindow search aims to find the optimal subimage which maximizes the score function of an object to be detected. After the development of the branch and bound (B&B) method called Efficient Subwindow Search (ESS), several algorithms (IESS, AESS, ARCS) have been proposed to improve the performance of ESS. For n×n images, IESS's time complexity is bounded by O(n3) which is better than ESS, but only applicable to linear score functions. Other work shows that Monge properties can hold in subwindow search and can be used to speed up the search to O(n3), but only applies to certain types of score functions. In this paper we explore the connection between submodular functions and the Monge property, and prove that sub-modular score functions can be used to achieve O(n3) time complexity for object detection. The time complexity can be further improved to be sub-cubic by applying B&B methods on row interval only, when the score function has a multivariate submodular bound function. Conditions for sub-modularity of common non-linear score functions and multivariate submodularity of their bound functions are also provided, and experiments are provided to compare the proposed approach against ESS and ARCS for object detection with some nonlinear score functions.

Finding a good color space is one of the main research goals for color face recognition. Existing research shows that RGB can improve over gray-scale, while some other color spaces (YQCr for instance) can improve over RGB. However, all developed color models consist of only three color components transformed linearly from RGB. Since three colors may not capture sufficient information for solving complex face recognition problems and non-linear transformed colors usually encode very different information, this paper investigates the feasibility and effectiveness of using more than three colors including some non-linear color spaces. We propose a novel color combination algorithm namely the Multiple Color Fusion (MCF) model to utilize multiple colors. Experiment 4 on FRGC2 is conducted to demonstrate the effectiveness of MCF. In particular, MCF outperforms any existing three-color based methods by at least 3% and improves over RGB by 8%.

Optimum subwindow search for object detection aims to find a subwindow so that the contained subimage is most similar to the query object. This problem can be formulated as a four dimensional (4D) maximum entry search problem wherein each entry corresponds to the quality score of the subimage contained in a subwindow. For n × n images, a naive exhaustive search requires O(n4) sequential computations of the quality scores for all subwindows. To reduce the time complexity, we prove that, for some typical similarity functions like Euclidian metric, χ2metric on image histograms, the associated 4D array carries some Monge structures and we utilise these properties to speed up the optimum subwindow search and the time complexity is reduced to O(n3). Furthermore, we propose a locally optimal alternating column and row search method with typical quadratic time complexity O(n2). Experiments on PASCAL VOC 2006 demonstrate that the alternating method is significantly faster than the well known efficient subwindow search (ESS) method whilst the performance loss due to local maxima problem is negligible.

In this paper we propose to convert the task of face hallucination into an image decomposition problem, and then use the morphological component analysis (MCA) for hallucinating a single face image, based on a novel three-step framework. Firstly, a low-resolution input image is up-sampled by interpolation. Then, the MCA is employed to decompose the interpolated image into a high-resolution image and an unsharp masking, as MCA can properly decompose a signal into special parts according to typical dictionaries. Finally, a residue compensation, which is based on the neighbor reconstruction of patches, is performed to enhance the facial details. The proposed method can effectively exploit the facial properties for face hallucination under the image decomposition perspective. Experimental results demonstrate the effectiveness of our method, in terms of the visual quality of the hallucinated face images.

We extend a recent Sparse Representation-based Classification (SRC) algorithm for face recognition to work on 2D images directly, aiming to reduce the computational complexity whilst still maintaining performance. Our contributions include: (1) a new 2D extension of SRC algorithm; (2) an incremental computing procedure which can reduce the eigen decomposition expense of each 2D-SRC for sequential input data; and (3) extensive numerical studies to validate the proposed methods.

Recently, a simple yet powerful branch-and-bound method called Efficient Subwindow Search (ESS) was developed to speed up sliding window search in object detection. A major drawback of ESS is that its computational complexity varies widely from O(n2) to O(n4) for n × n matrices. Our experimental experience shows that the ESS's performance is highly related to the optimal confidence levels which indicate the probability of the object's presence. In particular, when the object is not in the image, the optimal subwindow scores low and ESS may take a large amount of iterations to converge to the optimal solution and so perform very slow. Addressing this problem, we present two significantly faster methods based on the linear-time Kadane's Algorithm for 1D maximum subarray search. The first algorithm is a novel, computationally superior branch-and-bound method where the worst case complexity is reduced to O(n3). Experiments on the PASCAL VOC 2006 data set demonstrate that this method is significantly and consistently faster (approximately 30 times faster on average) than the original ESS. Our second algorithm is an approximate algorithm based on alternating search, whose computational complexity is typically O(n2). Experiments shows that (on average) it is 30 times faster again than our first algorithm, or 900 times faster than ESS. It is thus well-suited for real time object detection.

Even if the class label information is unknown, side information represents some equivalence constraints between pairs of patterns, indicating whether pairs originate from the same class. Exploiting side information, we develop algorithms to preserve both the intra-class and inter-class local structures. This new type of locality preserving projection (LPP), called LPP with side information (LPPSI), preserves the data’s local structure in the sense that the close, similar training patterns will be kept close, whilst the close but dissimilar ones are separated. Our algorithms balance these conflicting requirements, and we further improve this technique using kernel methods. Experiments conducted on popular face databases demonstrate that the proposed algorithm significantly outperforms LPP. Further, we show that the performance of our algorithm with partial side information (that is, using only small amount of pair-wise similarity/dissimilarity information during training) is comparable with that when using full side information. We conclude that exploiting side information by preserving both similar and dissimilar local structures of the data significantly improves performance.

This paper addresses the limitation of current multilinear techniques (multilinear PCA, multilinear ICA) when applied to face recognition for handling faces in unseen illumination and viewpoints. We propose a new recognition method, exploiting the interaction of all the subspaces resulting from multilinear decomposition (for both multilinear PCA and ICA), to produce a new basis called multilinear-eigenmodes. This basis offers the flexibility to handle face images at unseen illumination or viewpoints. Experiments on benchmarked datasets yield superior performance in terms of both accuracy and computational cost.

Two Dimensional Locality Preserving Projection (2D-LPP) is a recent extension of LPP, a popular face recognition algorithm. It has been shown that 2D-LPP performs better than PCA, 2D-PCA and LPP. However, the computational cost of 2D-LPP is high. This paper proposes a novel algorithm called Ridge Regression for Two Dimensional Locality Preserving Projection (RR-2DLPP), which is an extension of 2D-LPP with the use of ridge regression. RR-2DLPP is comparable to 2D-LPP in performance whilst having a lower computational cost. The experimental results on three benchmark face data sets - the ORL, Yale and FERET databases - demonstrate the effectiveness and efficiency of RR-2DLPP compared with other face recognition algorithms such as PCA, LPP, SR, 2D-PCA and 2D-LPP.

Two Dimensional Linear Discriminant Analysis (2DLDA) has received much interest in recent years. However, 2DLDA could make pairwise distances between any two classes become significantly unbalanced, which may affect its performance. Moreover 2DLDA could also suffer from the small sample size problem. Based on these observations, we propose two novel algorithms called Regularized 2DLDA and Ridge Regression for 2DLDA (RR-2DLDA). Regularized 2DLDA is an extension of 2DLDA with the introduction of a regularization parameter to deal with the small sample size problem. RR-2DLDA integrates ridge regression into Regularized 2DLDA to balance the distances among different classes after the transformation. These proposed algorithms overcome the limitations of 2DLDA and boost recognition accuracy. The experimental results on the Yale, PIE and FERET databases showed that RR-2DLDA is superior not only to 2DLDA but also other state-of-the-art algorithms.

This paper addresses the limitation of current multilinear PCA based techniques, in terms of prohibitive computational cost of testing and poor generalisation in some scenarios, when applied to large training databases. We define person-specific eigenmodes to obtain a set of projection bases, wherein a particular basis captures variation across lightings and viewpoints for a particular person. A new recognition approach is developed utilizing these bases. The proposed approach performs on a par with the existing multilinear approaches, whilst significantly reducing the complexity order of the testing algorithm.

In this paper, we present novel ridge regression (RR) and kernel ridge regression (KRR) techniques for multivariate labels and apply the methods to the problem efface recognition. Motivated by the fact that the regular simplex vertices are separate points with highest degree of symmetry, we choose such vertices as the targets for the distinct individuals in recognition and apply RR or KRR to map the training face images into a face subspace where the training images from each individual will locate near their individual targets. We identify the new face image by mapping it into this face subspace and comparing its distance to all individual targets. An efficient cross-validation algorithm is also provided for selecting the regularization and kernel parameters. Experiments were conducted on two face databases and the results demonstrate that the proposed algorithm significantly outperforms the three popular linear face recognition techniques (Eigenfaces, Fisher faces and Laplacian faces) and also performs comparably with the recently developed Orthogonal Laplacian faces with the advantage of computational speed. Experimental results also demonstrate that KRR outperforms RR as expected since KRR can utilize the nonlinear structure of the face images. Although we concentrate on face recognition in this paper, the proposed method is general and may be applied for general multi-category classification problems.

In this paper we address the spatial activity recognition problem with an algorithm based on Smith-Waterman (SW) local alignment. The proposed SW approach utilises dynamic programming with two dimensional spatial data to quantify sequence similarity. SW is well suited for spatial activity recognition as the approach is robust to noise and can accommodate gaps, resulting from tracking system errors. Unlike other approaches SW is able to locate and quantify activities embedded within extraneous spatial data. Through experimentation with a three class data set, we show that the proposed SW algorithm is capable of recognising accurately and inaccurately segmented spatial sequences. To benchmark the techniques classification performance we compare it to the discrete hidden markov model (HMM). Results show that SW exhibits higher accuracy than the HMM, and also maintains higher classification accuracy with smaller training set sizes. We also confirm the robust property of the SW approach via evaluation with sequences containing artificially introduced noise.

Login access control refers to securing the entry point to a computing system. The login system is responsible for this service and must provide a number of tasks, the main task being that of user authentication. Traditionally, the authentication in Linux and other operating systems is achieved through password verification. This research focused on overcoming the vulnerabilities of password verification, which has been identified as the greatest gap in computer security, through the development of a biometric (facial recognition) login system. The application of biometrics, the automated use of physiological or behavioral characteristics to determine or verify identity, was applied via the eigenface algorithm successfully in this research, resulting in a functional prototype biometric login system for Linux

Given n training examples, the training of a kernel Fisher discriminant (KFD) classifier corresponds to solving a linear system of dimension n. In cross-validating KFD, the training examples are split into 2 distinct subsets for a number of times (L) wherein a subset of m examples is used for validation and the other subset of (n - m) examples is used for training the classifier. In this case L linear systems of dimension (n - m) need to be solved. We propose a novel method for cross-validation of KFD in which instead of solving L linear systems of dimension (n - m), we compute the inverse of an n /spl times/ n matrix and solve L linear systems of dimension 2m, thereby reducing the complexity when L is large and/or m is small. For typical 10-fold and leave-one-out cross-validations, the proposed algorithm is approximately 4 and ( 4/9 n ) times respectively as efficient as the naive implementations. Simulations are provided to demonstrate the efficiency of the proposed algorithms.

In this paper, a new deniable signature scheme, i.e. the committal deniable signature scheme is proposed. This scheme has been constructed by use of bilinear pairings over elliptic curves. In addition to the general property possessed by other deniable signature schemes, there are some new features for this new scheme. One important of them is that signer is not able to forge this type of deniable signatures on behalf of the verifier or any third party. Another feature is that any third party, even though she can obtain the committal deniable signatures by tapping, cannot distinguish the actual signer between the verifier and the signer in the underlying system.

In this paper, a novel approach is developed to evaluate the overall performance of a local area network as well as to monitor some possible intrusion detections. The data is obtained via system utility 'ping' and huge data is analyzed via statistical methods. Finally, an overall performance index is defined and simulation experiments in three months proved the effectiveness of the proposed performance index. A software package is developed based on these ideas.

We present a novel generic tool for data compression and filtering: the generalized Karhunen-Loeve (GKL) transform. The GKL transform minimizes a distance between any given reference and a transformation of some given data where the transform has a predetermined maximum possible rank. The GKL transform is also a generalization of the relative Karhunen-Loeve (RKL) transform by Yamashita and Ogawa (see IEEE Trans. Signal Processing, vol.44, p.661-72, Mar. 1996) where the latter assumes that the given data consist of the given reference (signal) and an independent noise. This letter provides a very simple and yet complete description of the GKL transform and shows useful engineering insights into the GKL transform.