Incomplete multi-view clustering (IMVC) has recently attracted increasing attention and achieved notable progress in computer vision. However, existing IMVC approaches still face several critical challenges. First, most methods fail to capture the inherent nonlinear structures of real-world data. Second, they fail to sufficiently exploit low-rankness and smoothness of multiple views. To overcome these limitations, we propose a novel method, termed Incomplete Multi-View Kernel Subspace Clustering with Tensor-Correlated Total Variation Regularization (KSC-TCTV), which integrates the ability of kernels to capture nonlinear separability with the strength of tensors in characterizing high-order correlations. Specifically, KSC-TCTV first effectively models nonlinear structures by embedding data into a kernel Hilbert space. And then, we introduce a log-based tensor-correlated total variation (Logt-CTV) regularizer in the kernel space, which jointly enforces global low-rankness for inter-view dependency modeling and local smoothness for intra-view structure preservation. The Logt-CTV employs logarithmic non- convex relaxation to mitigate the estimation bias. Experiments on several public benchmark datasets demonstrate that KSC-TCTV outperforms the state-of-the-art IMVC methods.

Color mapping transforms visual data between color representations to enhance perceptual quality, improve compression efficiency by decorrelating color components, ensure consistent reproduction across devices, and facilitate analysis and visualization in computer vision applications. Integer-invertible color mappings enable exact reversibility for lossless image and video coding. Certain integer-invertible variants of the YCuCv and YCoCg color mappings have been adopted in standards such as JPEG-LS, JPEG-2000, JPEG-XL. However, these specialized formulations cannot be readily generalized to other widely used color mappings. Therefore, it is highly desirable to develop a general integer-invertible framework for arbitrary color transforms that achieves lower transformation errors than ad hoc integer-based formulations. This work proposes a lifting-based three triangular factorization, termed ULU0, which enables general integer-invertible color mappings for any non-singular transform matrix. Numerical experiments on five representative mappings (YIQ, YCuCv, YCoCg, YCbCr, and YUV) across forty high-definition and ultra-high definition images demonstrate that the proposed approach achieves complete integer-invertibility in the RGB domain while more accurately preserving the characteristics of the original non-invertible mappings.

Automatic modulation recognition (AMR) is a pivotal technology in modern communication systems. While deep learning(DL) has significantly advanced AMR, effectively leveraging the complementary information in different signal representations(e.g., in-phase/quadrature (I/Q) and amplitude/phase (A/P)) remains challenging. This letter proposes a novel dual-branch network with cooperative supervised contrastive learning (CSCL) for AMR. The CSCL is composed of two main components: 1) a modality-level cooperative learning (CL) module that aligns the projected features of I/Q and A/P sequences onto a common latent space, and 2) an extended supervised contrastive learning (ESCL) module that employs a queue memory mechanism to enrich positive and negative pairs for enhanced feature discriminability. The experimental results on three datasets demonstrate the effectiveness of our proposals and its superiority over other methods.

Video Synthetic Aperture Radar enables high-resolution, continuous imaging of observed scenes under all-weather and day-night conditions. Nevertheless, video SAR image reconstruction remains challenged by substantial data volumes and high computational complexity. This study addresses these limitations by exploiting temporal redundancy in sequential frame data. Through systematic analysis of video SAR data characteristics, we formulate video SAR imaging as a sparse tensor recovery problem by introducing a tailored correlation function to leverage inter-frame dependencies. An iterative solution is derived by integrating the alternating direction method of multipliers (ADMM) and proximal-alternating inexact minimization (P-AIM) frameworks. Based on this formulation, we propose an imaging network (ViSAR-UTNet) by unfolding the iterative process into a Transformer architecture. ViSAR-UTNet comprises two core modules: a weighted self-attention (WSA) mechanism that learns inter-frame correlations and a linearized ADMM (LADMM) operator for sparse tensor recovery. By leveraging the unfolded Transformer structure, ViSAR-UTNet effectively exploits data redundancy, thereby enabling high-quality video reconstruction from reduced measurements. Experiments on synthetic and real datasets are conducted to validate ViSAR-UTNet. The results demonstrate enhanced reconstruction accuracy and computational efficiency of the proposed method.

The occurrence of bubbles in pipeline parallelism is an inherent limitation that can account for more than 40% of the large language model (LLM) long training times and is one of the main reasons for the under-utilization of GPU resources in LLM training. Harvesting these bubbles for GPU side tasks can increase resource utilization and reduce training costs but comes with challenges. First, because bubbles are discontinuous with various shapes, programming side tasks becomes difficult while requiring excessive engineering effort. Second, a side task can compete with pipeline training for GPU resources and incur significant overhead. To address these challenges, we propose FreeRide, a middleware system that harvests the hard-to-utilize bubbles in pipeline parallelism systems to run generic GPU side tasks. FreeRide provides programmers with interfaces to implement side tasks easily, manages bubbles and side tasks during pipeline training, and controls access to GPU resources by side tasks to reduce overhead. We demonstrate that FreeRide achieves almost 8% average cost savings with a negligible overhead of about 1% in training LLMs while serving model training, graph analytics, and image processing side tasks.

Background: Large Language Models often struggle with multi-step reasoning due to cascading errors, rigid prompt structures, and underutilized intermediate reasoning steps. While prompting strategies such as Chain-of-Thought , CoT with Self-Consistency, and Least-to-Most offer partial improvements, they typically lack mechanisms for feedback-driven learning or structured reuse of prior thought sequences. Objectives: This work introduces Recursive Decomposition of Logical Thoughts (RDoLT), a cognitively inspired prompting framework that enhances LLM reasoning through hierarchical decomposition, multi-feature scoring, and knowledge propagation. The framework aims to overcome linear reasoning limitations by enabling structured, memory-aware exploration of thought spaces. Methods: RDoLT executes a three-stage iterative reasoning process across Easy, Intermediate, and Final tiers. At each level, multiple candidate thoughts are generated and scored on Logical Validity, Coherence, Simplicity, and Adaptiveness. The Knowledge Propagation Module (KPM) persistently tracks both selected and rejected thoughts, allowing future reasoning stages to reuse contextually relevant but previously discarded knowledge. The framework supports adaptive thresholding, controlled reasoning depth, and edge-case regeneration through structured feedback loops. Results: Extensive evaluation across five reasoning benchmarks demonstrates that RDoLT outperforms the most competitive prompting strategies in both accuracy and stability. On GSM8K, RDoLT achieves 90.98% accuracy with ChatGPT-4o, surpassing CoT-SC (89.4%) and ReAct (90.5%). It improves Gemma 2 (27B) performance on SVAMP from 69.86% (Vanilla) to 75.27%, and on MultiArith from 67.96% (Vanilla) to 72.49%. Across all benchmarks, RDoLT outperforms or matches the strongest baseline in over 60% of settings, highlighting its robustness across diverse reasoning tasks and model scales. Ablation studies reveal that generating three thoughts per stage yields the best trade-off between performance and efficiency, while the Knowledge Propagation Module (KPM) consistently reduces reasoning variance by leveraging both accepted and discarded thoughts across stages. Conclusions: RDoLT presents a scalable reasoning paradigm grounded in cognitive principles. Its integration of hierarchical decomposition, structured scoring, and selective memory propagation enables more reliable and adaptive reasoning in LLMs. These results establish RDoLT as a robust prompt engineering framework with broad applicability, and future work will focus on optimizing token efficiency and extending to domain-specific use cases.

Diffusion spline filters have been widely studied due to their low computational complexity. The current diffusion spline methods, aimed at external input flows at one moment, are categorized into single spline activation function schemes and multiple parallel spline activation function schemes. The former exhibits limited nonlinearity, whereas the latter, despite improvements, does not procure additional learning information to significantly augment its nonlinear capabilities. Hence, this paper proposes a Diffusiion Pipelined Spline Adaptive Filter (D-PNSF). The filter enhances the model's nonlinear identification capability by cascading spline modules, achieves synchronous processing of asynchronous information, and minimally increases time consumption. Experiments demonstrate that compared to recently proposed diffusion spline filters, D-PNSF can better learn potential nonlinear systems and achieve lower steady-state errors.

Recently, Conformer has achieved tremendous success in speaker verification task. It demonstrates that Transformer-based model can achieve remarkable performance in this domain, bypassing the need for intricate pre-training procedures. However, its special macaron-style feed-forward module introduced prohibitive computing and memory overhead. Speaker verification is often applied in resource-constrained embedded environments like smartphones, where only low memory is available. In light of this, we proposed two approaches to compress the size of the Conformer-based system while maintaining its performance. First, we introduced a lightweight Convolutional Neural Network (CNN) front-end with channel-frequency attention to substitute shallow Conformer blocks. This substitution is aimed at extracting more informative speaker characteristics for subsequent processing. Secondly, we introduced a light Feed-forward Network (FFN) based on depth-wise separable convolution to decrease the model size of Conformer blocks. To better demonstrate the effectiveness of our model, we conducted the evaluation in three different test sets. By incorporating these two approaches, we achieved an Equal Error Rate (EER) of 0.61% on VoxCeleb-O, surpassing the previous state-of-the-art Transformer-based model MFA-Conformer. Moreover, our model has achieved a 60.6% reduction in parameters and a 36.8% reduction in FLOPs compared with MFA-Conformer.

In this work, we focus on the performance benchmarking problem of storage services in cloud-native database systems, which are widely used in various cloud applications. The core idea of these systems is to separate computation and storage in traditional monolithic OLTP databases. Specifically, we first present the characteristics of two representative real I/O workloads at the storage tier of ByteDance's cloud-native database veDB. We then elaborate the limitations of using standard benchmarks such as TPC-C and YCSB to resemble these workloads. To overcome these limitations, we devise a learning-based I/O workload benchmark called CDS-Ben. We demonstrate the superiority of CDSBen by deploying it at ByteDance and showing that its generated I/O traces accurately resemble the real I/O traces in production. Additionally, we verify the accuracy and flexibility of CDSBen by generating a wide range of I/O workloads with different I/O characteristics.

As a popular distributed data warehouse system, Apache Hive has been widely used for big data analytics in many organizations. Meanwhile, exploiting the massive parallelism of GPU to accelerate online analytical processing (OLAP) has been extensively explored in the database community. In this paper, we present GHive, which enhances CPU-based Hive via CPU-GPU heterogeneous computing. GHive is designed for the business intelligence applications and provides the same API as Hive for compatibility. To run SQL queries jointly on both CPU and GPU, GHive comes with three key techniques: (i) a novel data model gTable, which is column-based and enables efficient data movement between CPU memory and GPU memory; (ii) a GPU-based operator library Panda, which provides a complete set of SQL operators with extensively optimized GPU implementations; (iii) a hardware-aware MapReduce job placement scheme, which puts jobs judiciously on either GPU or CPU via a cost-based approach. In the experiments, we observe that GHive outperforms Hive in both query processing speed and operating expense on the Star Schema Benchmark (SSB).

As a distributed, fault-tolerant data warehouse system for large-scale data analytics, Apache Hive has been used for various applications in many organizations (e.g., Facebook, Amazon, and Huawei). Exploiting the large degrees of parallelism of GPU to improve the performance of online analytical processing (OLAP) in database system is a common practice in the industry. Meanwhile, it is a common practice to exploit the large degrees of parallelism of GPU to improve the performance of online analytical processing (OLAP) in database systems. This demo presents GHive, which enables Apache Hive to accelerate OLAP queries by jointly utilizing CPU and GPU in intelligent and efficient ways. The takeaways for SIGMOD attendees include: (1) the superior performance of GHive compared with vanilla Hive that only uses CPU; (2) intuitive visualizations of execution statistics for Hive and GHive to understand where the acceleration of GHive comes from; (3) detailed profiling of the time taken by each operator on CPU and GPU to show the advantages of GPU execution.

This paper introduces two combined-sample multiband-structured subband adaptive filters (MSAFs). In the design, an adaptive convex combination scheme of two self-reliant multi-sampled MSAF (MS-MSAF) with different sampled periods is firstly developed, which leads to the so-called CTMS-MSAF algorithm. Secondly, based on an adaptive filter, the combined-sample MS-MSAF (CMS-MSAF) algorithm is proposed via designing a time-varying sampled period, which possesses lower computational complexity than the former. Then, the convergence behaviors of the CTMS-MSAF and CMS-MSAF algorithms are investigated using standard mean-square deviation analysis. Finally, the simulation study in the system identification and acoustic echo cancellation applications shows that at the same steady-state error, the CMS-MSAF method provides a faster convergence rate than the improved convex combination of two MSAFs, combined-step-size MSAF and CTMS-MSAF algorithms.

Sensor networks are an indispensable part of the Internet of Things (IoT), where sensors perform data acquisition and information processing tasks to obtain the parameters of interest so that IoT-based monitoring, diagnosis and other systems respond quickly to the changing conditions, instantaneous faults, etc. Distributed estimation algorithms are usually employed to estimate the parameters of interest in these IoT-based applications. However, when sensor networks have highly correlated input signals and nonstationary behavior in which the parameters of interest are time-varying, conventional distributed estimation algorithms suffer from severely degraded learning performance due to the large eigenvalue spread in the covariance matrix of the input signals and the random perturbation of the parameters of interest. To address these problems, this paper proposes two diffusion Bayesian subband adaptive filter (DBSAF) algorithms from a Bayesian learning perspective. As the highly-correlated input signal is whitened in a multiband structure and an estimate of the uncertainty in the parameters of interest is obtained by performing Bayesian inference, the proposed DBSAF algorithms are able to achieve better learning performance in comparison with the competing diffusion algorithms. The transient and steady-state mean square error performance of the proposed DBSAF algorithms are analyzed, and are verified by numerical simulations. A lower bound on the time-varying step-size is derived to maintain the optimal steady-state performance in nonstationary scenarios. A new method for the estimation of the noise variance is also proposed. Numerical simulations demonstrate the excellent learning performance of the proposed algorithms in comparison with benchmark algorithms.

In this paper, we present a robust variable step-size decorrelation normalized least-mean-square (RVSSDNLMS) algorithm. A new constrained minimization problem is developed by minimizing the l2norm of the a decorrelated posteriori error signal with a constraint on the filter coefficients in the l2norm sense. Solving this minimization problem gives birth to the efficient RVSSDNLMS algorithm. The convergence performance and computational complexity of RVSSDNLMS algorithm are analyzed. Finally, simulations show that the proposed RVSSDNLMS considerably outperforms the normalized least-mean-square (NLMS), robust variable step-size NLMS, and pseudoaffine projection algorithms in terms of convergence rate and steady-state error in Gaussian noise and impulsive noise environments.

The Ice, Cloud and land Elevation Satellite-2 (ICESat-2) mission of the National Aeronautics and Space Administration is scheduled to launch in 2017. This upcoming mission aims to provide data to determine the temporal and spatial changes of ice sheet elevation, sea ice freeboard, and vegetation canopy height. A photon-counting lidar onboard ICESat-2 yields point clouds resulting from surface returns and noise. In support of the ICESat-2 mission, this letter derives an adaptive density-based model that is capable of detecting the ground surface and vegetation canopy in photon-counting laser altimeter data. Based on results from point clouds generated by a first principle simulation and those observed by the Multiple Altimeter Beam Experimental Lidar, the ground and canopy returns can be reliably extracted using the proposed approach. Further study on performance assessment shows that smoother surfaces will result in improved accuracy of ground height estimation. In addition, the proposed detection approach has better performance in environments with lower noise, although the performance evaluation metric F-measure does not vary significantly over a range of noise rates (0.5-5 MHz). This proposed approach is generally applicable for surface and canopy finding from photon-counting laser altimeter data.

In order to provide high-security remote authentication, the three-factor authentication scheme combining biometric with smart card and password has been proposed. With a careful review of the recently proposed Lee-Hsu's scheme, this paper points out some design flaws in it. In order to withstand them, a new scheme based on Fuzzy extractors is proposed. With a detail analysis, this paper demonstrates that the proposed scheme is more practical and reasonable. It also has higher security and deals with biometric more appropriately in spite of higher computation cost at client than Lee-Hsu's scheme. Furthermore, an access control method has been introduced in it for the purpose of making different users enjoy different access privileges with regard to the data. What is more, the proposed scheme can also achieve key agreement. Copyright © 2014 John Wiley & Sons, Ltd.

Observations from satellite lidar instruments have provided evidence in the remarkable changes in polar ice sheets on a global scale. The Ice, Cloud and land Elevation Satellite-2 (ICESat-2) is scheduled for launch by NASA in 2017 and will monitor the elevation changes of polar ice sheets and vegetation canopy. To validate ICESat-2's approach of photon-counting laser altimetry, measurements obtained from the Multiple Altimeter Beam Experimental Lidar (MABEL) instrument are critical. In support of the ICESat-2 mission, this paper derives an algorithm for the detection of ground and vegetation canopy in photon-counting laser altimeter data. This approach uses a density-based clustering model and modifies the shape of search area. Based on results from MABEL observations, the proposed approach is seen to be robust in detecting ground and vegetation canopy as well as background noise reduction. In addition, this approach can be quickly implemented and adaptive to photon-counting lidar data sets with different point cloud densities.

Conventional local preserving projection (LPP) is sensitive to outliers because its objective function is based on the L2-norm distance criterion and suffers from the small sample size (SSS) problem. To improve the robustness of LPP against outliers, LPP-L1 uses L1-norm distance metric. However, LPP-L1 does not work ideally when there are larger outliers. We propose a more robust version of LPP, called LPP-MCC, which formulates the objective problem based on maximum correntropy criterion (MCC). The objective problem is efficiently solved via a half-quadratic optimization procedure and the complicated non-linear optimization procedure can thereby be reduced to a simple quadratic optimization at each iteration. Moreover, LPP-MCC avoids the SSS problem because the generalized eigenvalues computation is not involved in the optimization procedure. The experimental results on both synthetic and real-world databases demonstrate that the proposed method can outperform LPP and LPP-L1 when there are large outliers in the training data.

The step-size in well-known variable step-size least mean square (VSSLMS) is updated as μn+1= αμn+ γen2with , γ > 0, and μn+1is set to μminor μmaxwhen it falls below or above these lower and upper bounds, respectively. It provides fast convergence at early stages of adaptation while ensuring small steady-state misalignment. This paper considers the steady-state performance of the VSSLMS in non-Gaussian noise environments. The contribution of the paper to the VSSLMS is threefold; (1) when γ ≪ 1 - α, the VSSLMS has low steady-state misalignment. (2) when α ≪ 1, the VSSLMS achieves different steady-state misalignments for different noise distributions. (3) In theory, there are different optimal values α for different noise distributions, i.e., 0.17 (Gaussian distribution), 0.21 (Student distribution), 0.38 (Laplace distribution), 0 (Binary and Uniform distributions). Analytical results are compared with simulations and are shown to agree well.

To advance the science of lidar sensing of complex surfaces as well as in support of the upcoming Ice, Cloud, and land Elevation Satellite-2 (ICESat-2) mission, this paper establishes a framework that simulates the performance of a spaceborne micropulse photon-counting detector system on a complex surface. A first-principle 3-D Monte Carlo approach is used to investigate returning photon distributions. The photomultiplier tube (PMT) detector simulation takes into account detector dead time and multiple pixels based on the latest ICESat-2 design, as well as photon detection efficiency for probabilistic modeling. To explore system behavior, Fourier synthesis is introduced to create a synthetic surface based on parameters derived from a real data set. A radiometric model using bidirectional reflection distribution functions is also applied in the synthetic scene. Such an approach allows the study of surface elevation retrieval accuracy for landscapes which have different shapes as well as reflectivities. As a case study, returning photon detection on an example snow surface is explored. Based on the simulation results for lidar sensing on synthetic complex surfaces with an elevation range of 10 m across the scene, the spaceborne photon-counting lidar system considered here is seen to have a derived elevation bias of up to 2 cm and a error standard deviation of 10 cm. Further study on multiple-pixel PMT performance for complex surfaces demonstrates that a less rough surface will result in higher accuracy and a surface with a smaller diffuse albedo will result in smaller bias.

Conventional discriminant locality preserving projection (DLPP) is a dimensionality reduction technique based on manifold learning, which has demonstrated good performance in pattern recognition. However, because its objective function is based on the distance criterion using L2-norm, conventional DLPP is not robust to outliers which are present in many applications. This paper proposes an effective and robust DLPP version based on L1-norm maximization, which learns a set of local optimal projection vectors by maximizing the ratio of the L1-norm-based locality preserving between-class dispersion and the L1-norm-based locality preserving within-class dispersion. The proposed method is proven to be feasible and also robust to outliers while overcoming the small sample size problem. The experimental results on artificial datasets, Binary Alphadigits dataset, FERET face dataset and PolyU palmprint dataset have demonstrated the effectiveness of the proposed method.

Sparse representation based classification (SRC) could not well classify the sample belonging to different classes distribute on the same direction. To solve the problem, a Volterra kernel sparse representation based classification (Volterra-SRC) algorithm is proposed in this paper. Firstly, the original face images are divided into non overlapped patches and then mapped into a high dimensional space by utilizing the Volterra kernels. During the training stage, following by the Fisher criteria, the objective function is defined to obtain the optimal Volterra kernels via maximizing inter-class distances and minimizing intra-class distances simultaneously. During the testing stage, a voting procedure is introduced in conjunction with a sparse representation based classification to decide to which class each individual patch belongs. Finally, the aggregate classification results of all patches in a face are used to determine the overall recognition outcome for the given face image. We demonstrate the experiments on ORL and Extended Yale B benchmark face databases and show that our proposed Volterra-SRC algorithm consistently outperforms the original SRC and the proposed has some advantages and robustness in case of small train number samples.

To advance the science of lidar sensing of complex ice and snow surfaces as well as in support of the upcoming ICESat- 2 mission, this paper establishes a framework to theoretically study a spaceborne micropulselidar returns from snow and ice surfaces. First, the anticipated lidar return characteristics for a sloped non-penetrating surface is studied when measured by a multiple-channel photon-counting detector. Second, an analytical snow reflectance model based on experimental observations is applied in synthetic scene. Based on the simulation results, the spaceborne photon-counting lidar system considered here is seen to have moderate detectability on snow surfaces. In addition, for the penetrating snow model considered here, it is shown that slightly sloped snow terrain with larger snow grain size will result in smaller elevation bias.

Linear discriminant analysis (LDA) is a well-known dimensionality reduction technique, which is widely used for many purposes. However, conventional LDA is sensitive to outliers because its objective function is based on the distance criterion using L2-norm. This paper proposes a simple but effective robust LDA version based on L1-norm maximization, which learns a set of local optimal projection vectors by maximizing the ratio of the L1-norm-based between-class dispersion and the L1-norm-based within-class dispersion. The proposed method is theoretically proved to be feasible and robust to outliers while overcoming the singular problem of the within-class scatter matrix for conventional LDA. Experiments on artificial datasets, standard classification datasets and three popular image databases demonstrate the efficacy of the proposed method.

Lidar sensing has been found to be a useful method of monitoring the dynamics and mass balance of glaciers, ice caps, and ice sheets. However, it is also known that ice surfaces can have complex 3-dimensional structure, which can challenge their accurate retrieval with lidar sensing. In support of future lidar sensing satellite missions, such as the upcoming ICESat-2, a joint research project was recently initiated between the Rochester Institute of Technology (RIT) and the University at Buffalo to study lidar sensing of complex ice surfaces. This effort is supported by NASA's Remote Sensing Theory program and is aimed at advancing the science of lidar sensing. The general approach is to 1) define realistic complex ice surfaces, 2) render lidar image simulations, and 3) compare the resulting data to the known surfaces to gain insight into the phenomenology of lidar sensing of snow and ice. The project will build on existing scientific understanding of light scattering from snow and ice as well as lidar sensor system modeling with a systems engineering end-to-end perspective. Initial results show the simulations capturing realistic scattering of photons in snow volumes and the resulting point clouds measured by a model spaceborne lidar system.

In this paper, we present the performance analysis of a self-recovery fragile watermarking scheme using block-neighbor- hood tamper characterization. This method uses a pseudorandom sequence to generate the nonlinear block-mapping and employs an optimized neighborhood characterization method to detect the tampering. Performance of the proposed method and its resistance to malicious attacks are analyzed. We also investigate three optimization strategies that will further improve the quality of tamper localization and recovery. Simulation results demonstrate that the proposed method allows image recovery with an acceptable visual quality (peak signal-to-noise ratio (PSNR) as 25 dB) up to 60% tampering.

Discriminant locality preserving projection (DLPP) is a linear approach that encodes discriminant information into the objective of locality preserving projection and improves its classification ability. To enhance the nonlinear description ability of DLPP, we can optimize the objective function of DLPP in reproducing kernel Hilbert space to form a kernel-based discriminant locality preserving projection (KDLPP). However, KDLPP suffers the following problems: 1) larger computational burden; 2) no explicit mapping functions in KDLPP, which results in more computational burden when projecting a new sample into the low-dimensional subspace; and 3) KDLPP cannot obtain optimal discriminant vectors, which exceedingly optimize the objective of DLPP. To overcome the weaknesses of KDLPP, in this paper, a direct discriminant locality preserving projection with Hammerstein polynomial expansion (HPDDLPP) is proposed. The proposed HPDDLPP directly implements the objective of DLPP in high-dimensional second-order Hammerstein polynomial space without matrix inverse, which extracts the optimal discriminant vectors for DLPP without larger computational burden. Compared with some other related classical methods, experimental results for face and palmprint recognition problems indicate the effectiveness of the proposed HPDDLPP.

To combat the linear and non-linear distortions for time-invariant and time-variant channels, a novel adaptive joint process equaliser based on a pipelined decision feedback recurrent neural network (JPDFRNN) is proposed in this paper. The JPDFRNN consists of a number of simple small-scale decision feedback recurrent neural network (DFRNN) modules and a linear combiner. The cascaded DFRNN provides pre-processing for the linear combiner. Moreover, each DFRNN can provide a local interpolation for M sample points; the final linear combiner presents a global interpolation with good localisation properties. Furthermore, since those modules of non-linear subsection can be performed simultaneously in a pipelined parallelism fashion, this would result in a significant improvement in the total computational efficiency. Simulation results show that the performance of the JPDFRNN using the modified real-time recurrent learning (RTRL) algorithm is superior to that of the DFRNN and RNN for the non-linear time-invariant and time-variant channels.

This paper, presents a novel unsupervised dimensionality reduction approach called variance difference embedding (VDE) for facial feature extraction. The proposed VDE method is derived from maximizing the difference between global variance and local variance, so it can draw the close samples closer and simultaneously making the mutually distant samples even more distant from each other. VDE utilizes the maximum variance difference criterion rather than the generalized Rayleigh quotient as a class separability measure, thereby avoiding the singularity problem when addressing the sample size problem. The results of the experiments conducted on ORL database, Yale database and a subset of PIE database indicate the effectiveness of the proposed VDE method on facial feature extraction and classification.

In this letter, a novel pipelined decision feedback RNN equalizer (PDFRNE) with low computational complexity is proposed. Since each module is a DFRNN with the decision feedback structure so that it can eliminate the past error remaining in the network. Moreover, the performance can be further improved. At the same time, it can overcome the unstableness due to its nature of the infinite impulse response (IIR) structure.

A novel nonlinear adaptive filter with pipelined Chebyshev functional link artificial recurrent neural network (PCFLARNN) is presented in this paper, which uses a modification real-time recurrent learning algorithm. The PCFLARNN consists of a number of simple small-scale Chebyshev functional link artificial recurrent neural network (CFLARNN) modules. Compared to the standard recurrent neural network (RNN), those modules of PCFLARNN can simultaneously be performed in a pipelined parallelism fashion, and this would lead to a significant improvement in its total computational efficiency. Furthermore, contrasted with the architecture of a pipelined RNN (PRNN), each module of PCFLARNN is a CFLARNN whose nonlinearity is introduced by enhancing the input pattern with Chebyshev functional expansion, whereas the RNN of each module in PRNN utilizing linear input and first-order recurrent term only fails to utilize the high-order terms of inputs. Therefore, the performance of PCFLARNN can further be improved at the cost of a slightly increased computational complexity. In addition, due to the introduced nonlinear functional expansion of each module in PRNN, the number of input signals can be reduced. Computer simulations have demonstrated that the proposed filter performs better than PRNN and RNN for nonlinear colored signal prediction, nonstationary speech signal prediction, and chaotic time series prediction.

This paper proposes a novel computational efficient adaptive nonlinear equalizer based on combination of finite impulse response (FIR) filter and functional link artificial neural network (CFFLANN) to compensate linear and nonlinear distortions in nonlinear communication channel. This convex nonlinear combination results in improving the speed while retaining the lower steady-state error. In addition, since the CFFLANN needs not the hidden layers, which exist in conventional neural-network-based equalizers, it exhibits a simpler structure than the traditional neural networks (NNs) and can require less computational burden during the training mode. Moreover, appropriate adaptation algorithm for the proposed equalizer is derived by the modified least mean square (MLMS). Results obtained from the simulations clearly show that the proposed equalizer using the MLMS algorithm can availably eliminate various intensity linear and nonlinear distortions, and be provided with better anti-jamming performance. Furthermore, comparisons of the mean squared error (MSE), the bit error rate (BER), and the effect of eigenvalue ratio (EVR) of input correlation matrix are presented.

This paper presents a new digital watermarking method based on chaotic maps. Different from most of the existing chaotic watermarking methods, two chaotic maps are incorporated into the watermarking system to resolve the finite word length effect and to improve the system's resistance to attacks. One map is used for watermark generation, another as the private key. Simulation results show that the proposed digital watermarking system is feasible and robust to common signal processing procedures. In addition, it performs better than existing watermarking schemes and exhibits better protection than watermarking systems using only one chaotic map

The work presents an automatic and efficient video object segmentation algorithm. Moving object extraction is carried out by adaptive background model and edge-based change detection techniques. The background is updated by use of pixel history and a moving object mask. Connected component analysis and morphological filtering are employed to obtain an accurate VOP (video object plane). Finally, a novel object tracking window scheme is applied to improve the processing speed. Experimental results for three different types of MPEG-4 video sequences are shown to demonstrate the effectiveness of the proposed algorithm.