Search

Freezing of gait (FoG) is identified as a sudden and brief episode of movement cessation despite the intention to continue walking. It is one of the most disabling symptoms of Parkinson's disease (PD) and often leads to falls and injuries. Many computer-aided FoG detection methods have been proposed to use data collected from unimodal sources, such as motion sensors, pressure sensors, and video cameras. However, there are limited efforts of multimodal-based methods to maximize the value of all the information collected from different modalities in clinical assessments and improve the FoG detection performance. Therefore, in this study, a novel end-to-end deep architecture, namely graph fusion neural network (GFN), is proposed for multimodal learning-based FoG detection by combining footstep pressure maps and video recordings. GFN constructs multimodal graphs by treating the encoded features of each modality as vertex-level inputs and measures their adjacency patterns to construct complementary FoG representations, thus reducing the representation redundancy among different modalities. In addition, since GFN is devised to process multimodal graphs of arbitrary structures, it is expected to achieve superior performance with inputs containing missing modalities, compared to the alternative unimodal methods. A multimodal FoG dataset was collected, which included clinical assessment videos and footstep pressure sequences of 340 trials from 20 PD patients. Our proposed GFN demonstrates a great promise of multimodal FoG detection with an area under the curve (AUC) of 0.882. To the best of our knowledge, this is one of the first studies to utilize multimodal learning for automated FoG detection, which offers significant opportunities for better patient assessments and clinical trials in the future.

The goal of hyperspectral image fusion (HIF) is to reconstruct high spatial resolution hyperspectral images (HR-HSI) via fusing low spatial resolution hyperspectral images (LR-HSI) and high spatial resolution multispectral images (HR-MSI) without loss of spatial and spectral information. Most existing HIF methods are designed based on the assumption that the observation models are known, which is unrealistic in many scenarios. To address this blind HIF problem, we propose a deep learning-based method that optimizes the observation model and fusion processes iteratively and alternatively during the reconstruction to enforce bidirectional data consistency, which leads to better spatial and spectral accuracy. However, general deep neural network inherently suffers from information loss, preventing us to achieve this bidirectional data consistency. To settle this problem, we enhance the blind HIF algorithm by making part of the deep neural network invertible via applying a slightly modified spectral normalization to the weights of the network. Furthermore, in order to reduce spatial distortion and feature redundancy, we introduce a Content-Aware ReAssembly of FEatures module and an SE-ResBlock model to our network. The former module helps to boost the fusion performance, while the latter make our model more compact. Experiments demonstrate that our model performs favorably against compared methods in terms of both nonblind HIF fusion and semiblind HIF fusion.

Continues field monitoring and searching sensor data remains an imminent element emphasizes the influence of the Internet of Things (IoT). Most of the existing systems are concede spatial coordinates or semantic keywords to retrieve the entail data, which are not comprehensive constraints because of sensor cohesion, unique localization haphazardness. To address this issue, we propose deep learning inspired sensor-rank consolidation (DLi-SRC) system that enables 3-set of algorithms. First, sensor cohesion algorithm based on Lyapunov approach to accelerate sensor stability. Second, sensor unique localization algorithm based on rank-inferior measurement index to avoid redundancy data and data loss. Third, a heuristic directive algorithm to improve entail data search efficiency, which returns appropriate ranked sensor results as per searching specifications. We examined thorough simulations to describe the DLi-SRC effectiveness. The outcomes reveal that our approach has significant performance gain, such as search efficiency, service quality, sensor existence rate enhancement by 91%, and sensor energy gain by 49% than benchmark standard approaches.

A common drawback of EEG-based emotion recognition is that volume conduction effects of the human head introduce interchannel dependence and result in highly correlated information among most EEG features. These highly correlated EEG features cannot provide extra useful information, and they actually reduce the performance of emotion recognition. However, the existing feature selection methods, commonly used to remove redundant EEG features for emotion recognition, ignore the correlation between the EEG features or utilize a greedy strategy to evaluate the interdependence, which leads to the algorithms retaining the correlated and redundant features with similar feature scores in the EEG feature subset. To solve this problem, we propose a novel EEG feature selection method for emotion recognition, termed global redundancy minimization in orthogonal regression (GRMOR). GRMOR can effectively evaluate the dependence among all EEG features from a global view and then select a discriminative and nonredundant EEG feature subset for emotion recognition. To verify the performance of GRMOR, we utilized three EEG emotional data sets (DEAP, SEED, and HDED) with different numbers of channels (32, 62, and 128). The experimental results demonstrate that GRMOR is a promising tool for redundant feature removal and informative feature selection from highly correlated EEG features.

Compared with traditional convolutions, grouped convolutional neural networks are promising for both model performance and network parameters. However, existing models with the grouped convolution still have parameter redundancy. In this article, concerning the grouped convolution, we propose a sharing grouped convolution structure to reduce parameters. To efficiently eliminate parameter redundancy and improve model performance, we propose a Bayesian sharing framework to transfer the vanilla grouped convolution to be the sharing structure. Intragroup correlation and intergroup importance are introduced into the prior of the parameters. We handle the Maximum Type II likelihood estimation problem of the intragroup correlation and intergroup importance by a group LASSO-type algorithm. The prior mean of the sharing kernels is iteratively updated. Extensive experiments are conducted to demonstrate that on different grouped convolutional neural networks, the proposed sharing grouped convolution structure with the Bayesian sharing framework can reduce parameters and improve prediction accuracy. The proposed sharing framework can reduce parameters up to 64.17%. For ResNeXt-50 with the sharing grouped convolution on ImageNet dataset, network parameters can be reduced by 96.875% in all grouped convolutional layers, and accuracies are improved to 78.86% and 94.54% for top-1 and top-5, respectively.

CNN-based salient object detection (SOD) methods achieve impressive performance. However, the way semantic information is encoded in them and whether they are category-agnostic is less explored. One major obstacle in studying these questions is the fact that SOD models are built on top of the ImageNet pre-trained backbones which may cause information leakage and feature redundancy. To remedy this, here we first propose an extremely light-weight holistic model tied to the SOD task that can be freed from classification backbones and trained from scratch, and then employ it to study the semantics of SOD models. With the holistic network and representation redundancy reduction by a novel dynamic weight decay scheme, our model has only 100K parameters, ∼ 0.2% of parameters of large models, and performs on par with SOTA on popular SOD benchmarks. Using CSNet, we find that a) SOD and classification methods use different mechanisms, b) SOD models are category insensitive, c) ImageNet pre-training is not necessary for SOD training, and d) SOD models require far fewer parameters than the classification models. The source code is publicly available at https://mmcheng.net/sod100k/.

Feature selection has attracted a lot of attention in obtaining discriminative and non-redundant features from high-dimension data. Compared with traditional filter and wrapper methods, embedded methods can obtain a more informative feature subset by fully considering the importance of features in the classification tasks. However, the existing embedded methods emphasize the above importance of features and mostly ignore the correlation between the features, which leads to retain the correlated and redundant features with similar scores in the feature subset. To solve the problem, we propose a novel supervised embedded feature selection framework, called feature selection under global redundancy minimization in orthogonal regression (GRMOR). The proposed framework can effectively recognize redundant features from a global view of redundancy among the features. We also incorporate the large margin constraint into GRMOR for robust multi-class classification. Compared with the traditional embedded methods based on least square regression, the proposed framework utilizes orthogonal regression to preserve more discriminative information in the subspace, which can help accurately rank the importance of features in the classification tasks. Experimental results on twelve public datasets demonstrate that the proposed framework can obtain superior classification performance and redundancy removal performance than twelve other feature selection methods.

Large deep neural network (DNN) models pose the key challenge to energy efficiency due to the significantly higher energy consumption of off-chip DRAM accesses than arithmetic or SRAM operations. It motivates the intensive research on model compression with two main approaches. Weight pruning leverages the redundancy in the number of weights and can be performed in a non-structured, which has higher flexibility and pruning rate but incurs index accesses due to irregular weights, or structured manner, which preserves the full matrix structure with a lower pruning rate. Weight quantization leverages the redundancy in the number of bits in weights. Compared to pruning, quantization is much more hardware-friendly and has become a "must-do" step for FPGA and ASIC implementations. Thus, any evaluation of the effectiveness of pruning should be on top of quantization. The key open question is, with quantization, what kind of pruning (non-structured versus structured) is most beneficial? This question is fundamental because the answer will determine the design aspects that we should really focus on to avoid the diminishing return of certain optimizations. This article provides a definitive answer to the question for the first time. First, we build ADMM-NN-S by extending and enhancing ADMM-NN, a recently proposed joint weight pruning and quantization framework, with the algorithmic supports for structured pruning, dynamic ADMM regulation, and masked mapping and retraining. Second, we develop a methodology for fair and fundamental comparison of non-structured and structured pruning in terms of both storage and computation efficiency. Our results show that ADMM-NN-S consistently outperforms the prior art: 1) it achieves 348× , 36× , and 8× overall weight pruning on LeNet-5, AlexNet, and ResNet-50, respectively, with (almost) zero accuracy loss and 2) we demonstrate the first fully binarized (for all layers) DNNs can be lossless in accuracy in many cases. These results provide a strong baseline and credibility of our study. Based on the proposed comparison framework, with the same accuracy and quantization, the results show that non-structured pruning is not competitive in terms of both storage and computation efficiency. Thus, we conclude that structured pruning has a greater potential compared to non-structured pruning. We encourage the community to focus on studying the DNN inference acceleration with structured sparsity.

It is often tricky in real-life classification applications to select model features that would ensure an adequate sample classification, given a large number of candidate features. Our main contribution is threefold: (1) Evaluate the relevance and redundancy of feature. (2) Define the feature selection problem as eigenvalue computation problem with linear constraint. (3) Select the best features in an efficient way. We considered 20 UCI benchmark datasets to validate and test our approach. The results were compared with those obtained using one of the more widely used approaches, namely mRMR, the conventional features and two moderns state-of-the-art approaches. The experimental results revealed that our approach could improve the classification task, using only 20 % of the conventional features.

Aiming at the problem of effectively selecting relevant features from heterogeneous data without decision, a novel feature selection approach is studied based on fuzzy mutual information in fuzzy rough set theory. First, the fuzzy relevance of each feature is defined by using fuzzy mutual information, and then the fuzzy conditional relevance is further given. Next, the fuzzy redundancy is defined by using the difference between the fuzzy relevance and the fuzzy conditional relevance. Thereby the evaluation index of the feature importance is obtained by using the idea of unsupervised minimum redundancy and maximum relevance. Finally, a fuzzy mutual information-based unsupervised feature selection (FMIUFS) algorithm is designed to select feature sequences. Extensive experiments are conducted on public datasets and six unsupervised feature selection algorithms are compared. The selected features are evaluated by classification, clustering, and outlier detection methods. Experimental results show that the proposed algorithm can select fewer heterogeneous features to maintain or improve the performance of learning algorithms.

In the natural language processing family, learning representations is a pioneering study, especially in sequence-to-sequence tasks where outputs are generated, totally relying on the learning representations of source sequence. Generally, classic methods infer that each word occurring in the source sequence, having more or less influence on the target sequence, should all be considered when generating outputs. As the summarization task requires the output sequence to only retain the essence, classic full consideration of the source sequence may not work well on it, which calls for more suitable methods with the ability to discard the misleading noise words. Motivated by this, with both relevance retaining and redundancy removal in mind, we propose a summarization learning model by implementing an encoder with copious contextual information represented and a decoder with a selecting mechanism integrated. Specifically, we equip the encoder with an asynchronous bi directional parallel structure, in order to obtain abundant semantic representation. The decoder, different from the classic attention-based works, employs a self-aware context selecting mechanism to generate summary in a more productive way. We evaluate the proposed methods on three benchmark summarization corpora. The experimental results demonstrate the effectiveness and applicability of the proposed framework in relation to several well-practiced and state-of-the-art summarization methods.

The day-to-day growth of online information necessitates intensive research in automatic text summarization (ATS). The ATS software produces summary text by extracting important information from the original text. With the help of summaries, users can easily read and understand the documents of interest. Most of the approaches for ATS used only local properties of text. Moreover, the numerous properties make the sentence selection difficult and complicated. So this article uses a graph based summarization to utilize structural and global properties of text. It introduces maximal clique based sentence selection (MCBSS) algorithm to select important and non-redundant sentences that cover all concepts of the input text for summary. The MCBSS algorithm finds novel information using maximal cliques (MCs). The experimental results of recall oriented understudy for gisting evaluation (ROUGE) on Timeline dataset show that the proposed work outperforms the existing graph algorithms Bushy Path (BP), Aggregate Similarity (AS), and TextRank (TR).

Deep neural networks (DNNs) have become a widely deployed model for numerous machine learning applications. However, their fixed architecture, substantial training cost, and significant model redundancy make it difficult to efficiently update them to accommodate previously unseen data. To solve these problems, we propose an incremental learning framework based on a grow-and-prune neural network synthesis paradigm. When new data arrive, the neural network first grows new connections based on the gradients to increase the network capacity to accommodate new data. Then, the framework iteratively prunes away connections based on the magnitude of weights to enhance network compactness, and hence recover efficiency. Finally, the model rests at a lightweight DNN that is both ready for inference and suitable for future grow-and-prune updates. The proposed framework improves accuracy, shrinks network size, and significantly reduces the additional training cost for incoming data compared to conventional approaches, such as training from scratch and network fine-tuning. For the LeNet-300-100 (LeNet-5) neural network architectures derived for the MNIST dataset, the framework reduces training cost by up to 64% (67%), 63% (63%), and 69% (73%) compared to training from scratch, network fine-tuning, and grow-and-prune from scratch, respectively. For the ResNet-18 architecture derived for the ImageNet dataset (DeepSpeech2 for the AN4 dataset), the corresponding training cost reductions against training from scratch, network fine-tunning, and grow-and-prune from scratch are 64% (67%), 60% (62%), and 72% (71%), respectively. Our derived models contain fewer network parameters but achieve higher accuracy relative to conventional baselines.

In this paper, we improve the inter coding performance of surveillance videos by simultaneously investigating the distinct characteristics of background and foreground redundancy, and introduce two novel reference frames in a complementary manner. On one hand, a block level background reference frame (BRF) is proposed to reduce the background redundancy. The proposed scheme incorporates semantic information into the compression process, and makes use of instance segmentation to facilitate the background block decision, making the generated BRF free from foreground pollution. On the other hand, in order to handle foreground redundancy, a foreground reference frame (FRF) is generated based on Surveillance Prediction Generative Adversarial Network (SP-GAN), which utilizes previous reconstructed frames, optical flow based prediction, as well as BRF to infer the foreground objects of the to-be-coded frame. We integrate the proposed scheme into HM-16.6 software and append BRF and FRF to the reference pictures list (RPS). Simulation results demonstrate considerable superiority of the proposed scheme. In particular, by adding the proposed BRF to RPS, 3% coding gains are observed compared with the state-of-the-art BRF method. When both BRF and FRF are incorporated into RPS, 5.8% gains are achieved for surveillance video coding.

The plethora of ways of data representation and their applications to system modeling is inherently associated with dimensionality reduction. In a nutshell, the result of dimensionality reduction should support efficient ways of constructing ensuing models (classifiers, predictors) as well as an interpretation of the data themselves. Furthermore, there should be a suitable measure quantifying the quality of data positioned in the reduced space. We advocate that what makes the reduced data interpretable, goes hand in hand with revealing a logic fabric of the data, suppressing redundancy, and finally arriving at a logic description of data. The anticipation is that the reduced data can be described in the form of logic expressions formed over the original highly dimensional data. Evidently, having these above stated points in mind, the aim of this study is two-fold: (i) to develop a logic-oriented data representation, and (ii) to quantify the quality of results of dimensionality reduction by incorporating a facet of information granularity. In other words, we argue that the result of dimensionality reduction gives rise to information granules whose level of granularity associates with the quality of processing completed by the autoencoder. In light of the recent surge of architectures of deep learning, the study is focused on the construction and analysis of logic-oriented autoencoders. We propose a two-level architecture composed of the logic-oriented processing units (and processing carried out at the first layer of the autoencoder) followed by the or processing completed at the second layer. As data representation provided by the autoencoder is not ideal, we augment the original architecture by granular parameters which give rise to granular logic-oriented autoencoders. A suite of experiments is also reported.

Band selection is one of the most effective methods to reduce the band redundancy of hyperspectral images (HSIs). Most existing band selection methods tend to regard each band as a whole, and then explore the band redundancy with the pixel-wise features directly. However, since the regions of HSIs corresponding to different objects have diverse spectral properties and spatial structure, such above scheme limits the performance of hyperspectral band selection due to the lack of spatial information. To address above issues, a novel band selection method via region-aware latent features fusion based clustering (RLFFC) is proposed. Specifically, we employ the superpixel segmentation to segment HSIs into multiple regions so that the spatial information of HSIs can be fully preserved. In order to capture the priori information, we construct its corresponding Laplacian matrix from which a group of low dimensional latent features are generated to further enhance the separability among different bands. Then, a shared latent feature representation of HSIs is obtained by fusing region-aware latent features to effectively capture the band redundancy of HSIs. Finally, the k -means clustering algorithm is utilized to obtain the index of the selected bands from the shared latent feature representation. As a result, the spectral and spatial properties are well exploited in the proposed method. Extensive experiments on four public hyperspectral datasets show that the proposed method achieves superior performance when compared with other state-of-the-art ones.

Crowdsourcing is a hotspot research field which can facilitate machine learning by collecting labels to train models. Consequently, the state-of-the-art research efforts in crowdsourcing focus on truth inference or label integration, to remove inconsistent labels or to alleviate biased labeling. In turn, the integrated labels will be used to fine-tune machine learning models. Particularly, in this paper, we change the target of truth inference in crowdsourcing from discrete labels to multiple comments given by online participants, that is, the integration of the crowdsourced comments. For such a goal, we propose a Self-play and Sentiment-Emphasized Comment Integration Framework (SSECIF), based on deep Q-learning, with three unique features. First, our framework SSECIF can generate the comment integration in a totally self-play way, without relying on the ground truth generated by human effort. Second, the integrated comment generated by SSECIF can include salient content with low redundancy. Third, the proposed framework SSECIF has emphasized, with a higher intensity, the sentiment in the integrated comment, in order to reflect the attitude or opinion more obviously. Extensive evaluation on real-world datasets demonstrates that SSECIF has achieved the best overall performance in terms of both effectiveness and efficiency, compared with the state-of-the-art methods. Index Terms: Crowdsourcing; Comment Integration; Reinforcement Learning; Deep Q-Learning; Sentiment Analysis.

Visual grounding (VG) aims to locate the most relevant object or region in an image, based on a natural language query. Generally, it requires the machine to first understand the query, identify the key concepts in the image, and then locate the target object by specifying its bounding box. However, in many real-world visual grounding applications, we have to face with ambiguous queries and images with complicated scene structures. Identifying the target based on highly redundant and correlated information can be very challenging, and often leading to unsatisfactory performance. To tackle this, in this paper, we exploit an attention module for each kind of information to reduce internal redundancies. We then propose an accumulated attention (A-ATT) mechanism to reason among all the attention modules jointly. In this way, the relation among different kinds of information can be explicitly captured. Moreover, to improve the performance and robustness of our VG models, we additionally introduce some noises into the training procedure to bridge the distribution gap between the human-labeled training data and the real-world poor quality data. With this "noised" training strategy, we can further learn a bounding box regressor, which can be used to refine the bounding box of the target object. We evaluate the proposed methods on four popular datasets (namely ReferCOCO, ReferCOCO+, ReferCOCOg, and GuessWhat?!). The experimental results show that our methods significantly outperform all previous works on every dataset in terms of accuracy.

Artificial Neural networks (ANNs) are widely used to solve classification problems for many machine learning applications. When errors occur in the computational units of an ANN implementation due to for example radiation effects, the result of an arithmetic operation can be changed, and therefore, the predicted classification class may be erroneously affected. This is not acceptable when ANNs are used in many safety-critical applications, because the incorrect classification may result in a system failure. Existing error-tolerant techniques usually rely on physically replicating parts of the ANN implementation or incurring in a significant computation overhead. Therefore, efficient protection schemes are needed for ANNs that are run on a processor and used in resource-limited platforms. A technique referred to as Selective Neuron Re-Computation (SNRC), is proposed in this paper. As per the ANN structure and algorithmic properties, SNRC can identify the cases in which the errors have no impact on the outcome; therefore, errors only need to be handled by re-computation when the classification result is detected as unreliable. Compared with existing temporal redundancy-based protection schemes, SNRC saves more than 60 percent of the re-computation (more than 90 percent in many cases) overhead to achieve complete error protection as assessed over a wide range of datasets. Different activation functions are also evaluated. This research was supported by the National Science Foundation Grants CCF-1953961 and 1812467, by the ACHILLES project PID2019-104207RB-I00 and the Go2Edge network RED2018-102585-T funded by the Spanish Ministry of Science and Innovation and by the Madrid Community research project TAPIR-CM P2018/TCS-4496. Publicado

The gastrointestinal polyp (GIP) is the abnormal growth of tissues in digestive organs. Identifying these polyps from endoscopy video or image is a tremendous task to reduce the future risk of gastrointestinal cancer. This paper proposes a proper diagnosis method of polyp using a fusion of contourlet transform and fine-tuned VGG19 pre-trained model from enhanced endoscopic 224 × 224 patch images. This study has used different fine-tuned models (Alexnet, ResNet50, VGG16, VGG19) as well as a few scratch models while fine-tuned VGG19 works better. Also, this research has used Principal Component Analysis (PCA) and Minimum Redundancy Maximum Relevance (MRMR) dimensionality reduction methods to collect the intuitive features for classification. In Support Vector Machine (SVM) based polyp detection, the prior method (PCA) performs better. Besides, a proposed algorithm marks polyp region from identified polyp patches and uses a binning strategy to process video. A set of experiments are performed on standard public data sets and found comparative improved performance with an accuracy of 99.59%, sensitivity of 99.74% and specificity of 99.44%. This work can be instrumental for the radiologist for diagnosis of polyps during real-time endoscopy.

The problem of detecting and identifying sensor faults is critical for efficient, safe, regulatory-compliant, and sustainable operations of modern industrial processing systems. The increasing complexity of such systems brings, however, new challenges for sensor fault detection and sensor fault isolation (SFD-SFI). One of the key enablers for any SFD-SFI method is analytical redundancy, which is provided by an analytical model of sensor observations derived from first principles or identified from historical data. As defective sensors generate measurements that are inconsistent with their expected behavior as defined by the model, SFD amounts to the generation and monitoring of residuals between sensor observations and model predictions. In this article, we introduce a disentangled recurrent neural network (RNN) with the objective to cope with the smearing-out effect, i.e., where the propagation of a sensor fault to nonfaulty sensor results in large and misleading residuals. The introduction of a probabilistic model for the residual generation allows us to develop a novel procedure for the identification of the faulty sensors. The computational complexity of the proposed algorithm is linear in the number of sensors as opposed to the combinatorial nature of the SFI problem. Finally, we empirically verify the performance of the proposed SFD-SFI architecture using a real data set collected at a petrochemical plant.

Quantitative ultrasound (QUS), which is commonly used to extract quantitative features from the ultrasound radiofrequency (RF) data or the RF envelope signals for tissue characterization, is becoming a promising technique for noninvasive assessments of liver fibrosis. However, the number of feature variables examined and finally used in the existing QUS methods is typically small, to some extent limiting the diagnostic performance. Therefore, this paper devises a new multiparametric QUS (MP-QUS) method which enables the extraction of a large number of feature variables from US RF signals and allows for the use of feature-engineering and machinelearning based algorithms for liver fibrosis assessment. In the MP-QUS, eighty-four feature variables were extracted from multiple QUS parametric maps derived from the RF signals and the envelope data. Afterwards, feature reduction and selection were performed in turn to remove the feature redundancy and identify the best combination of features in the reduced feature set. Finally, a variety of machine-learning algorithms were tested for classifying liver fibrosis with the selected features, based on the results of which the optimal classifier was established and used for final classification. The performance of the proposed MPQUS method for staging liver fibrosis was evaluated on an animal model, with histologic examination as the reference standard. The mean accuracy, sensitivity, specificity and area under the receiver-operating-characteristic curve achieved by MP-QUS are respectively 83.38%, 86.04%, 80.82% and 0.891 for recognizing significant liver fibrosis, and 85.50%, 88.92%, 85.24% and 0.924 for diagnosing liver cirrhosis. The proposed MP-QUS method paves a way for its future extension to assess liver fibrosis in human subjects.

Single image dehazing, which aims at restoring a haze-free image from its correspondingly unconstrained hazy scene, is highly challenging and has gained immense popularity in recent years. However, the images generated using existing haze-removal methods often contain haze, artifacts, and color distortions, which severely degrade the visual quality of the final images. To this end, we propose a network combining multi-scale hierarchical feature fusion and mixed convolution attention to progressively and adaptively enhance the dehazing performance. The haze levels and image structure information are accurately estimated by fusing multi-scale hierarchical features, thus the model restores images with less remaining haze. The proposed attention mechanism is capable of reducing feature redundancy, learning compact internal representations, highlighting task-relevant features and further helping the model to estimate images with sharper textural details and more vivid colors. Therefore, with the application of multi-scale features extracted from both diverse layers and filters, the dehazing performance is significantly improved. Furthermore, a deep semantic loss function is proposed to highlight more semantic information in deep features. The experimental results show that the proposed method outperforms state-of-the-art haze removal algorithms.

For the existing repetitive motion generation (RMG) schemes for kinematic control of redundant manipulators, the position error always exists and fluctuates. This article gives an answer to this phenomenon and presents the theoretical analyses to reveal that the existing RMG schemes exist a theoretical position error related to the joint angle error. To remedy this weakness of existing solutions, an orthogonal projection RMG (OPRMG) scheme is proposed in this article by introducing an orthogonal projection method with the position error eliminated theoretically, which decouples the joint space error and Cartesian space error with joint constraints considered. The corresponding new recurrent neural networks (NRNNs) are structured by exploiting the gradient descent method with the assistance of velocity compensation with theoretical analyses provided to embody the stability and feasibility. In addition, simulation results on a fixed-based redundant manipulator, a mobile manipulator, and a multirobot system synthesized by the existing RMG schemes and the proposed one are presented to verify the superiority and precise performance of the OPRMG scheme for kinematic control of redundant manipulators. Moreover, via adjusting the coefficient, simulations on the position error and joint drift of the redundant manipulator are conducted for comparison to prove the high performance of the OPRMG scheme. To bring out the crucial point, different controllers for the redundancy resolution of redundant manipulators are compared to highlight the superiority and advantage of the proposed NRNN. This work greatly improves the existing RMG solutions in theoretically eliminating the position error and joint drift, which is of significant contributions to increasing the accuracy and efficiency of high-precision instruments in manufacturing production.

Inter prediction is a crucial part of hybrid video coding frameworks, utilized to exploit the temporal redundancy in video sequences and improve the coding performance. During inter prediction, a predicted block is typically derived from reference pictures using motion estimation and motion compensation. To improve the coding performance of inter prediction, a neural network based enhancement to inter prediction (NNIP) is proposed in this paper. NNIP is composed of three networks, namely residue estimation network, combination network, and deep refinement network. Specifically, first, a residue estimation network is designed to estimate the residue between current block and its predicted block using their available spatial neighbors. Second, the feature maps of the estimated residue and the predicted block are extracted and concatenated in a combination network. Finally, the concatenated feature maps are fed into a deep refinement network to generate a refined residue, which is added back to the predicted block to derive a more accurate predicted block. NNIP is integrated in HEVC to evaluate its efficiency. The experimental results demonstrate that NNIP can achieve 4.6%, 3.0%, and 2.7% BD-rate reduction on average under LDP, LDB, and RA configurations compared to HEVC.

The use of deep learning methods in hyperspectral image (HSI) classification has been a promising approach due to its powerful ability to automatically extract features in recent years. This article proposes a novel deep framework for HSI classification problems, referred to as feature-grouped network based on spectral-spatial connected attention mechanism (FG-SSCA). Different from the existing deep learning methods, the proposed framework integrates the spectral attention module and spatial attention module continuously from the raw HSI input, which is embedded into convolutional neural networks and could enhance the distinguishing ability of spectral bands and learn the spatial relevance between the neighboring pixels together. Meanwhile, the generating feature maps are sliced into a series of small groups in sequence along the direction of spectral bands and each group sequentially extracts spatial-spectral features through multiple spectral and spatial residual blocks. This feature-grouped strategy could fully utilize the redundancy and difference of bands and obtain more available and valuable information. The proposed FG-SSCA method could greatly improve generalization performance and make tremendous successes in HSI classification. Experimental results on several HSI benchmark data sets verify the effectiveness and superiority of the proposed method in comparison with the state-of-the-art approaches for HSI classification.

A synthetic air data system (SADS) is an analytical redundancy technique that is crucial for unmanned aerial vehicles (UAVs) and is used as a backup system during air data sensor failures. Unfortunately, the existing state-of-the-art approaches for SADS require GPS signals or high-fidelity dynamic UAV models. To address this problem, a novel synthetic airspeed estimation method that leverages deep learning and an unscented Kalman filter (UKF) for analytical redundancy is proposed. Our novel fusion-based method only requires an inertial measurement unit (IMU), elevator control input, and airflow angles while GPS, lift/drag coefficients, and complex aircraft dynamic models are not required. Additionally, we demonstrate that our proposed temporal convolutional network (TCN) is a more efficient model for airspeed estimation than the renowned models, such as ResNet or bidirectional long short-term memory (LSTM). Our deep learning-aided UKF was experimentally verified on long-duration real flight data and has promising performance compared with the state-of-the-art methods. In particular, it is confirmed that our proposed method robustly estimates the airspeed even under dynamic flight conditions where the performance of conventional methods is degraded.

Neuromorphic vision sensors (NVS) can enable energy savings due to their event-driven that exploits the temporal redundancy in video streams from a stationary camera. However, noise-driven events lead to the false triggering of the object recognition processor. Image denoise operations require memoryintensive processing leading to a bottleneck in energy and latency. In this paper, we present in-memory filtering (IMF), a 6TSRAM in-memory computing based image denoising for eventbased binary image (EBBI) frame from an NVS. We propose a non-overlap median filter (NOMF) for image denoising. An inmemory computing framework enables hardware implementation of NOMF leveraging the inherent read disturb phenomenon of 6T-SRAM. To demonstrate the energy-saving and effectiveness of the algorithm, we fabricated the proposed architecture in a 65nm CMOS process. As compared to fully digital implementation, IMF enables > 70x energy savings and a > 3x improvement of processing time when tested with the video recordings from a DAVIS sensor and achieves a peak throughput of 134.4 GOPS. Furthermore, the peak energy efficiencies of the NOMF is 51.3 TOPS/W, comparable with state of the art inmemory processors. We also show that the accuracy of the images obtained by NOMF provide comparable accuracy in tracking and classification applications when compared with images obtained by conventional median filtering., 13 pages

Bubble size distribution (BSD) is a crucial indicator of operating conditions in froth flotation. However, current BSD-based recognition methods have some limitations in the BSD estimation and don't achieve satisfactory results. Therefore, an alternative operating condition recognition method is proposed. Firstly, the temporal cumulative distribution function (TCDF) is designed to represent the BSD on the froth video. Then, the piecewise basis function is established to fit the TCDF. Secondly, the TCDF-Minimum Redundancy Maximum Relevancy (MRMR) feature is generated by MRMR to increase the sensitivity of TCDF features to the operating conditions. Lastly, the proposed TCDF-MRMR feature together with froth velocity and texture features are fed into an AdaBoost-extreme learning machine (ELM) model to recognize the operating conditions. Experiments in a zinc flotation process demonstrate that the TCDF-MRMR feature is more effective than other BSD-based features, and the AdaBoost-ELM shows better performance than traditional single recognition models.

Many pioneering approaches have verified the effectiveness of utilizing the global temporal and local object information for video understanding tasks and have achieved significant progress. However, existing methods utilize object detectors to extract all objects overall video frames. This may bring performance degradation due to the information redundancy both spatially and temporally. To address this problem, we propose an adaptive spatial location module for the video captioning task which dynamically predicts an important position of each video frame in the procedure of generating the description sentence. The proposed adaptive spatial location method not only makes our model focus on local object information, but also reduces time and memory consumption brought by the temporal redundancy in extensive video frames and improves the accuracy of generated description. Besides, we propose a balanced loss function to address the class imbalance problem existing in training data. The proposed balanced loss assigns different weight to each word of ground-truth sentence in the training process which can generate more diversified description sentences. Extensive experimental results on the MSVD and MSR-VTT dataset show that the proposed method achieves competitive performance compared to state-of-the-art methods.

Hyperspectral image (HSI) registers hundreds of spectral bands, whose intraclass variability and interclass similarity are resourceful information to be mined. Intraclass variability reflects the nonuniform and redundancy of the spatial and semantic features extracted from HSI. Interclass similarity represents the inherent relationship between adjacent features and snapshots. Existing models extract the superficial correlation representation for HSI to tackle the classification task but fail to embed the interclass and intraclass correlations due to these models' intrinsic bottlenecks. Confronting the challenges of capturing interrelation for complex data in practice, we propose a Context-Aware Attentional Graph U-Net (CAGU) to improve these two modes of representation, which is more flexible in feature enhancement. In this method, attentional Graph U-Net is capable of extracting the intraclass embeddings within a non-Euclidean space by combining similar distributing feature vertices. The gated recurrent unit (GRU) is another critical component of our model to capture the context-aware dynamic interclass embeddings. Extensive experiments demonstrate that our model can efficiently outperform state-of-the-art methods across-the-board on five wide-adopted public data sets, namely, Pavia University, Indian Pines, Salinas Scene-show, Houston 2013, and Houston 2018, on par with the same scale of model parameters.

We address the problem of storing and analyzing large datasets of passenger trips over public transportation networks that are of interest to network administrators trying to balance transportation offers (e.g., frequency of vehicles) according to the historical demand. We exploit the fact that all passenger trips made within the same vehicle share the same trajectories to reduce their redundancy and provide a representation, based on well-known compact data structures, that not only reduces the space requirements of the original passenger’s trajectories but also efficiently supports querying. Our solution uses two complementary representations: T-Matrices which excels at querying the aggregated network load, and TTCTR which represents all passenger trips and aims at counting the trips following a given pattern (i.e., how many passengers started/ended a trip at a given location or moved from a given location to another). In addition, we propose XCTR, a variant of TTCTR, which efficiently answers a wider range of queries at the cost of a moderate performance loss for some queries and some space overhead. Overall, our representation can handle a dataset of ten million trips within approximately 65% of its original size while supporting a wide range of queries in the order of microseconds.

The key to efficient person search is jointly localizing pedestrians and learning discriminative representation for person re-identification (re-ID). Some recently developed models are built with separate detection and re-ID branches on top of shared region feature extraction networks. There are two factors that are detrimental to re-ID feature learning. One is the background information redundancy resulting from the large receptive field of neurons. The other is the body part missing and background clutter caused by inaccurate localization. In this work, a bottom-up fusion (BUF) subnet is proposed to fuse the bounding box features pooled from multiple network stages. With a few parameters introduced, BUF leverages the multi-level features with various sizes of receptive fields to mitigate the background-bias problem. To further suppress the non-pedestrian regions, the newly introduced segmentation head generates a foreground probability map as guidance for the network to focus on the foreground regions. The resulting foreground attention module (FAM) enhances the foreground features. Moreover, for robust feature learning in practical person search, we propose to adaptively smooth the labels of the pedestrian boxes with consideration of the detection quality. Extensive experiments on PRW and CUHK-SYSU validate the effectiveness of the proposals. Our Bottom-Up Foreground-Aware Feature Fusion (BUFF) network with ALS achieves considerable gains over the state-of-the-art on PRW and competitive performance on CUHK-SYSU.

Telemedicine and various tele-medicinal applications are revolutionizing a variety of healthcare departments through its innovative means of remote diagnosing and faster first aid administration. Digital images play an important role in these applications for ensuring better and faster health carfigure. These digital images, which usually contain confidential and diagnostic information about the patients, are usually transmitted through public networks among hospitals, doctors and patients. Consequently, there is a need to secure them while being stored and in transit to guarantee the privacy of the patient. However, unique properties of digital image data, such as high redundancy and correlation between the pixels of the image and their large size, make the conventional cryptographical algorithms insufficient for ensuring proper security while encrypting them. As the conventional algorithms seize to be a reliable solution, the need to develop improved and dedicated image encryption algorithms also arises. Chaotic systems are systems that appear random and unpredictable from the outside but are governed by deterministic equations or rules on the inside. Due to these properties, along with ergodicity and its heightened sensitivity to the initial conditions, chaotic systems are devised to be one of the best candidates for securing the storage and transmission of digital images. However, the security of a chaotic image encryption system depends upon the chaotic behavior demonstrated by the chaotic map applied in the image encryption system. Because of this, different attacks can be used to break a chaotic encryption system if the scheme is not well-structured. This paper introduces a chaotic image encryption scheme that incorporates two chaotic maps, namely, Arnold's Cat Map and 2D Logistic-Sine-Coupling Map(2D-LSCM), for increased randomness and security of the encrypted image. We also analyze the performance and security of our scheme and compare it with other prominent chaotic image encryption schemes.

Although deep learning is gaining more widespread use in hyperspectral image analysis, it is challenging to train high-capacity models in a supervised way--ground-truth sets are expensive to obtain, and they are practically always extremely imbalanced. To deal with the problem of missing ground-truth data, its high dimensionality and potential redundancy, we introduce a novel unsupervised feature learning technique to extract discriminative features from the original data. It exploits recurrent neural network-based asymmetric autoencoders (AEs) to learn the compressed representation of unlabeled data, and can elaborate both spectral and spectral-spatial features. Our extractors can be incorporated into the unsupervised segmentation pipeline--they can be followed by any clustering algorithm. The experiments revealed that our approaches deliver high-quality segmentation without any prior class labels, and are one order of magnitude faster than 3-D convolutional AEs. Our algorithms outperform or work on par with other approaches while allowing for significant data reduction.

Multi-label learning algorithms have significant challenges due to high-dimensional feature space and noises in multi-label datasets. Feature selection methods are effective techniques to deal with these problems. ParetoCluster is an effective multi-label feature selection algorithm based on Pareto dominance and cluster analysis concepts which considers each label an objective function. This algorithm loses its effectiveness to differentiate features when dealing with high labeled datasets and makes most features incomparable (e.g., when most features fall into the first layer). Thus, a cluster analysis criterion in ParetoCluster will play a decisive role in determining the most relevant features. Bearing this in mind, in this paper, we have modeled the multi-label feature selection problem into a bi-objective optimization problem regarding the relevancy and redundancy degree of the features. We then handle it using Pareto dominance in this bi-objective domain. To illustrate the optimality and efficiency of the proposed method, we have compared our approach against some similar techniques.

Neural network architecture search automatically configures a set of network architectures according to the targeted rules. Thus, it relieves the human-dependent effort and repetitive resources consumption for designing neural network architectures and makes the task of finding the optimum network architecture with better performance much more accessible. Network architecture search methods based on differentiable architecture search (DARTS), however, introduces parameter redundancy. To address this issue, this work presents a novel method for optimizing network architectures that combines DARTS with generative adversarial learning (GAL). We first find the module structures utilizing the DARTS algorithm. Afterwards, the retrieved modules are stacked to derive the initial neural network architecture. Next, the GAL is used to prune some branches of the initial neural network, thereby obtaining the final neural network architecture. The proposed DARTS-GAL method re-optimizes the network architecture searched by DARTS to simplify the network connection and reduce network parameters without compromising network performance. Experimental results on benchmark datasets, i.e., Mixed National Institute of Standards and Technology (MNIST), FashionMNIST, Canadian Institute for Advanced Research10 (CIFAR10), Canadian Institute for Advanced Research100 (CIAFR100), Cats vs Dogs, and voiceprint recognition datasets, indicate that the test accuracies of the DARTS-GAL are higher than those of the DARTS in the majority of the cases. In particular, the proposed solution exhibits an improvement in accuracy by 7.35% on CIFAR10 compared with DARTS, attaining the state-of-the-art result of 99.60%. Additionally, the number of network parameters derived by the DARTS-GAL is significantly lower than that by the DARTS method, with a pruning rate of 62.3% at the highest case.

It has long been recognized that synthetic aperture radar (SAR) images suffer in many applications from the speckle noise. Video SAR has a high frame rate of imaging and contains redundant information among frames. The temporal redundancy in video SAR images has been found useful for suppressing the speckle noise. However, the motion and the local differences between frames make it difficult to employ the temporal redundancy to suppress speckle noise. This article presents a video SAR image despeckling framework based on a new unsupervised training strategy referred to as DualNoise2Noise. This developed framework consists of a registration network and a denoising network. The registration network first compensates for the motion between two adjacent frames in video SAR in real time. After image registration, two adjacent frames with random speckle noise can be considered as the observations of the same region with local differences. The denoising network adopts the DualNoise2Noise training strategy to suppress speckle noise by using the temporal redundancy and to remove the negative impact caused by the local differences. The proposed approach has been used to process the real video SAR data, and the experimental results are convincing.

Redundancy manipulators need favorable redundancy resolution to obtain suitable control actions to guarantee accurate kinematic control. Among numerous kinematic control applications, some specific tasks such as minimally invasive manipulation/surgery require the distal link of a manipulator to translate along such fixed point. Such a point is known as remote center of motion (RCM) to constrain motion planning and kinematic control of manipulators. Recurrent neural network (RNN) which possesses parallel processing ability, is a powerful alternative and has achieved success in conventional redundancy resolution and kinematic control with physical constraints of joint limits. However, up to now, there still is few related works on the RNNs for redundancy resolution and kinematic control of manipulators with RCM constraints considered yet. In this paper, for the first time, an RNN-based approach with a simplified neural network architecture is proposed to solve the redundancy resolution issue with RCM constraints, with a new and general dynamic optimization formulation containing the RCM constraints investigated. Theoretical results analyze and convergence properties of the proposed simplified RNN for redundancy resolution of manipulators with RCM constraints. Simulation results further demonstrate the efficiency of the proposed method in end-effector path tracking control under RCM constraints based on a redundant manipulator.

Mining knowledge from imbalanced data is challenging due to the uneven distribution of classes and increasing dimensionality of data accumulated from real-life applications. Selecting informative features from imbalanced data is especially important for building an effective learning method. The global redundancy and the effect of imbalanced distribution need to be considered simultaneously. In this study, a feature selection method that considers the imbalanced distribution of classes in data is investigated by embedding the weighted constraint on the majority class into the global redundancy minimization GRM framework. Global redundancy minimization is acquired through an objective function that contains a feature redundancy matrix and feature scores. A new form of regularization to a within-class scatter matrix is first presented, which emphasizes the minority class and replaces the redundancy measurement approach. Then, after employing this new form of a within-class scatter matrix in GRM and taking the between-class distance as the GRM input score, a GRM-based discriminant feature selection algorithm (GRM-DFS) is proposed. Comparison studies on a within-class scatter matrix with different forms of regularization indicate that the proposed form of a within-class scatter matrix is effective when dealing with imbalanced data. Experiments on public imbalanced datasets are performed. The experimental results indicate that GRM-DFS is effective.

Convolutional Neural Networks (CNNs) have achieved great success due to the powerful feature learning ability of convolution layers. Specifically, the standard convolution traverses the input images/features using a sliding window scheme to extract features. However, not all the windows contribute equally to the prediction results of CNNs. In practice, the convolutional operation on some of the windows (e.g., smooth windows that contain very similar pixels) can be very redundant and may introduce noises into the computation. Such redundancy may not only deteriorate the performance but also incur the unnecessary computational cost. Thus, it is important to reduce the computational redundancy of convolution to improve the performance. To this end, we propose a Content-aware Convolution (CAC) that automatically detects the smooth windows and applies a 1 ×1 convolutional kernel to replace the original large kernel. In this sense, we are able to effectively avoid the redundant computation on similar pixels. By replacing the standard convolution in CNNs with our CAC, the resultant models yield significantly better performance and lower computational cost than the baseline models with the standard convolution. More critically, we are able to dynamically allocate suitable computation resources according to the data smoothness of different images, making it possible for content-aware computation. Extensive experiments on various computer vision tasks demonstrate the superiority of our method over existing methods.

As deep CNNs get larger, it becomes more challenging to deploy them on resource-restricted mobile devices. Filter-level pruning is one of the most popular methods to compress deep models for mobile deployment. It prunes unimportant filters in the pre-trained CNN to reduce its storage and computational cost, yielding a smaller and more efficient model. Though having made some progress, most filter-level pruning methods still suffer from at least one of the following two problems: 1) High redundancy: some methods pick out unimportant filters without considering their correlations, and thus many highly correlated filters are not pruned, yielding a model still with high redundancy. 2) Sub-optimal: all existing pruning methods achieve sub-optimal pruning plan because they neglect that a high parameters reduction ratio (PRR) does not always mean a high FLOPs reduction ratio (FRR). In this paper, we propose our customized correlation-based pruning (COP) to solve these problems. In particular, we observe redundant filters through their correlations. Moreover, to achieve the optimal pruning plan, PRR and FRR are considered when evaluating filters’ importance. Besides, we also propose a new pruning pipeline, which improves the accuracy of the pruned model. Extensive experiments show that our proposed method has outperformed the state-of-the-art on several popular architectures and datasets.

Band selection (BS) can mitigate the “curse of dimensionality” problem and improve the performance of hyperspectral image (HSI) classification. Genetic algorithms (GAs) have been applied to the task of hyperspectral BS showing significant advantages compared with other literature methods. However, the traditional GAs-based methods often select sets of bands having residual redundancy due to the large search space related to hyperspectral BS and the limitation of premature convergence in GAs. Moreover, existing GAs-based methods often are supervised, and that needs a large number of labeled samples to compute the fitness value for assessing the quality of selected bands. In this article, an unsupervised BS approach based on an improved GA is proposed. A fitness function based on the fisher score combined with superpixel is designed for evaluating the discriminability of band subsets considering both spectral and spatial information. Then, modified genetic operations are constructed to restrain the search space and reduce the redundancy of selected bands. The performance of the proposed spectral-spatial GA-based BS method is evaluated on three HSIs. The experimental results demonstrate that the proposed method is superior to the traditional GA-based method and seven state-of-the-art unsupervised methods.

An elastic neural network is developed and evolved towards effectively re-configurable hardware in fully parallel on chip. The original prototype of DiaNet is organized as a symmetrical bisection neural network, which is feasible to be partitioned into arbitrary pieces of neural networks (NNs) without redundancy. To prevent the depth explosion in implementing complex tasks (complicated pattern recognition for instance), the evolution of DiaNets is investigated in this work. By using the I/O layer integration technology which enables all neurons in the hidden layer of DiaNet to receive inputs, the number of layers is reduced to 8.8 % of DiaNet prototype. In this manner, the DiaNet topology is feasible to implement complex NNs without the risk of depth explosion. Moreover, the skip connection technology is proposed to avoid the gradient vanishing due to deep learning, which is significant to DiaNets especially. Compared with the LeNet5 model as state-of-the-art, the evolved DiaNet topology achieves the parameter reduction of 90.86 % for MNIST recognition with the negligible loss of accuracy. To reduce hardware utilization, the sensitivity to the decline of computational precision and bit-width is investigated to suggest the guideline for efficient hardware implementations. Finally, the effectiveness of DiaNet is verified by the proposed re-configurable architecture on FPGA with the power reduction of 10.8 % compared to state-of-the-art implementations.

This study considers the problem of text-based person search, which aims to find the corresponding person of a given text description in an image gallery. Existing methods usually learn a similarity mapping of local parts between image and text, or embed the whole image and text into a unified embedding space. However, the relevance between local and the whole is largely underexplored. In this paper, we design a Divide-and-Merge Embedding (DME) learning framework for text-based person search. DME explicitly 1) models the relations between local parts and global embedding, 2) incorporates local details into global embedding. Specifically, we design a Feature Dividing Network (FDN) to embed the input into K locally guided semantic representations by self-attentive embedding, each representation depicts a local part of the person. Then, we propose a Relevance based Subspace Projection (RSP) method for merging diverse local representations to a compact global embedding. RSP helps the model to obtain discriminative embedding by jointly minimizing the redundancy of local parts and maximizing the relevance between local parts and global embedding. Extensive experimental results on three challenging benchmarks, i.e., CUHK-PEDES, CUB and Flowers datasets, have demonstrated the effectiveness of the proposed method.

With the rapid increase in GPS data collection through pervasive use of mobile devices, it has become an important problem to discover significant places of moving objects from complex spatial and temporal trajectories. This problem is challenging mainly because such trajectory data suffer from several issues including incompleteness, low quality, high redundancy, and oftentimes trajectory points do not follow Gaussian distribution . We propose a clustering-based method with temporal and spatial semantics, referred to as Stops and Moves of Trajectories using Attribute Selection (SMoTAS), whose technical advantages are multifold. Firstly, it improves data availability by using a self-adaptive algorithm to correct the deviation in traditional speed-based methods. Secondly, it improves place mining accuracy by filtering multi-label clustering results when there is a lack of detailed geographic data. Thirdly, it employs feature selection to exploit the core attributes of clustering and simplify the clustering results with Grubbs criterion. Experimental results on real-life datasets show that SMoTAS not only achieves substantial improvement of accuracy over existing methods in discovering significant places, but also exhibits superior adaptability to different trajectories and application scenarios.

As the electromagnetic environment in battlefields is more and more complex, automatic modulation recognition for radar signals is becoming vital and challenging. Traditional methods are more likely to cause lower recognition accuracy with higher computational complexity in low signal-to-noise ratio (SNR). Feature redundancy especially for handcrafted features is one of the shortcomings of deep-learning-based methods. In this article, a novel end-to-end sequence-based network that consists of a shallow convolutional neural network, a bidirectional long short-term memory (Bi-LSTM) network strengthening with a self-attention mechanism, and a dense neural network is constructed to recognize eight kinds of intrapulse modulations of radar signals. The autocorrelation functions of received radar signals are first calculated as autocorrelation features. Then, these features are employed as inputs of the proposed network which owns significant sequence processing advantages and adaptive selection ability of features. Finally, the proposed network outputs prediction modulations directly. The simulation results verify the robustness and effectiveness of autocorrelation features. And the proposed network achieves about 61.25% accuracy at −20 dB and more than 95% accuracy at −10 dB. Compared with four state-of-the-art networks, the proposed network has better recognition performance especially at low SNRs with much lower computational complexity. Results on measured signals also demonstrate that the proposed network outperforms these four networks.

Active learning aims to select the most valuable unlabeled instances for annotations and thus to maximize the performance of a learner with the selected instances. Batch-mode active learning methods can select a batch of unlabeled instances with informativeness and low redundancy at each iteration, they are more efficient than myopic active learning methods, which typically have to retrain the model after querying each instance. However, Batch-mode Multi-label Active Learning is not well explored yet, since it is quite difficult to select a batch of instances or instance-label pairs with high informativeness but low redundancy for query and budget saving. In this paper, we propose a novel approach CBMAL (Cost-effective Batch-mode Multi-label Active Learning) to address this challenge. CBMAL firstly selects a batch of informative instance-label pairs using uncertainty, label correlation and label space sparsity. Next, CBMAL leverages the information from the feature and label dimensions to pick out a small batch of instance-label pairs with the highest information and lowest redundancy from the first batch. These instance-label pairs are then queried and used to update the learner for the next iteration. Experimental results on six benchmark datasets demonstrate that CBMAL can reduce both the query and time costs while achieve a better performance than state-of-the-art methods.

Palm Vein Identification (PVI) is a modern biometric security technique used for enhancing security and\ud authentication systems. The key characteristics of palm vein patterns include its uniqueness to each individual, its\ud unforgettability, non-intrusiveness and its ability for disallowing unauthorized persons. However, the extracted\ud features from the palm vein patterns are huge with high redundancy. In this paper, we propose a combined model of\ud two-Dimensional Discrete Wavelet Transform, Principal Component Analysis (PCA), and Particle Swarm\ud Optimization (PSO) (2D-DWTPP) that feeds wrapper model with an optimal subset of features to enhance the\ud prediction accuracy of -palm vein patterns. The 2D-DWT extract features from palm vein images, using the PCA to\ud reduce the redundancy in palm vein features. The system has been trained to select high recognition features based on\ud the wrapper model. The proposed system uses four classifiers as an objective function to determine PVI which include\ud Support Vector Machine (SVM), K Nearest Neighbor (KNN), Decision Tree (DT) and Naïve Bayes (NB). The\ud empirical results proved that the proposed model has the best results with SVM. Moreover, our proposed 2D-DWTPP\ud model has been evaluated and the results show remarkable efficiency in comparison with AlexNet and other classifiers\ud without feature selection. Experimentally, the proposed model has better accuracy as reflected by 98.65% whereas\ud AlexNet has 63.5% accuracy and the classifier without feature selection process has 78.79% accuracy.