Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks306 from public data
- Non-Salient Region Object Mining for Weakly Supervised Semantic Segmentation223
A graph-based global reasoning unit is introduced to strengthen the classification network’s ability to capture global relations among disjoint and distant regions and a potential object mining module is proposed to reduce the false-negative rate in pseudo labels.
- Minimum cost attribute reduction in decision-theoretic rough set models210
A new definition of attribute reduct for decision-theoretic rough set models is provided and a heuristic approach, a genetic approach and a simulated annealing approach to the new problem are proposed.
- Jo-SRC: A Contrastive Approach for Combating Noisy Labels206
A noise-robust approach named Jo-SRC (Joint Sample Selection and Model Regularization based on Consistency) that trains the network in a contrastive learning manner to advance the model generalization performance by introducing consistency regularization.
- 194
- On an optimization representation of decision-theoretic rough set model147
An optimization representation of decision-theoretic rough set model is proposed by considering the minimization of the decision cost and two significant inferences can be drawn from the solution of the optimization problem.
- A theorem on the uncorrelated optimal discriminant vectors109
It is proved that the classical optimal discriminant vectors are equivalent to UODV, which can be used to extract (L−1) uncorrelated discriminant features for L-class problems without losing any discriminant information in the meaning of Fisher discriminant criterion function.
- Hierarchical initialization approach for K-Means clustering98
The experimental results show that the proposed hierarchical initialization approach to the K-Means clustering problem needs less iteration time compared with existing approaches and has better performance in terms of convergence speed and ability to reduce the impact of noises.
- Feature Affinity-Based Pseudo Labeling for Semi-Supervised Person Re-Identification89
This paper proposes a novel feature affinity-based pseudo labeling method with two possible label encodings, and is the first study that employs pseudo-labeling by measuring the affinity of unlabeled samples with the underlying clusters of labeled data samples using the intermediate feature representations from deep networks.
- Improving protein-ATP binding residues prediction by boosting SVMs with random under-sampling81
A modified AdaBoost ensemble scheme based on random under-sampling is developed and a method for objectively reporting evaluation results under the imbalanced learning scenario is discussed, showing that the proposed method achieves higher prediction accuracy.
- WLBP: Weber local binary pattern for local image description77
A local descriptor, called Weber Local Binary Pattern (WLBP 1), which effectively combines the advantages of WLD and LBP, which is evaluated on the face recognition problem under different challenges and demonstrates that WLBP is a powerful texture descriptor.
- TargetATPsite: A template‐free method for ATP‐binding sites prediction with residue evolution image sparse representation and classifier ensemble76
A new sequence‐based template‐free predictor (TargetATPsite) to identify the Adenosine‐5′‐triphosphate (ATP) binding sites with machine‐learning approaches is reported, which is effective for dealing with imbalance phenomenon between the positive and negative training samples.
- Web-Supervised Network with Softly Update-Drop Training for Fine-Grained Visual Classification66
A novel approach to remove irrelevant samples from the real-world web images during training, and only utilize useful images for updating the networks is proposed, which can alleviate the harmful effects caused by irrelevant noisy web images to achieve better performance.
- 62
- Semantically Meaningful Class Prototype Learning for One-Shot Image Segmentation61
This paper proposes to leverage the multi-class label information during the episodic training to encourage the network to generate more semantically meaningful features for each category and proposes a fused prototype guidance branch for the segmentation of the query image.
- 59
- Robust gait recognition using hybrid descriptors based on Skeleton Gait Energy Image55
A new model-based representation, Skeleton Gait Energy Image (SGEI), has been proposed in this paper and the experimental performances indicate that the proposed methods are more robust to the cloth changes, and contribute to increasing the robustness of gait recognition in the unconstrained environments with view variances and clothes changes.
- 55
- Modified particle filter-based infrared pedestrian tracking54
A modified particle filter algorithm that combines intensity and edge cues is proposed that can track the infrared pedestrian more effectively and reliably than the traditional particle filter algorithms.
- PLDA modeling in i-vector and supervector space for speaker verification51
The probabilistic linear discriminant analysis (PLDA) is employed to handle speaker and session variability for speaker verification task and the Gaussianized rank-norm is introduced, as an alternative to whitening, for feature normalization prior to PLDA modeling.
- Feature mask network for person re-identification50
A Feature Mask Network (FMN) is proposed that takes advantage of ResNet high-level features to predict a feature map mask and then imposes it on the low- level features to dynamically re-weight different object parts for a complementary feature representation.
- 48
- 47
- 47
- 46
- Dispersion based Clustering for Unsupervised Person Re-identification.44
A Dispersion based Clustering (DBC) approach which performs better at discovering the underlying data patterns and can automatically prioritize standalone data points and prevents poor clustering.
- A Domain Robust Approach For Image Dataset Construction44
This work proposes a novel image dataset construction framework which can generalize well to unseen target domains and formulate image filtering as a multi-instance learning (MIL) problem with constrained positive bags.
- 42
- 40
- 40
- 39
- 37
- 36
- Feature extraction using local structure preserving discriminant analysis32
An efficient feature extraction method, named local structure preserving discriminant analysis (LSPDA), is presented, which constructs the local scatter and the between-class scatter to characterize the sub- and multi-manifold information respectively.
- 32
- 31
- QMOR: QoS-Aware Multi-sink Opportunistic Routing for Wireless Multimedia Sensor Networks30
This paper proposes a QoS-aware multi-sink opportunistic routing (QMOR) to efficiently deliver multimedia information under QoS constraints for WMSNs and demonstrates that the QMOR performs better than typical WMSN QoS routing algorithm, in terms of the video transmission quality and energy utilization efficiency.
- 30
- 29
- 28
- Data-driven Meta-set Based Fine-Grained Visual Recognition26
This work proposes a data-driven meta-set based approach to deal with noisy web images for fine-grained recognition that can alleviate the harmful effects caused by out-of-distribution noise and properly exploit the in-dist distribution noisy samples for training.
- 26
- 25
- 25
- Discovering and Distinguishing Multiple Visual Senses for Polysemous Words25
This work presents a novel framework that solves the problem of polysemy by allowing sense-specific diversity in search results by discovering a list of possible semantic senses and training a visual classifier for each selected semantic sense.
- 25
- A new web-supervised method for image dataset constructions24
A novel automatic image dataset construction framework is proposed by employing multiple query expansions, which indicates the effectiveness of the method is superior to weak supervised and web supervised state-of-the-art methods.
- 23
- 23
- Sparse representation-based ECG signal enhancement and QRS detection22
The proposed algorithm models the ECG signal as the superposition of a few inner structures plus additive random noise, where these structures can be learned from the input signal or a training set and used to detect and locate the QRS complexes in the enhanced ECG signals.
- 22
- Information-theoretic measures of uncertainty for interval-set decision tables21
This paper addresses the uncertainty measurement problem in interval-set decision tables by extending the accuracy and the roughness, and proposes an alternative uncertainty measure based on the conditional information entropy, called the interval-decision entropy.
- Strong barrier coverage of directional sensor networks with mobile sensors20
Through extensive simulations, the results show that the energy-efficient barrier repair algorithm improves the barrier coverage and prolongs the network lifetime by minimizing the maximum sensor moving distance.
- Fusing hierarchical multi-scale local binary patterns and virtual mirror samples to perform face recognition20
A framework that fuses virtual mirror synthesized training samples as bases and hierarchical multi-scale local binary patterns (LBP) features for classification that has superior competitiveness that the accuracy has been improved, while the complexity has been fallen.
- Learning protein multi-view features in complex space20
A novel framework for protein multi-view feature fusion is proposed: first, features from different views are parallely combined to form complex feature vectors; then, the classic principal component analysis is extended to the generalized principle component analysis for further feature extraction from the parallely Combined complex features, which lie in a complex space.
- An Adaptive Opportunistic Network Coding Mechanism in Wireless Multimedia Sensor Networks20
This work proposes an Adaptive Opportunistic Network Coding mechanism (AONC) to improve the transmission quality of video stream in wireless multimedia sensor networks and designs an opportunistic forwarding strategy based on dynamic priority to ensure that packets have a better chance to be coded and transmitted, thus achieving much higher throughput.
- Design and Implementation of an Audio Classification System Based on SVM20
The results of experiments show that the audio classification system based on SVM designed in the paper can classify audio effectively, and the average identification accuracy reaches more than 90%.
- Extracting Useful Knowledge from Noisy Web Images via Data Purification for Fine-Grained Recognition19
This work investigates the noisy label problem and proposes a method that can specifically distinguish in- and out-of-distribution noisy samples and train the model on a less noisy web training set to achieve better robustness and performance.
- Road segmentation with image-LiDAR data fusion in deep neural network19
A novel structure to fuse image and LiDAR point cloud in an end-to-end semantic segmentation network, in which the fusion is performed at decoder stage instead of at, more commonly, encoder stage to improve the road prediction compared with transpose convolution with skip layers.
- Deep representation learning for road detection using Siamese network19
A Siamese deep neural network based on FCN-8s to detect road region using data collected from a monocular color camera and a Velodyne-64 LiDAR sensor to demonstrate the effectiveness of this proposed approach.
- Extracting Privileged Information from Untagged Corpora for Classifier Learning19
A new instance-level multi-instance learning (MIL) model is proposed to enhance classifier learning by extracting PI from untagged corpora, which can effectively eliminate the dependency on manually labeled data.
- Priority-based IEEE 802.15.4 CSMA/CA mechanism for WSNs19
A priority-based IEEE 802.15.4 carrier sense multiple access with collision avoidance (CSMA/CA) mechanism for WSNs by modeling the proposed mechanism using a Markov chain, and compute the successful transmission probability, throughput and energy consumption for nodes with different priority levels.
- 18
- Uncertainty measures for interval set information tables based on interval δ-similarity relation18
Four new combinations ofuncertainty measures by considering proposed granularity measures and δ-accuracy andδ-roughness are defined and analyzed and illustrate that the proposed measures are effective and accurate for interval set information tables.
- 17
- 17
- Exploiting Universum data in AdaBoost using gradient descent17
This paper addresses a novel boosting algorithm called UAdaBoost that can improve the classification performance of AdaBoost with Universum data and describes the practical conditions for the effectiveness of Universum learning.
- Locality constrained representation based classification with spatial pyramid patches17
Based on the locality constrained representation, three algorithms are presented which outperform the state-of-the-art on the AR and Extended Yale B datasets for both the occlusion and single sample per person (SSPP) problems.
- Exploiting textual queries for dynamically visual disambiguation16
A novel framework that resolves visual polysemy by dynamically matching candidate text queries with retrieved images by employing the proposed saliency-guided deep multi-instance learning (MIL) network to remove outliers and learn classification models for visual disambiguation.
- Discriminant similarity and variance preserving projection for feature extraction16
A novel supervised dimensionality reduction algorithm called discriminant similarity and variance preserving projection (DSVPP) is presented for feature extraction and recognition, which redefine the intrinsic graph and penalty graph to model the intra-class compactness and inter-class separability of data points.
- Kernel-based improved discriminant analysis and its application to face recognition16
This paper presents a variant of KDA called kernel-based improved discriminant analysis (KIDA), which can effectively deal with the small sample size problem and experiments show that the proposed method is superior to the traditional KDA and LDA.
- 15
- 15
- A Hierarchical Framework Approach for Voice Activity Detection and Speech Enhancement15
Effectiveness of the proposed approach is compared and evaluated to other VAD techniques by using two well-known databases, namely, TIMIT database and NOISEX-92 database.
- Asymptotic stability of bidirectional associative memory neural networks with time-varying delays via delta operator approach15
A new delay-dependent criterion for analyzing the asymptotic stability of BAM neural networks with time-varying delays via delta operator approach is obtained based on the Lyapunov–Krasovskii functional in delta domain.
- 14
- 14
- Exploiting textual and visual features for image categorization14
This work forms image selection and classifier learning as a multi-instance learning problem and proposes to solve the employed problem by the cutting-plane algorithm to suppress the search error induced noisy images.
- Fast approximate L∞ minimization: Speeding up robust regression14
It is shown how the L∞ norm minimization problem can be broken up into smaller sub-problems, which can be solved extremely efficiently, and the radical reduction in computation time is demonstrated, along with robustness against large numbers of outliers in a few model-fitting problems.
- Local maximal margin discriminant embedding for face recognition14
A manifold learning based method named local maximal margin discriminant embedding (LMMDE) is developed for feature extraction that takes consideration of intra-class compactness and inter-class separability of samples lying in each manifold.
- 14
- Infrared human tracking with improved mean shift algorithm based on multicue fusion14
Compared with the traditional mean shift algorithm, the presented method greatly improves the accuracy and effectiveness of infrared human tracking under complex scenes, and the tracking results are satisfactory.
- 14
- Intelligent Service Recommendation for Cold-Start Problems in Edge Computing13
This work proposes an intelligent recommendation method named Inverse_CF_Rec, which first search for his/her opposite users, infer the possible friends of the target user indirectly according to Social Balance Theory, and optimal services are recommended to the target users based on the derived possibleFriends of thetarget user.
- Local Competition-Based Superpixel Segmentation Algorithm in Remote Sensing13
This paper proposes a superpixel segmentation algorithm called Superpixel Segmentation with Local Competition (SSLC), which utilizes a local competition mechanism to construct energy terms and label pixels and is less sensitive to the diversity of image content and scene layout.
- Fast total variation deconvolution for blurred image contaminated by Poisson noise13
A fast non-blind deconvolution method for restoring blurred images contaminated by Poisson noise that runs very fast and the quality of the restored image is comparable with that of some state of the art methods.
- 13
- 基于分维和三阶特征量的红外小目标检测13
The experimental results show that the proposed algorithm is robust and efficient for IR dim target detection and the performance and efficiency comparisons between the presented algorithm and the traditional fractal detection method on real IR images are illustrated.
- 12
- Speech Enhancement Based on NMF Under Electric Vehicle Noise Condition12
Speech enhancement and speaker verification experiments showed that the proposed ImNMF can effectively enhance speech signal in the noise environment of electric vehicles and further can reduce the equal error rate of the speaker verification system.
- Deep Learning for Person Reidentification Using Support Vector Machines12
Different from previous works, this work represents the pairs of pedestrian images as new resized input and use linear Support Vector Machine to replace softmax activation function for similarity learning and dropout and data augmentation techniques are employed in this model to prevent the network from overfitting.
- 12
- Proportional fair resource allocation based on hybrid ant colony optimization for slow adaptive OFDMA system12
A proportional fair resource allocation problem based on chance-constrained programming for slow adaptive OFDMA system which maximizes the average sum-rate in an adaptive time window and guarantees the Jain fair index (JFI) requirement with the target probability is formulated.
- Optimal Deployment and Scheduling with Directional Sensors for Energy-Efficient Barrier Coverage12
This paper addresses the efficient sensor deployment (ESD) problem and energy-efficient barrier coverage (EEBC) problem for directional sensor networks and proposes an efficient scheduling algorithm to prolong the network lifetime when the target area is barrier covered.
- Cooperative Search and Rescue with Artificial Fishes Based on Fish-Swarm Algorithm for Underwater Wireless Sensor Networks12
A fish-swarm based algorithm is designed requiring local information at each fish node and maximizing the joint detection probabilities of distress signals and it is showed that the control scheme has adaptive and effective properties.
- Robust subspace clustering via joint weighted Schatten-p norm and Lq norm minimization11
An improved LRR method is proposed, which achieves low rank property via the new formulation with weighted Schatten-p norm and Lq norm (WSPQ), and an efficient algorithm based on the inexact augmented Lagrange multiplier method is designed for the formulated problem.
- 11
- Deep Convex NMF for Image Clustering11
An efficient iterative optimization algorithm is developed, which can enhance the stability via layer-by-layer factorization and fine-tuning of feature representation in a novel deep convex non-negative matrix factorization method.
- Abnormal behavior recognition system for ATM monitoring by RGB-D camera11
The experimental results show that the real-time system for ATM intelligent monitoring by using Kinect has the advantages of robustness, and provides a new mean for preventing financial crimes.
- A Visibility-Based Algorithm for Multi-Robot Boundary Coverage11
A visibility-based decomposition algorithm for single robot boundary coverage and a corresponding multi-robot algorithm in unknown environment is proposed and a graph data structure is exploited for completeness of coverage and incremental description of partially observed world.
- A modification of kernel discriminant analysis for high-dimensional data—with application to face recognition10
This paper presents a variant of KDA that deals with both of the shortcomings in an efficient and cost effective manner and uses simultaneous diagonalization technique for optimization and a modified Fisher criterion that is more closely related to classification error.
- Study on Safety Speed Protection of the High Maglev Train10
The solving process and algorithm of safety speed protection curve are presented and it is pointed out that Ma- glev traia must operate in the range of the safety speed Protection curve even in the worst condition.
- 9
- 9
- 9
- 9
- 9
- Nearest-neighbor classifier motivated marginal discriminant projections for face recognition9
Experimental results show that NN-MDP not only avoids the intractability, and high expense of neighborhood parameter selection, but is also more applicable to face recognition with NN classifier than other methods.
- Single-Sample Face Recognition Based on Shared Generative Adversarial Network8
This paper proposes a shared generative adversarial network (SharedGAN) to expand the gallery dataset, Benefiting from the shared decoding network, which requires only a small number of training samples.
- 8
- Incomplete Variable Multigranulation Rough Sets Decision8
The incomplete variable multigranulation rough set (VMGRS) is proposed, and the relationships among VMGRS, OMGRS and PMGRS are deeply explored, and it can be found that OMG RS and PMgRS are the special cases compared to the VMG RS.
- Proportional fairness resource allocation scheme based on quantised feedback for multiuser orthogonal frequency division multiplexing system8
The proposed algorithm satisfies the constraint condition of the proportional fair very well, but also reduces the feedback overhead of the resource allocation result greatly, and the average throughput of the proposed algorithm is very close to that of the optimal resource allocation algorithm with full feedback when the equivalent channel gain in every subcarrier is quantised by 4 bit.
- Hybrid‐genetic‐algorithm‐based resource allocation for slow adaptive OFDMA system under channel uncertainty8
The neural network and the genetic algorithm are integrated to develop a hybrid GA (HGA) which could satisfy the user data rate requirement with the target outage probability and yields a higher long-term system throughput than the Li algorithm with the Bernstein approximation.
- 8
- A Variable Muitlgranulation Rough Sets Approach8
The concept of the variable multigranulation rough set is proposed, which is a generalization of both optimistic and pessimistic multigraphic rough set and which is meaningful for the development of multigraphene rough set theory.
- 8
- 8
- New algorithm of automatic water level measurement based on image processing8
This paper presents a new algorithm of automatic water lever detection based on image processing that has achieved high accuracy and practical application in hydrology.
- 8
- 8
- 7
- A Novel Multi-Sample Data Augmentation Method for Oriented Object Detection in Remote Sensing Images7
- Extracting Visual Knowledge from the Internet: Making Sense of Image Data7
This work evaluates the proposed webly supervised approach on the benchmark Pascal VOC 2007 dataset and the results demonstrates the superiority of the method over many other state-of-the-art methods in image data collection.
- 6
- Dynamic Re-Weighting and Cross-Camera Learning for Unsupervised Person Re-Identification6
Comprehensive experiments on the Market-1501 and DukeMTMC-reID datasets demonstrate that the proposed method vastly improves the performance of unsupervised domain adaptability.
- 6
- C2EM: cloud-assisted complex event monitoring in wireless multimedia sensor networks6
C2EM, a cloud-assisted complex event monitoring architecture that involves scalar sensors, camera sensors, cloudlets, and clouds to leverage computation offloading reliability, service response time, coverage efficiency, and energy efficiency is proposed.
- New Layered SOA-Based Architecture for Multi-Robots Cooperative Online SLAM6
A new layered multi-robots architecture for cooperative online simultaneous localization and map- ping based on Service-oriented architecture (SOA) is constructed by the concepts of service-oriented and loosely coupled center scheduling to make robots' underlying function be transparent in cooperation and avoid the impact affected by robot heterogeneous characteristics.
- Extensions of dominance-based rough set approach in incomplete information system6
An extension of the dominance-based rough set approach in incomplete information system is proposed by assuming the condition attributes, which are not only positively but also negatively related with classification analysis, and the concept of valued dominance relation is presented.
- 6
- 6
- 5
- Maximum Likelihood Estimation of Monocular Optical Flow Field for Mobile Robot Ego-Motion5
An optimized scheme of monocular ego-motion estimation to provide location and pose information for mobile robots with one fixed camera is presented and the estimated ego- motion parameters closely follow the GPS/INS ground truth in complex outdoor road scenarios.
- L2OR: Low-Cost Low-Complexity Opportunistic Routing for Wireless Sensor Networks5
A low-cost low-complexity opportunistic routing protocol to achieve an energy-efficient data delivery over WSNs and can achieve performance improvement in terms of data transmission efficiency as well as energy consumption.
- Fast Single Image Super Resolution Reconstruction via Image Separation5
This work proposes a fast single image super resolution reconstruction (SRR) approach via image separation based on the assumption that the edges, corners, and textures in the image have different mathematical models and applies different image SRR algorithms to process them individually.
- Distance Measurement in Visual Navigation of Monocular Autonomous Robots5
A monocular camera calibration technique based on B-dual-spacial geometry is brought forward and applied to visual distance-measurement of monocular autonomous robots and can compute camera's inner parameters more veraciously, and it can obviously reduce the total error in monocular visual Distance Measurement.
- 5
- 5
- Study on the Operation Control System Framework for High-Speed MAGLEV Train5
Based on the analysis of the OCS framework of German and Japanese system and the research experience of railway system, a new-style O CS framework for high-speed MAGLEV traffic was put forward and its function disassembly scheme was described.
- A Descriptive Model of Robot Team and the Dynamic Evolution of Robot Team Cooperation5
The dynamic evolution model of robot team and cooperative relationships between robot teams proposed and described in this paper can not only generalize the robot team as a whole, but also depict the dynamic evolving process quantitatively.
- 5
- Local Self-Attention on Fine-grained Cross-media Retrieval4
An attention space training method that utilizes local self-attention layers to learn the common attention space between different media data and a similarity concatenation method to understand the content relationship between features to effectively reduce the gap between different feature distributions on cross-media retrieval tasks.
- Exploiting Category Similarity-Based Distributed Labeling for Fine-Grained Visual Classification4
A simple yet effective approach named category similarity-based distributed labeling (CSDL) to tackle the problem of diversity within subcategories of fine-grained visual classification (FGVC).
- Uniform Generic Representation for Single Sample Face Recognition4
A uniform generic representation (UGR) method to solve the single sample per person (SSPP) problem in face recognition, which aims to find consistency between the global and local generic representations.
- 4
- Speaker Recognition of Noisy Short Utterance Based on Speech Frame Quality Discrimination and Three-stage Classification Model4
Experiments show that, the dual discrimination quality algorithms can be more precise to classify speech frame, and combining it with GMM-UBM three-stage classification model can make full use of limited corpus of short utterance and can improve the speaker recognition rate of the noisyShort utterance.
- Integrating virtual samples and fuzzy discriminant analysis for sparse representation-based face classification4
A framework that fuses virtual synthesized training samples as bases and fuzzy discriminant sparse residuals measurement in face classification and designed by the new weighted value, which can be merged into the typical discriminant analysis criterion.
- A novel SRC fusion method using hierarchical multi-scale LBP and greedy search strategy4
This paper proposes a novel SRC fusion method using hierarchical multi-scale local binary patterns (LBP) and greedy search strategy (SRC-GSLBP) for face recognition and develops a heuristic learning strategy to achieve this goal.
- The Speaker Recognition of Noisy Short Utterance4
Noise separation algorithm based on constrained Non-negative matrix factorization (CNMF) is proposed, use it to separate pure speech from noisy speech and speech frames are classified to high quality class and low quality class using differences detection and discrimination algorithm (DDADA).
- Optimal Multilevel Thresholding using Improved Gravitational Search Algorithm for Image Segmentation4
- A new approach of formation navigation derived from multi-robots cooperative online FastSLAM4
A new formation navigation approach derived from multi-robots cooperative online FastSLAM is proposed, in which the posteriori estimation of the leader robot state is treated as a relative reference for all follower robots to correct their state priori estimations.
- ACO Algorithm-based Parallel Job Scheduling Investigation on Hadoop4
This paper innovatively proposes that ACO(Ant Colony Optimization) can improve job scheduling performance dynamically with cost matrix, hormone matrix and probability matrix, and then simulates large amounts of jobs scheduling on cloud computing scene.
- 4
- 4
- 4
- RESEARCH ON OBSTACLE AVOIDANCE PLANNING FOR MOBILE ROBOTS IN DYNMIC WORLD4
This paper deals with coordination and obstacle (or collision) avoidance between mobile robots in conflict areas, that is, intersections, and an emulation result is given.
- 3
- 3
- 3
- A Fast Generalized Low Rank Representation Framework Based on $L_{2,p}$ Norm Minimization for Subspace Clustering3
Fast generalized LRR is proposed, which requires no SVD and thus has low computational cost and an efficient algorithm based on alternating direction method is designed to solve the proposed problem.
- 3
- Local and global regularized concept factorization for image clustering3
This work proposes an improved CF method, namely local and global regularized concept factorization (LGCF), by considering the local andglobal structures simultaneously, and establishes an unsupervised discriminant criterion, which characterizes the between-class scatter and the total scatter of the data with the help of latent features in LGCF.
- 3
- Robust Principal Component Analysis Based Speaker Verification Under Additive Noise Conditions3
The robust principal component analysis (RPCA) is introduced into the TVS modeled speaker verification system, called RPCA-TVS, which regards the noise spectrum as the low-rank component and the speech spectrum as a sparse component in short-time Fourier transform (SFT) domain.
- 3
- 3
- Local Similarity Based Linear Discriminant Analysis for Face Recognition with Single Sample per Person3
A novel method coined local similarity based linear discriminant analysis (LS_LDA) to solve the single sample per person problem (SSPP) and has strong robustness to expression, illumination, occlusion and time variation.
- 3
- 3
- 3
- Salient Object Detection Based on Regional Contrast and Relative Spatial Compactness3
Experiments show that the proposed algorithm outperforms several existing salient object detection methods with higher precision and better recall rates when evaluated using public datasets.
- 3
- Research on the Speaker Identification Based on Short Utterance3
A new approach which combines division and fuzzy k-means clustering is used, in order to optimize GMM initialization parameters, and the improved method is more effective in improving performance of the system than traditional initialization methods.
- Research on GMM text-independent speaker recognition3
This paper improves the traditional method of Gaussian Mixture Mode parameters initialization at the time of GMM training by presenting a new approach which combines division and K-means clustering and shows that the system can achieve the recognition rate with above 90% under the provided experimental condition.
- Analysis of Similarities and Differences of Operation Control Systems among Several Typical Modes of Rail Traffic3
Though there are great differences in the technical characteristics among the high speed railway,mass transit and high speed maglev trains, the main design idea and technical implementation methods of their operation control systems can be used for mutual reference.
- Combining wavelet packet with higher-order statistics for small IR targets detection3
Aiming to the problem of detecting small targets in a single frame infrared image, a new method combining wavelet packet with higher-order statistics (HOS) was presented and is robust and effective for small IR targets detection.
- Research of Vowel Mapping Theory for Speaker Identification of Chinese Mandarin3
A novel framework of speaker identification based on Chinese vowel mapping theory succeeds in separating personal identification information from semantic information based on classifying the Chinese vowel, which will be a new way to transform the text-independent system into text-dependent speaker recognition system.
- Speaker identification using multi-step clustering algorithm with transformation-based GMM3
A new model structure is presented that integrates a multi-step cluster algorithm into the estimating process of GMM with the embedded transformation according to a maximum likelihood criterion and shows that this method outperforms the traditionalGMM with cluster EM algorithm.
- 3
- SCGRNet: shape completion-guided registration network for face point cloud2
A shape completion-guided registration network for face point cloud registration is proposed, which takes face shape as a prior knowledge to guide the network to learn a complete face features at the low overlap.
- Rapid Person Re-Identification via Sub-space Consistency Regularization2
A novel Sub-space Consistency Regularization (SCR) algorithm that can speed up the ReID procedure by 0.25 times than real-value features under same dimensions whilst maintain a competitive accuracy, especially under short codes.
- 2
- 2
- 2
- Patch based Semi-supervised Linear Regression for Face Recognition2
A patch based semi-supervised linear regression (PSLR) algorithm, which draws facial variation information from unlabeled samples, and a multistage PSLR (MPSLR) algorithm, which iteratively updates the training dataset by adding those reliably labeled probe samples into it.
- 2
- 2
- 2
- Idology and Its Applications in Public Security and Network Security2
The authors propose the concept of idology - a systematic and scientific study of identifications of persons and things, and give the definitions of a symmetric identity and an asymmetric identity, and compare asymmetric RFIDs with BFIDs.
- 2
- Exploiting Visual Saliency and Bag-of-Words for Road Sign Recognition2
A novel method for road sign detection and recognition in complex scene real world images using a regional contrast based bottom-up visual saliency method to highlight the traffic sign regions, which usually have dominant color contrast against the background.
- Research on Noise Processing and Speech Frame Classification Model of Noisy Short UtteranceSignal Processing2
The improved SNR discrimination algorithm and the differences detection and discrimination algorithm are proposed and combined with the GMM-UBM three-stage classification model proposed so that the noisy short utterance can make full use of the limit corpus.
- 2
- 2
- 2
- 2
- Architecture Design for Intelligent Vehicle Computing Platform Based on Internet of Vehicles2
The intelligent vehicle computing technology is discussed and the architecture of Intelligent vehicle computing platform is proposed, which is key component of Internet of Vehicles.
- 2
- Secure XML keyword search based on compression policies.2
This method with compacted policies provides important basis for secure and efficient XML Key words query and adopts the extensible Dewey coding which makes the query easy.
- 2
- ASM and Color Gabor Features for Facial Feature Extraction2
In order to improve active shape Model (ASM) accuracy in facial feature points location in facial images, Color Gabor features was used to extract edges and corner points for feature, so the authors could get the key facial feature Points quickly and accurately.
- Rough Set Theory in Incomplete Interval-valued Information System2
These two types of reduction are minimal subsets of the attributes that can preserve the lower and upper approximations of all decision classes respectively and with these reductions, the simplified decision rules can be obtained.
- 2
- 2
- Transformation-Based GMM with Improved Cluster Algorithm for Speaker Identification2
A new model is presented that integrates an improved cluster algorithm into the estimating process of GMM with the embedded transformation and shows that this method outperforms the traditionalGMM with cluster EM algorithm.
- 2
- Design of ATS Simulation System of Beijing Line One Subway2
It was expatiated the total network structure of ATS Simulation System for beijing's first one subway based on this, mostly researched on the Management Database Subsystem of the train running graph information.
- One effective method for ship recognition in ship locks2
The paper presents one effective method for ship recognition in the ship lock that analyzes their characters on gray level and structure, proposes a new method to form a special histogram of only those pixels besides the boundary, fit for small object segment and also large.
- 1
- Generative Adversarial and Self-Attention Based Fine-Grained Cross-Media Retrieval1
This work proposes a novel fine-grained cross-media retrieval algorithm, which is based on the generative adversarial network and self-attention mechanism, and demonstrates the improving results compared to other counterpart methods.
- 1
- 1
- 1
- A Semi-supervised Learning Algorithm Based on Low Rank and Weighted Sparse Graph for Face Recognition1
A novel low rank and weighted sparse graph is proposed that improves the accuracy of semi-supervised learning and achieves the state-of-the-art performance.
- Food Safety Evaluation System Construction Based on Artificial Neural Network1
The results showed that the BP neural network was high-precision, fast and objective, which could be used to food safety evaluation of circulation links of production, processing and sales.
- 1
- 1
- 1
- Local Geometry Feature Extraction for Face Sketch Recognition1
A new feature extraction approach for face sketch recognition based on the assumption that small image patches in the photo and sketch images form manifold with similar local geometry in two different image spaces that outperforms state-of-the-art methods.
- 1
- 1
- Autonomous Navigation Trajectory Evaluation Method Based on Uncertainty Analysis1
Experiment results show that this trajectory analysis method can effectively be used in trajectory evaluation of autonomous navigation, and it can compensate shortcomings of reinforcement learning methods in stability evaluation.
- 1
- Supervised Kernel Self-Organizing Map1
This work generalizes the traditional supervised self-organizing map to supervised kernel self- Organizing map by incorporating the kernel function to further improve its capability of solving non-linear problems.
- A GOP-aware multipath source routing protocol for MPEG video delivery over mobile ad hoc networks1
A group of picture (GOP)-aware multipath source routing protocol (GMSR) for real-time moving picture experts group (MPEG) video delivery in mobile ad hoc networks (MANETs) and introduces a GOP-aware flow control strategy (GFC) to coordinate video delivery, working together with the routing protocol.
- The Algorithm for Unstructured Road to Abstract Road Surface based on Color Feature1
The experiments show that this algorithm could identify a wide range of unstructured roads base on only part of the image color characteristics and could basically meet the real-time requirement of the mobile robot during outdoor industry.
- Multi-EKF Localization Algorithm for Multi-Robots Formation Navigation1
This research presents a dynamic multi-robots formation navigation algorithm named multi-EKF localization algorithm, which derived from single EKF, which shows the feasibility, efficiency and stability of the proposed algorithm.
- Negative decision rules and knowledge reduction in incomplete information system1
An approach of discernibility matrix is proposed to obtain the reduction of the negative rules from the incomplete decision information system, and the results of the illustrative examples show the effectiveness of the approach.
- Dewey Encoding Generation Algorithm of XML Document1
Experimental result shows that the SAX event generation algorithm is with higher speed and lower memory consumption for very large XML document.
- 1
- Application of Particle Filter Based on Feature Fusion in Small IR Target Tracking1
This algorithm combines the gray feature with fractal feature, and then uses the fusion results to calculate the particle weights, which greatly improves the robustness of the tracking algorithm.
- 1
- 1
- Speaker Identification Algorithm of Semi-Supervised Learning Mechanism1
Experimental shows that this new method of complex Gaussian Mixture Model based on Semi-Supervised Learning can take full advantage of unlabeled data on the self-adaptive updating, improve the performance, get higher recognition rate compared with traditional GMM, and effectively enhance the generalization ability of system.
- A study on speaker identification based on weighted LS-SVM1
In order to improve the pairwise LS-SVM and make it applicable to multi-speaker identification system, a new notion of classification weight for pairwise LTVMs and the corresponding algorithm is proposed, named as pairwise LSVM based on classification weight, i.e., the m-ωLS-S VM method, which can be used in multi- Speakers identification system.
- Camera calibration based on vanishing points in B-dual-space geometry1
A camera calibration method based on vanishing points based on corresponding subject in B-dual-space geometry rooted in computational geometry that has the merits of simpleness and high precision in experimentation is introduced.
- Non-parametric transform of color image for stereo matching1
A new stereo matching approach based on matching primitives from improved rank transform about color image with absolute-value exponent method to solve the problems such as image noise and brightness difference in stereo images.
- New method for automatic road rut detection1
A rapid and effective calibration method based on broad is proposed to calculate rut depth, which has high coherence comparing with that obtained by people, and has much higher precision and efficiency than other similar instrument.
- 1
- Generate Fuzzy Rules of Controller Parameters Automatically Using PSO1
A novel controller with fuzzy non-linear parameter was proposed that was proved to be global stable with Lyapunov theory and to solve the problems of time consuming and hard to satisfy the control demands when seeking the parameters.
- Study on Urban Rail Transit Model Based on the Max-Plus Algebraic Theory1
An open loop urban rail transit linear model built by using the max-plus algebraic approach is shown to be non-blocking and robust to the perturbation of the time parameters when the delivery headway wasn't smaller than the maximum value of the train operation time matrix.
- 1
- Method for the traffic volume detection based on Bayes parameter estimation and weighted data fusion algorithm1
It is proved that the weighted data fusion algorithm based on Bayes parameter estimation is not only simple and efficient, but also convenient for programming.
- Design of Data Analysis System of railway station graph1
It was described the basic principle of deriving and collecting from the static data of the railway station graph, and given the basic methods of the System.
- Study on Operation Safety Information System of Urban Rail Transit1
The operation safety information system built was beneficial in reducing somewhat the loss caused by accident and improving the management efficiency and is an innovational exploration in raising the level of automation of the urban rail transit.
- –
- Enhancement of the Learning Experience of Foreign Graduate Exchange students and Visiting Scholars–
George Karady was appointed to Salt River Project Chair Professor at Arizona State University in 1986, where he is responsible for the electrical power education and performs research in Power Electronics, High Voltage Techniques and Electric Power.
- –
- Towards robust visual explanations for deep convolutional networks with weight-wise perturbations–
The introduction of a novel approach that injects noise into network weights to enhance visualization, involving image gradient updates and average gradient computations, and a new indicator for evaluating interpretability - the center of gravity are proposed.
- –
- –
- –
- Pedestrian Attribute Recognition with Occlusion in Low Resolution Surveillance Scenarios–
Experiments show that this improved pedestrian attribute recognition method proposed performs excellent when the pedestrian images suffer occlusion problem and the method is robust to low resolution problem.
- Robust Speaker Verification via RPCA under Additive Noise.–
Robust principal component analysis (RPCA) is introduced into a TVS modeled speaker verification system, called the RPCA-TVS, to improve the robustness of speaker verification under additive noisy environments, particularly for non-stationary and unseen noises.
- –
- –
- A Real-Time local path planning method based on SVM for UGV–
In this approach, given grid map from 64-beam LiDAR, obstacles on both sides of the road are regarded as two classes fed to Support Vector Machine (SVM) to generate an initial safe path.
- –
- –
- –
- Nonnegative matrix factorization with region sparsity learning for hyperspectral unmixing–
This paper proposes an improved NMF method, namely NMF with region sparsity learning (RSLNMF), to simultaneously consider both spectral and spatial information, and defines a new sparsitylearning model based on a small homogeneous region obtained via the graph cut algorithm.
- –
- –
- A Method for Three-Dimensional Reconstruction of a Train Accident Scene using Photographs–
This work proposes a photo-based three-dimensional reconstruction framework of vehicles for measuring the positions and poses of carriages involved in an accident and implements and examines two case studies to validate this reconstruction method, which performs well in the assigned tasks.
- Consensus Filtering Algorithm with Incomplete Measurements–
The numerical simulation demonstrates that the algorithm proposed in this paper can improve the accuracy and consensus of estimation, and it is more robust with incomplete measurements.
- The Application of Electronic Forensics Technology in Food Safety Traceability System under the Environment of Cloud Computing–
The realization of the forensics technology in the food safety traceability system under the environment of cloud computing is discussed, so as to maximize the benefits of consumer's safety.
- An OCL-Based Formal Model for Cloud Forensics–
An extensible semantic model using OCL (Object Constraint Language) for cloud digital forensics data is proposed and a feature tree can be built with semantic logical relation in order to get the overall semantic description of features in the forensic domain of digital data.
- Real-Time Head Pose Estimation by RGB-D Camera–
Experimental results have shown that the use of both depth data and RGB data in the method overcomes the shortcomings of single depth or RGB data, and is still robust when there is only one type of data reliable.
- Robust FHPD Features from Speech Harmonic Analysis for Speaker Identification–
A robust speech feature extraction method, which is based on the harmonic structure of voiced segments, which reflects that under low SNR (signal-to-noise ratio) environments new features achieve better performance than conventional MFCC (Mel-Frequency Cepstral Coefficients) parameters.
- –
- –
- –
- –
- Design of PG digital protection and monitor device–
A design method of protection and monitor device based on the graphics and network that can collect the protected data timely in the workstation and display the device's various kinds of configuration information and real-time running state is presented.
- –
- Detection and tracking based on motion template under strong disturbance–
It has as a quality of practical navigation and control about self-determined robot and unpiloted vehicle based on real-time computer vision about moving object's detection and tracking.
- Implementation of Anycast Service in Mobile IPv6 Network–
A new kind of Anycast communication model in mobile IPv6 networks which is built on a tree topology is proposed which can perform IPSec which is difficult to implement in the existing mobile IP communication models.
- Self-adaptable Granularity Road Network Moving Objects' Clustering Algorithm–
This paper introduced a novel method to train historical data with BP network, and then got clustering distance granularity and clustering time granularity dynamically and these granularity values can be made to train BP network further.
- –
- Application of improved fractal algorithm in small target detection–
The region to be processed by the fractal algorithm is reduced to a small range that is estimated by local entropy,thus overcoming the defect of huge computational cost and poor anti-noise capability of traditional fractal methods.
- DRSA and reductions in incomplete fuzzy information system–
This paper presents a general framework for the study of dominance-based rough set approach to the incomplete fuzzy information systems, and proposes two types of knowledge reductions, relative lower and upper approximate reducts, which can be used to induce simplified decision rules from the complete fuzzy decision table.
- Scheme on implementing interconnection between wireless sensor networks and IPv6 networks–
The author proposed and designed a model to implement a seamless communication between a Wireless Sensor Network (WSN) and an IPv6 network that creates an automatic address configuration algorithm in a sensor network and based on the algorithm this model performs automatic routing and addressing.
- –
- A Maglev Transportation Simulation Platform Using UML–
A visible model of Maglev simulation platform is build in the environment of Rational Rose using Unified Modeling Language (UML) and the idea of the Object-Oriented Analysis (OOA) to confirm the correctness and feasibility of the platform.
- A Convex Partition Algorithm for a Polygon with Holes–
The experiment and analysis show that the average complexity of the algorithm approximates O(nlgn); the algorithm makes convex partition of a polygon with holes using monotone chains.
- Solving Anycast Scalability with Multicast Tree in IPv6–
A new kind of Anycast communication model was proposed on the basis of multicast tree technology that achieves dynamic Anycast group, and the experimental data in IPv6 simulation demonstrated that the TRT of one Anycast service acquired through this communication model is shorter than the one through the current communication model.
- DISCRIMINATIVE FEATURE EXTRACTION BASED ON SELF-ADAPTIVE FREQUENCY WARPING FOR ROBUST SPEAKER IDENTIFICATION–
It is shown that the theory of this feature is reasonable and understandable, which is insensitive to spoken content and thus more discriminative and robust in comparison to the conventional Mel frequency cepstral coefficients.
- Vector quantization technology based on clustering features in speaker recognition–
A method in which speaker feature based on speech clustering to vector quantization in speaker recognition is proposed, which shows that it could reduce the number of codebook from 64 with simple vector quantification to 8 with VQ based on clustering features.
- Implementation of serial-production-line operations based on maximum-algebra with Matlab–
It was implemented the operating in the modeling and characteristic analyzing of a serial production line with unlimited buffers based on maximum algebra by Matlab and provided more advantages for the application of maximum algebra theory.
- Frame for multi-robot motion planning based on local sensing and communication–
An integrated frame based on local sensing and communication is described and simulated for motion planning which is demanded by many Multi-Robot tasks and includes an environmental model, a distributed architecture, a dynamic communication network, planning and coordination algorithms.
- –
- Planning method of Beijing future public transit network based on small world network–
It was mainly introduced the efficiency definition of small world network and its application in public transit planning, and an example to explain the application inpublic transit planning is given.
- –
- Simulation of Maglev ATO Control System based on MATLAB–
Based on the simulation toolbox of MATLAB a simulation method about Maglev’s ATO Control System showed that the control effect was better in control performance.
- –
- A Micro-Simulation-based Evaluation System for Thruway Traffic Accidents–
Based on vehicle microscopic behavior model, the paper developed the evaluation system for effect of thruway traffic accident by adopting object-oriented software technology that can simulate movement of vehicle, state of traffic accident and data detector.
- –
- –
- Inquire on System of Analysis for Expressway Operation–
The characters of expressway economy and technology decide the high standard and the high starting point and how to collect the data and give them statistics and analysis is the key point.
- Study on the Mechanism of Safe Drawing Exchange Format(DXF) Fife Information–
A reliable arithmetic has been put forward through analyzing the framework of DXF file and the result in word file format has been given.
- Dynamic Planning of Obstacle Avoidance in Multiple Robots System–
This paper deals with the cordination and obstacle avoidance among multiple robots in conflict areas and the planning approach both concentrated and distributed is adopted.
- Real-time Human's Shape Monitoring & Recognition System Using Moment Invariants–
This paper firstly pretreats the image, which includes other object besides background that is, getting rid of background image and obtaining the shape image of the object and calculates the seven moment invariants of that image.
- Split-and-merge segmentation using relation stable state–
Experiments show the proposed segmentation method, which is based on the relation stable-state, can be applied to extract objects with multi-level perfectly even if the image is non-uniform illumination, thus it is more general and practical.
- –
- Image Preprocessing for Ship Detection Under Complex Conditions–
The image preprocess methods for ship detection method under complex conditions is introduced, which has been successfully running in the temporary ship lock of Three Gorges.
- A NEW APPROACH OF ALV GLOBAL SPEED CONTROL–
Some new ideas are proposed such as speed plan and hierarchical strategy of global speed control and common sense and experiential procedure knowledge can be easily expressed and utilized.
- Novel approach of better understanding the complicated environment from the laser radar's range data–
This work proposes a novel approach by using two matrixes from the range data to solve the problem of understanding the front environment more precise and better.
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-15. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.