Visible-infrared Cross-modality Person Re-identification(VI-ReID)is a critical technology in smart public facilities such as cities,campuses and libraries.It aims to match pedestrians in visible light and infrared ima...Visible-infrared Cross-modality Person Re-identification(VI-ReID)is a critical technology in smart public facilities such as cities,campuses and libraries.It aims to match pedestrians in visible light and infrared images for video surveillance,which poses a challenge in exploring cross-modal shared information accurately and efficiently.Therefore,multi-granularity feature learning methods have been applied in VI-ReID to extract potential multi-granularity semantic information related to pedestrian body structure attributes.However,existing research mainly uses traditional dual-stream fusion networks and overlooks the core of cross-modal learning networks,the fusion module.This paper introduces a novel network called the Augmented Deep Multi-Granularity Pose-Aware Feature Fusion Network(ADMPFF-Net),incorporating the Multi-Granularity Pose-Aware Feature Fusion(MPFF)module to generate discriminative representations.MPFF efficiently explores and learns global and local features with multi-level semantic information by inserting disentangling and duplicating blocks into the fusion module of the backbone network.ADMPFF-Net also provides a new perspective for designing multi-granularity learning networks.By incorporating the multi-granularity feature disentanglement(mGFD)and posture information segmentation(pIS)strategies,it extracts more representative features concerning body structure information.The Local Information Enhancement(LIE)module augments high-performance features in VI-ReID,and the multi-granularity joint loss supervises model training for objective feature learning.Experimental results on two public datasets show that ADMPFF-Net efficiently constructs pedestrian feature representations and enhances the accuracy of VI-ReID.展开更多
Distributed denial of service(DDoS)attacks launch more and more frequently and are more destructive.Feature representation as an important part of DDoS defense technology directly affects the efficiency of defense.Mos...Distributed denial of service(DDoS)attacks launch more and more frequently and are more destructive.Feature representation as an important part of DDoS defense technology directly affects the efficiency of defense.Most DDoS feature extraction methods cannot fully utilize the information of the original data,resulting in the extracted features losing useful features.In this paper,a DDoS feature representation method based on deep belief network(DBN)is proposed.We quantify the original data by the size of the network flows,the distribution of IP addresses and ports,and the diversity of packet sizes of different protocols and train the DBN in an unsupervised manner by these quantified values.Two feedforward neural networks(FFNN)are initialized by the trained deep belief network,and one of the feedforward neural networks continues to be trained in a supervised manner.The canonical correlation analysis(CCA)method is used to fuse the features extracted by two feedforward neural networks per layer.Experiments show that compared with other methods,the proposed method can extract better features.展开更多
In the era of Big data,learning discriminant feature representation from network traffic is identified has as an invariably essential task for improving the detection ability of an intrusion detection system(IDS).Owin...In the era of Big data,learning discriminant feature representation from network traffic is identified has as an invariably essential task for improving the detection ability of an intrusion detection system(IDS).Owing to the lack of accurately labeled network traffic data,many unsupervised feature representation learning models have been proposed with state-of-theart performance.Yet,these models fail to consider the classification error while learning the feature representation.Intuitively,the learnt feature representation may degrade the performance of the classification task.For the first time in the field of intrusion detection,this paper proposes an unsupervised IDS model leveraging the benefits of deep autoencoder(DAE)for learning the robust feature representation and one-class support vector machine(OCSVM)for finding the more compact decision hyperplane for intrusion detection.Specially,the proposed model defines a new unified objective function to minimize the reconstruction and classification error simultaneously.This unique contribution not only enables the model to support joint learning for feature representation and classifier training but also guides to learn the robust feature representation which can improve the discrimination ability of the classifier for intrusion detection.Three set of evaluation experiments are conducted to demonstrate the potential of the proposed model.First,the ablation evaluation on benchmark dataset,NSL-KDD validates the design decision of the proposed model.Next,the performance evaluation on recent intrusion dataset,UNSW-NB15 signifies the stable performance of the proposed model.Finally,the comparative evaluation verifies the efficacy of the proposed model against recently published state-of-the-art methods.展开更多
Symbolic Aggregate approXimation (SAX) is an efficient symbolic representation method that has been widely used in time series data mining. Its major limitation is that it relies exclusively on the mean values of segm...Symbolic Aggregate approXimation (SAX) is an efficient symbolic representation method that has been widely used in time series data mining. Its major limitation is that it relies exclusively on the mean values of segmented time series to derive the symbols. So, many important features of time series are not considered, such as extreme value, trend, fluctuation and so on. To solve this issue, we propose in this paper an improved Symbolic Aggregate approXimation based on multiple features and Vector Frequency Difference (SAX_VFD). SAX_VFD discriminates between time series by adopting an adaptive feature selection method. Furthermore, SAX_VFD is endowed with a new distance that takes into account the vector frequency difference between the symbolic sequence. We demonstrate the utility of the SAX_VFD on the time series classification task. The experimental results show that the proposed method has a better performance in terms of accuracy and dimensionality reduction compared to the so far published SAX based reduction techniques.展开更多
Feature representation is one of the key issues in data clustering. The existing feature representation of scientific data is not sufficient, which to some extent affects the result of scientific data clustering. Ther...Feature representation is one of the key issues in data clustering. The existing feature representation of scientific data is not sufficient, which to some extent affects the result of scientific data clustering. Therefore, the paper proposes a concept of composite text description(CTD) and a CTD-based feature representation method for biomedical scientific data. The method mainly uses different feature weight algorisms to represent candidate features based on two types of data sources respectively, combines and finally strengthens the two feature sets. Experiments show that comparing with traditional methods, the feature representation method is more effective than traditional methods and can significantly improve the performance of biomedcial data clustering.展开更多
Unsupervised feature selection has become an important and challenging problem faced with vast amounts of unlabeled and high-dimension data in machine learning. We propose a novel unsupervised feature selection method...Unsupervised feature selection has become an important and challenging problem faced with vast amounts of unlabeled and high-dimension data in machine learning. We propose a novel unsupervised feature selection method using Structured Self-Representation( SSR) by simultaneously taking into account the selfrepresentation property and local geometrical structure of features. Concretely,according to the inherent selfrepresentation property of features,the most representative features can be selected. Mean while,to obtain more accurate results,we explore local geometrical structure to constrain the representation coefficients to be close to each other if the features are close to each other. Furthermore,an efficient algorithm is presented for optimizing the objective function. Finally,experiments on the synthetic dataset and six benchmark real-world datasets,including biomedical data,letter recognition digit data and face image data,demonstrate the encouraging performance of the proposed algorithm compared with state-of-the-art algorithms.展开更多
Single-feature methods are unable to effectively track a target in an underground coal mine video due to the high background noise, low and uneven illumination, and drastic light fluctuation in the video. In this stud...Single-feature methods are unable to effectively track a target in an underground coal mine video due to the high background noise, low and uneven illumination, and drastic light fluctuation in the video. In this study, we propose an underground coal mine personnel target tracking method using multi-feature joint sparse representation. First, with a particle filter framework, the global and local multiple features of the target template and candidate particles are extracted. Second, each of the candidate particles is sparsely represented by a dictionary template, and reconstruction is achieved after solving the sparse coefficient. Last, the particle with the lowest reconstruction error is deemed the tracking result. To validate the effectiveness of the proposed algorithm, we compare the proposed method with three commonly employed tracking algorithms. The results show that the proposed method is able to reliably track the target in various scenarios, such as occlusion and illumination change, which generates better tracking results and validates the feasibility and effectiveness of the proposed method.展开更多
Unconstrained face images are interfered by many factors such as illumination,posture,expression,occlusion,age,accessories and so on,resulting in the randomness of the noise pollution implied in the original samples.I...Unconstrained face images are interfered by many factors such as illumination,posture,expression,occlusion,age,accessories and so on,resulting in the randomness of the noise pollution implied in the original samples.In order to improve the sample quality,a weighted block cooperative sparse representation algorithm is proposed based on visual saliency dictionary.First,the algorithm uses the biological visual attention mechanism to quickly and accurately obtain the face salient target and constructs the visual salient dictionary.Then,a block cooperation framework is presented to perform sparse coding for different local structures of human face,and the weighted regular term is introduced in the sparse representation process to enhance the identification of information hidden in the coding coefficients.Finally,by synthesising the sparse representation results of all visual salient block dictionaries,the global coding residual is obtained and the class label is given.The experimental results on four databases,that is,AR,extended Yale B,LFW and PubFig,indicate that the combination of visual saliency dictionary,block cooperative sparse representation and weighted constraint coding can effectively enhance the accuracy of sparse representation of the samples to be tested and improve the performance of unconstrained face recognition.展开更多
Fine-grained image search is one of the most challenging tasks in computer vision that aims to retrieve similar images at the fine-grained level for a given query image.The key objective is to learn discriminative fin...Fine-grained image search is one of the most challenging tasks in computer vision that aims to retrieve similar images at the fine-grained level for a given query image.The key objective is to learn discriminative fine-grained features by training deep models such that similar images are clustered,and dissimilar images are separated in the low embedding space.Previous works primarily focused on defining local structure loss functions like triplet loss,pairwise loss,etc.However,training via these approaches takes a long training time,and they have poor accuracy.Additionally,representations learned through it tend to tighten up in the embedded space and lose generalizability to unseen classes.This paper proposes a noise-assisted representation learning method for fine-grained image retrieval to mitigate these issues.In the proposed work,class manifold learning is performed in which positive pairs are created with noise insertion operation instead of tightening class clusters.And other instances are treated as negatives within the same cluster.Then a loss function is defined to penalize when the distance between instances of the same class becomes too small relative to the noise pair in that class in embedded space.The proposed approach is validated on CARS-196 and CUB-200 datasets and achieved better retrieval results(85.38%recall@1 for CARS-196%and 70.13%recall@1 for CUB-200)compared to other existing methods.展开更多
Background Deep 3D morphable models(deep 3DMMs)play an essential role in computer vision.They are used in facial synthesis,compression,reconstruction and animation,avatar creation,virtual try-on,facial recognition sys...Background Deep 3D morphable models(deep 3DMMs)play an essential role in computer vision.They are used in facial synthesis,compression,reconstruction and animation,avatar creation,virtual try-on,facial recognition systems and medical imaging.These applications require high spatial and perceptual quality of synthesised meshes.Despite their significance,these models have not been compared with different mesh representations and evaluated jointly with point-wise distance and perceptual metrics.Methods We compare the influence of different mesh representation features to various deep 3DMMs on spatial and perceptual fidelity of the reconstructed meshes.This paper proves the hypothesis that building deep 3DMMs from meshes represented with global representations leads to lower spatial reconstruction error measured with L_(1) and L_(2) norm metrics and underperforms on perceptual metrics.In contrast,using differential mesh representations which describe differential surface properties yields lower perceptual FMPD and DAME and higher spatial fidelity error.The influence of mesh feature normalisation and standardisation is also compared and analysed from perceptual and spatial fidelity perspectives.Results The results presented in this paper provide guidance in selecting mesh representations to build deep 3DMMs accordingly to spatial and perceptual quality objectives and propose combinations of mesh representations and deep 3DMMs which improve either perceptual or spatial fidelity of existing methods.展开更多
In this paper, we present a novel and efficient scheme for extracting, indexing and retrieving color images. Our motivation was to reduce the space overhead of partition-based approaches taking advantage of the fact t...In this paper, we present a novel and efficient scheme for extracting, indexing and retrieving color images. Our motivation was to reduce the space overhead of partition-based approaches taking advantage of the fact that only a relatively low number of distinct values of a particular visual feature is present in most images. To extract color feature and build indices into our image database we take into consideration factors such as human color perception and perceptual range, and the image is partitioned into a set of regions by using a simple classifying scheme. The compact color feature vector and the spatial color histogram, which are extracted from the seqmented image region, are used for representing the color and spatial information in the image. We have also developed the region-based distance measures to compare the similarity of two images. Extensive tests on a large image collection were conducted to demonstrate the effectiveness of the proposed approach.展开更多
Human body feature extraction based on 2D images provides an efficient method for many applications, e.g. non-contact body size measurements, constructing 3D human model and recognizing human actions. In this paper a ...Human body feature extraction based on 2D images provides an efficient method for many applications, e.g. non-contact body size measurements, constructing 3D human model and recognizing human actions. In this paper a systematic approach is proposed to detect feature points of human body automatically from its front and side images. Firstly, an efficient approach for silhouette and contour detection is used to represent the contour curves of a human body shape with Freeman’s 8-connected chain codes. The contour curves are considered as a number of segments connected together. Then, a series of feature points on human body are extracted based on the specified rules by measuring the differences between the directions of the segments. In total, 101 feature points with clearly geometric properties (that rather accurately reflect the bump or turning of the contours) are extracted automatically, including 27 points corresponding to the definitions of the landmarks about garment measurements. Finally, the proposed approach was tested on ten human subjects and the entire 101 feature points with specific geography geometrical characteristics were correctly extracted, indicating an effective and robust performance.展开更多
因为它的 insensitivity, 3D 脸识别吸引越来越多的注意到照明和姿势的变化。有在这个话题要解决的许多关键问题,例如 3D 脸表示和有效多特征熔化。在这份报纸,一个新奇 3D 脸识别算法被建议,它的性能在 BJUT-3D 脸数据库上被表明...因为它的 insensitivity, 3D 脸识别吸引越来越多的注意到照明和姿势的变化。有在这个话题要解决的许多关键问题,例如 3D 脸表示和有效多特征熔化。在这份报纸,一个新奇 3D 脸识别算法被建议,它的性能在 BJUT-3D 脸数据库上被表明。这个算法选择脸表面性质和相对关系矩阵的原则部件为脸表示特征。为每个特征的类似公制被定义。特征熔化策略被建议。它基于菲希尔是线性加权的策略线性判别式分析。最后,介绍算法在 BJUT-3D 脸数据库上被测试。算法和熔化策略的表演是令人满意的,这被结束。展开更多
Current Chinese event detection methods commonly use word embedding to capture semantic representation,but these methods find it difficult to capture the dependence relationship between the trigger words and other wor...Current Chinese event detection methods commonly use word embedding to capture semantic representation,but these methods find it difficult to capture the dependence relationship between the trigger words and other words in the same sentence.Based on the simple evaluation,it is known that a dependency parser can effectively capture dependency relationships and improve the accuracy of event categorisation.This study proposes a novel architecture that models a hybrid representation to summarise semantic and structural information from both characters and words.This model can capture rich semantic features for the event detection task by incorporating the semantic representation generated from the dependency parser.The authors evaluate different models on kbp 2017 corpus.The experimental results show that the proposed method can significantly improve performance in Chinese event detection.展开更多
With the development of modern industry, sheet-metal parts in mass production have been widely applied in mechanical, communication, electronics, and light industries in recent decades; but the advances in sheet-metal...With the development of modern industry, sheet-metal parts in mass production have been widely applied in mechanical, communication, electronics, and light industries in recent decades; but the advances in sheet-metal part design and manufacturing remain too slow compared with the increasing importance of sheet-metal parts in modern industry. This paper proposes a method for automatically extracting features from an arbitrary solid model of sheet-metal parts; whose characteristics are used for classification and graph-based representation of the sheet-metal features to extract the features embodied in a sheet-metal part. The extracting feature process can be divided for valid checking of the model geometry, feature matching, and feature relationship. Since the extracted features include abundant geometry and engineering information, they will be effective for downstream application such as feature rebuilding and stamping process planning.展开更多
This paper proposed a feature extraction scheme based on sparse representation considering the nonstationary property of surface electromyography( sEMG). Sparse Bayesian Learning( SBL) algorithm was introduced to extr...This paper proposed a feature extraction scheme based on sparse representation considering the nonstationary property of surface electromyography( sEMG). Sparse Bayesian Learning( SBL) algorithm was introduced to extract the feature with optimal class separability to improve recognition accuracies of multimovement patterns. The SBL algorithm exploited the compressibility( or weak sparsity) of sEMG signal in some transformed domains. The proposed feature extracted by using the SBL algorithm was named SRC. The feature SRC represented time-varying characteristics of sEMG signal very effectively. We investigated the effect of the feature SRC by comparing with other fourteen individual features and eighteen multi-feature sets in offline recognition. The results demonstrated the feature SRC revealed the important dynamic information in the sEMG signals. And the multi-feature sets formed by the feature SRC and other single features yielded more superior performance on recognition accuracy. The best average recognition accuracy of 91. 67% was gained by using SVM classifier with the multi-feature set combining the feature SRC and the feature wavelength( WL). The proposed feature extraction scheme is promising for multi-movement recognition with high accuracy.展开更多
Remaining time prediction of business processes plays an important role in resource scheduling and plan making.The structural features of single process instance and the concurrent running of multiple process instance...Remaining time prediction of business processes plays an important role in resource scheduling and plan making.The structural features of single process instance and the concurrent running of multiple process instances are the main factors that affect the accuracy of the remaining time prediction.Existing prediction methods does not take full advantage of these two aspects into consideration.To address this issue,a new prediction method based on trace representation is proposed.More specifically,we first associate the prefix set generated by the event log to different states of the transition system,and encode the structural features of the prefixes in the state.Then,an annotation containing the feature representation for the prefix and the corresponding remaining time are added to each state to obtain an extended transition system.Next,states in the extended transition system are partitioned by the different lengths of the states,which considers concurrency among multiple process instances.Finally,the long short-term memory(LSTM)deep recurrent neural networks are applied to each partition for predicting the remaining time of new running instances.By extensive experimental evaluation using synthetic event logs and reallife event logs,we show that the proposed method outperforms existing baseline methods.展开更多
Automated defect detection in woven fabrics for quality control is still a challenging novelty detection problem,while the efficient representation of fabric texture is essential for it.This paper presents a novel met...Automated defect detection in woven fabrics for quality control is still a challenging novelty detection problem,while the efficient representation of fabric texture is essential for it.This paper presents a novel method for fabric texture representation.Benefiting from the characteristics of the weaving process,the major texture information of woven fabric is concentrated in the warp and weft directions.Thus,the proposed method is firstly to project the image patch along warp and weft directions to obtain projected vectors containing warp and weft informations.Secondly,the obtained vectors instead of image patch,are used to extract the features that are able to represent fabric texture.Finally,the t-test is applied to verifying the usefulness of the proposed method in discriminating defective and normal fabric textures.The experiments on various defective samples demonstrate that the method yields a robust and good performance in representing fabric texture and discriminating defects.展开更多
Geoscience knowledge graph(GKG)can organize various geoscience knowledge into a machine understandable and computable semantic network and is an effective way to organize geoscience knowledge and provide knowledge-rel...Geoscience knowledge graph(GKG)can organize various geoscience knowledge into a machine understandable and computable semantic network and is an effective way to organize geoscience knowledge and provide knowledge-related services.As a result,it has gained significant attention and become a frontier in geoscience.Geoscience knowledge is derived from many disciplines and has complex spatiotemporal features and relationships of multiple scales,granularities,and dimensions.Therefore,establishing a GKG representation model conforming to the characteristics of geoscience knowledge is the basis and premise for the construction and application of GKG.However,existing knowledge graph representation models leverage fixed tuples that are limited in fully representing complex spatiotemporal features and relationships.To address this issue,this paper first systematically analyzes the categorization and spatiotemporal features and relationships of geoscience knowledge.On this basis,an adaptive representation model for GKG is proposed by considering the complex spatiotemporal features and relationships.Under the constraint of a unified spatiotemporal ontology,this model adopts different tuples to adaptively represent different types of geoscience knowledge according to their spatiotemporal correlation.This model can efficiently represent geoscience knowledge,thereby avoiding the isolation of the spatiotemporal feature representation and improving the accuracy and efficiency of geoscience knowledge retrieval.It can further enable the alignment,transformation,computation,and reasoning of spatiotemporal information through a spatiotemporal ontology.展开更多
基金supported in part by the National Natural Science Foundation of China under Grant 62177029,62307025in part by the Startup Foundation for Introducing Talent of Nanjing University of Posts and Communications under Grant NY221041in part by the General Project of The Natural Science Foundation of Jiangsu Higher Education Institution of China 22KJB520025,23KJD580.
文摘Visible-infrared Cross-modality Person Re-identification(VI-ReID)is a critical technology in smart public facilities such as cities,campuses and libraries.It aims to match pedestrians in visible light and infrared images for video surveillance,which poses a challenge in exploring cross-modal shared information accurately and efficiently.Therefore,multi-granularity feature learning methods have been applied in VI-ReID to extract potential multi-granularity semantic information related to pedestrian body structure attributes.However,existing research mainly uses traditional dual-stream fusion networks and overlooks the core of cross-modal learning networks,the fusion module.This paper introduces a novel network called the Augmented Deep Multi-Granularity Pose-Aware Feature Fusion Network(ADMPFF-Net),incorporating the Multi-Granularity Pose-Aware Feature Fusion(MPFF)module to generate discriminative representations.MPFF efficiently explores and learns global and local features with multi-level semantic information by inserting disentangling and duplicating blocks into the fusion module of the backbone network.ADMPFF-Net also provides a new perspective for designing multi-granularity learning networks.By incorporating the multi-granularity feature disentanglement(mGFD)and posture information segmentation(pIS)strategies,it extracts more representative features concerning body structure information.The Local Information Enhancement(LIE)module augments high-performance features in VI-ReID,and the multi-granularity joint loss supervises model training for objective feature learning.Experimental results on two public datasets show that ADMPFF-Net efficiently constructs pedestrian feature representations and enhances the accuracy of VI-ReID.
基金supported by the National Natural Science Foundation of Hainan(2018CXTD333,617048)National Natural Science Foundation of China(61762033,61702539)+4 种基金The National Natural Science Foundation of Hunan(2018JJ3611)Social Development Project of Public Welfare Technology Application of Zhejiang Province(LGF18F020019)Hainan University Doctor Start Fund Project(kyqd1328)Hainan University Youth Fund Project(qnjj1444)State Key Laboratory of Marine Resource Utilization in South China Sea Funding.
文摘Distributed denial of service(DDoS)attacks launch more and more frequently and are more destructive.Feature representation as an important part of DDoS defense technology directly affects the efficiency of defense.Most DDoS feature extraction methods cannot fully utilize the information of the original data,resulting in the extracted features losing useful features.In this paper,a DDoS feature representation method based on deep belief network(DBN)is proposed.We quantify the original data by the size of the network flows,the distribution of IP addresses and ports,and the diversity of packet sizes of different protocols and train the DBN in an unsupervised manner by these quantified values.Two feedforward neural networks(FFNN)are initialized by the trained deep belief network,and one of the feedforward neural networks continues to be trained in a supervised manner.The canonical correlation analysis(CCA)method is used to fuse the features extracted by two feedforward neural networks per layer.Experiments show that compared with other methods,the proposed method can extract better features.
基金This work was supported by the Research Deanship of Prince Sattam Bin Abdulaziz University,Al-Kharj,Saudi Arabia(Grant No.2020/01/17215).Also,the author thanks Deanship of college of computer engineering and sciences for technical support provided to complete the project successfully。
文摘In the era of Big data,learning discriminant feature representation from network traffic is identified has as an invariably essential task for improving the detection ability of an intrusion detection system(IDS).Owing to the lack of accurately labeled network traffic data,many unsupervised feature representation learning models have been proposed with state-of-theart performance.Yet,these models fail to consider the classification error while learning the feature representation.Intuitively,the learnt feature representation may degrade the performance of the classification task.For the first time in the field of intrusion detection,this paper proposes an unsupervised IDS model leveraging the benefits of deep autoencoder(DAE)for learning the robust feature representation and one-class support vector machine(OCSVM)for finding the more compact decision hyperplane for intrusion detection.Specially,the proposed model defines a new unified objective function to minimize the reconstruction and classification error simultaneously.This unique contribution not only enables the model to support joint learning for feature representation and classifier training but also guides to learn the robust feature representation which can improve the discrimination ability of the classifier for intrusion detection.Three set of evaluation experiments are conducted to demonstrate the potential of the proposed model.First,the ablation evaluation on benchmark dataset,NSL-KDD validates the design decision of the proposed model.Next,the performance evaluation on recent intrusion dataset,UNSW-NB15 signifies the stable performance of the proposed model.Finally,the comparative evaluation verifies the efficacy of the proposed model against recently published state-of-the-art methods.
文摘Symbolic Aggregate approXimation (SAX) is an efficient symbolic representation method that has been widely used in time series data mining. Its major limitation is that it relies exclusively on the mean values of segmented time series to derive the symbols. So, many important features of time series are not considered, such as extreme value, trend, fluctuation and so on. To solve this issue, we propose in this paper an improved Symbolic Aggregate approXimation based on multiple features and Vector Frequency Difference (SAX_VFD). SAX_VFD discriminates between time series by adopting an adaptive feature selection method. Furthermore, SAX_VFD is endowed with a new distance that takes into account the vector frequency difference between the symbolic sequence. We demonstrate the utility of the SAX_VFD on the time series classification task. The experimental results show that the proposed method has a better performance in terms of accuracy and dimensionality reduction compared to the so far published SAX based reduction techniques.
基金supported by the Agridata,the sub-program of National Science and Technology Infrastructure Program(Grant No.2005DKA31800)
文摘Feature representation is one of the key issues in data clustering. The existing feature representation of scientific data is not sufficient, which to some extent affects the result of scientific data clustering. Therefore, the paper proposes a concept of composite text description(CTD) and a CTD-based feature representation method for biomedical scientific data. The method mainly uses different feature weight algorisms to represent candidate features based on two types of data sources respectively, combines and finally strengthens the two feature sets. Experiments show that comparing with traditional methods, the feature representation method is more effective than traditional methods and can significantly improve the performance of biomedcial data clustering.
基金Sponsored by the Major Program of National Natural Science Foundation of China(Grant No.13&ZD162)the Applied Basic Research Programs of China National Textile and Apparel Council(Grant No.J201509)
文摘Unsupervised feature selection has become an important and challenging problem faced with vast amounts of unlabeled and high-dimension data in machine learning. We propose a novel unsupervised feature selection method using Structured Self-Representation( SSR) by simultaneously taking into account the selfrepresentation property and local geometrical structure of features. Concretely,according to the inherent selfrepresentation property of features,the most representative features can be selected. Mean while,to obtain more accurate results,we explore local geometrical structure to constrain the representation coefficients to be close to each other if the features are close to each other. Furthermore,an efficient algorithm is presented for optimizing the objective function. Finally,experiments on the synthetic dataset and six benchmark real-world datasets,including biomedical data,letter recognition digit data and face image data,demonstrate the encouraging performance of the proposed algorithm compared with state-of-the-art algorithms.
文摘Single-feature methods are unable to effectively track a target in an underground coal mine video due to the high background noise, low and uneven illumination, and drastic light fluctuation in the video. In this study, we propose an underground coal mine personnel target tracking method using multi-feature joint sparse representation. First, with a particle filter framework, the global and local multiple features of the target template and candidate particles are extracted. Second, each of the candidate particles is sparsely represented by a dictionary template, and reconstruction is achieved after solving the sparse coefficient. Last, the particle with the lowest reconstruction error is deemed the tracking result. To validate the effectiveness of the proposed algorithm, we compare the proposed method with three commonly employed tracking algorithms. The results show that the proposed method is able to reliably track the target in various scenarios, such as occlusion and illumination change, which generates better tracking results and validates the feasibility and effectiveness of the proposed method.
基金Natural Science Foundation of Jiangsu Province,Grant/Award Number:BK20170765National Natural Science Foundation of China,Grant/Award Number:61703201+1 种基金Future Network Scientific Research Fund Project,Grant/Award Number:FNSRFP2021YB26Science Foundation of Nanjing Institute of Technology,Grant/Award Numbers:ZKJ202002,ZKJ202003,and YKJ202019。
文摘Unconstrained face images are interfered by many factors such as illumination,posture,expression,occlusion,age,accessories and so on,resulting in the randomness of the noise pollution implied in the original samples.In order to improve the sample quality,a weighted block cooperative sparse representation algorithm is proposed based on visual saliency dictionary.First,the algorithm uses the biological visual attention mechanism to quickly and accurately obtain the face salient target and constructs the visual salient dictionary.Then,a block cooperation framework is presented to perform sparse coding for different local structures of human face,and the weighted regular term is introduced in the sparse representation process to enhance the identification of information hidden in the coding coefficients.Finally,by synthesising the sparse representation results of all visual salient block dictionaries,the global coding residual is obtained and the class label is given.The experimental results on four databases,that is,AR,extended Yale B,LFW and PubFig,indicate that the combination of visual saliency dictionary,block cooperative sparse representation and weighted constraint coding can effectively enhance the accuracy of sparse representation of the samples to be tested and improve the performance of unconstrained face recognition.
文摘Fine-grained image search is one of the most challenging tasks in computer vision that aims to retrieve similar images at the fine-grained level for a given query image.The key objective is to learn discriminative fine-grained features by training deep models such that similar images are clustered,and dissimilar images are separated in the low embedding space.Previous works primarily focused on defining local structure loss functions like triplet loss,pairwise loss,etc.However,training via these approaches takes a long training time,and they have poor accuracy.Additionally,representations learned through it tend to tighten up in the embedded space and lose generalizability to unseen classes.This paper proposes a noise-assisted representation learning method for fine-grained image retrieval to mitigate these issues.In the proposed work,class manifold learning is performed in which positive pairs are created with noise insertion operation instead of tightening class clusters.And other instances are treated as negatives within the same cluster.Then a loss function is defined to penalize when the distance between instances of the same class becomes too small relative to the noise pair in that class in embedded space.The proposed approach is validated on CARS-196 and CUB-200 datasets and achieved better retrieval results(85.38%recall@1 for CARS-196%and 70.13%recall@1 for CUB-200)compared to other existing methods.
基金Supported by the Centre for Digital Entertainment at Bournemouth University by the UK Engineering and Physical Sciences Research Council(EPSRC)EP/L016540/1 and Humain Ltd.
文摘Background Deep 3D morphable models(deep 3DMMs)play an essential role in computer vision.They are used in facial synthesis,compression,reconstruction and animation,avatar creation,virtual try-on,facial recognition systems and medical imaging.These applications require high spatial and perceptual quality of synthesised meshes.Despite their significance,these models have not been compared with different mesh representations and evaluated jointly with point-wise distance and perceptual metrics.Methods We compare the influence of different mesh representation features to various deep 3DMMs on spatial and perceptual fidelity of the reconstructed meshes.This paper proves the hypothesis that building deep 3DMMs from meshes represented with global representations leads to lower spatial reconstruction error measured with L_(1) and L_(2) norm metrics and underperforms on perceptual metrics.In contrast,using differential mesh representations which describe differential surface properties yields lower perceptual FMPD and DAME and higher spatial fidelity error.The influence of mesh feature normalisation and standardisation is also compared and analysed from perceptual and spatial fidelity perspectives.Results The results presented in this paper provide guidance in selecting mesh representations to build deep 3DMMs accordingly to spatial and perceptual quality objectives and propose combinations of mesh representations and deep 3DMMs which improve either perceptual or spatial fidelity of existing methods.
文摘In this paper, we present a novel and efficient scheme for extracting, indexing and retrieving color images. Our motivation was to reduce the space overhead of partition-based approaches taking advantage of the fact that only a relatively low number of distinct values of a particular visual feature is present in most images. To extract color feature and build indices into our image database we take into consideration factors such as human color perception and perceptual range, and the image is partitioned into a set of regions by using a simple classifying scheme. The compact color feature vector and the spatial color histogram, which are extracted from the seqmented image region, are used for representing the color and spatial information in the image. We have also developed the region-based distance measures to compare the similarity of two images. Extensive tests on a large image collection were conducted to demonstrate the effectiveness of the proposed approach.
文摘Human body feature extraction based on 2D images provides an efficient method for many applications, e.g. non-contact body size measurements, constructing 3D human model and recognizing human actions. In this paper a systematic approach is proposed to detect feature points of human body automatically from its front and side images. Firstly, an efficient approach for silhouette and contour detection is used to represent the contour curves of a human body shape with Freeman’s 8-connected chain codes. The contour curves are considered as a number of segments connected together. Then, a series of feature points on human body are extracted based on the specified rules by measuring the differences between the directions of the segments. In total, 101 feature points with clearly geometric properties (that rather accurately reflect the bump or turning of the contours) are extracted automatically, including 27 points corresponding to the definitions of the landmarks about garment measurements. Finally, the proposed approach was tested on ten human subjects and the entire 101 feature points with specific geography geometrical characteristics were correctly extracted, indicating an effective and robust performance.
基金Supported by National Natural Science Foundation of China (60533030) and Beijing Natural Science Foundation (4061001)
文摘因为它的 insensitivity, 3D 脸识别吸引越来越多的注意到照明和姿势的变化。有在这个话题要解决的许多关键问题,例如 3D 脸表示和有效多特征熔化。在这份报纸,一个新奇 3D 脸识别算法被建议,它的性能在 BJUT-3D 脸数据库上被表明。这个算法选择脸表面性质和相对关系矩阵的原则部件为脸表示特征。为每个特征的类似公制被定义。特征熔化策略被建议。它基于菲希尔是线性加权的策略线性判别式分析。最后,介绍算法在 BJUT-3D 脸数据库上被测试。算法和熔化策略的表演是令人满意的,这被结束。
基金973 Program,Grant/Award Number:2014CB340504The State Key Program of National Natural Science of China,Grant/Award Number:61533018+3 种基金National Natural Science Foundation of China,Grant/Award Number:61402220The Philosophy and Social Science Foundation of Hunan Province,Grant/Award Number:16YBA323Natural Science Foundation of Hunan Province,Grant/Award Number:2020JJ4525Scientific Research Fund of Hunan Provincial Education Department,Grant/Award Number:18B279,19A439。
文摘Current Chinese event detection methods commonly use word embedding to capture semantic representation,but these methods find it difficult to capture the dependence relationship between the trigger words and other words in the same sentence.Based on the simple evaluation,it is known that a dependency parser can effectively capture dependency relationships and improve the accuracy of event categorisation.This study proposes a novel architecture that models a hybrid representation to summarise semantic and structural information from both characters and words.This model can capture rich semantic features for the event detection task by incorporating the semantic representation generated from the dependency parser.The authors evaluate different models on kbp 2017 corpus.The experimental results show that the proposed method can significantly improve performance in Chinese event detection.
文摘With the development of modern industry, sheet-metal parts in mass production have been widely applied in mechanical, communication, electronics, and light industries in recent decades; but the advances in sheet-metal part design and manufacturing remain too slow compared with the increasing importance of sheet-metal parts in modern industry. This paper proposes a method for automatically extracting features from an arbitrary solid model of sheet-metal parts; whose characteristics are used for classification and graph-based representation of the sheet-metal features to extract the features embodied in a sheet-metal part. The extracting feature process can be divided for valid checking of the model geometry, feature matching, and feature relationship. Since the extracted features include abundant geometry and engineering information, they will be effective for downstream application such as feature rebuilding and stamping process planning.
文摘This paper proposed a feature extraction scheme based on sparse representation considering the nonstationary property of surface electromyography( sEMG). Sparse Bayesian Learning( SBL) algorithm was introduced to extract the feature with optimal class separability to improve recognition accuracies of multimovement patterns. The SBL algorithm exploited the compressibility( or weak sparsity) of sEMG signal in some transformed domains. The proposed feature extracted by using the SBL algorithm was named SRC. The feature SRC represented time-varying characteristics of sEMG signal very effectively. We investigated the effect of the feature SRC by comparing with other fourteen individual features and eighteen multi-feature sets in offline recognition. The results demonstrated the feature SRC revealed the important dynamic information in the sEMG signals. And the multi-feature sets formed by the feature SRC and other single features yielded more superior performance on recognition accuracy. The best average recognition accuracy of 91. 67% was gained by using SVM classifier with the multi-feature set combining the feature SRC and the feature wavelength( WL). The proposed feature extraction scheme is promising for multi-movement recognition with high accuracy.
基金supported by National Natural Science Foundation of China(No.U1931207 and No.61702306)Sci.&Tech.Development Fund of Shandong Province of China(No.ZR2019LZH001,No.ZR2017BF015 and No.ZR2017MF027)+4 种基金the Humanities and Social Science Research Project of the Ministry of Education(No.18YJAZH017)Shandong Chongqing Science and technology cooperation project(No.cstc2020jscx-lyjsAX0008)Sci.&Tech.Development Fund of Qingdao(No.21-1-5-zlyj-1-zc)the Taishan Scholar Program of Shandong ProvinceSDUST Research Fund(No.2015TDJH102 and No.2019KJN024).
文摘Remaining time prediction of business processes plays an important role in resource scheduling and plan making.The structural features of single process instance and the concurrent running of multiple process instances are the main factors that affect the accuracy of the remaining time prediction.Existing prediction methods does not take full advantage of these two aspects into consideration.To address this issue,a new prediction method based on trace representation is proposed.More specifically,we first associate the prefix set generated by the event log to different states of the transition system,and encode the structural features of the prefixes in the state.Then,an annotation containing the feature representation for the prefix and the corresponding remaining time are added to each state to obtain an extended transition system.Next,states in the extended transition system are partitioned by the different lengths of the states,which considers concurrency among multiple process instances.Finally,the long short-term memory(LSTM)deep recurrent neural networks are applied to each partition for predicting the remaining time of new running instances.By extensive experimental evaluation using synthetic event logs and reallife event logs,we show that the proposed method outperforms existing baseline methods.
基金National Natural Science Foundations of China(Nos.61501209,61379011)the Foundamental Research Funds for the Central Universities of China(Nos.JUSRP51631A,JUSRP11506)
文摘Automated defect detection in woven fabrics for quality control is still a challenging novelty detection problem,while the efficient representation of fabric texture is essential for it.This paper presents a novel method for fabric texture representation.Benefiting from the characteristics of the weaving process,the major texture information of woven fabric is concentrated in the warp and weft directions.Thus,the proposed method is firstly to project the image patch along warp and weft directions to obtain projected vectors containing warp and weft informations.Secondly,the obtained vectors instead of image patch,are used to extract the features that are able to represent fabric texture.Finally,the t-test is applied to verifying the usefulness of the proposed method in discriminating defective and normal fabric textures.The experiments on various defective samples demonstrate that the method yields a robust and good performance in representing fabric texture and discriminating defects.
基金supported by the National Natural Science Foundation of China(Grant No.42050101)the National Key Research and Development Program of China(Grant Nos.2022YFB3904200&2021YFB00903)supported by the International Big Science Program of Deeptime Digital Earth(DDE)。
文摘Geoscience knowledge graph(GKG)can organize various geoscience knowledge into a machine understandable and computable semantic network and is an effective way to organize geoscience knowledge and provide knowledge-related services.As a result,it has gained significant attention and become a frontier in geoscience.Geoscience knowledge is derived from many disciplines and has complex spatiotemporal features and relationships of multiple scales,granularities,and dimensions.Therefore,establishing a GKG representation model conforming to the characteristics of geoscience knowledge is the basis and premise for the construction and application of GKG.However,existing knowledge graph representation models leverage fixed tuples that are limited in fully representing complex spatiotemporal features and relationships.To address this issue,this paper first systematically analyzes the categorization and spatiotemporal features and relationships of geoscience knowledge.On this basis,an adaptive representation model for GKG is proposed by considering the complex spatiotemporal features and relationships.Under the constraint of a unified spatiotemporal ontology,this model adopts different tuples to adaptively represent different types of geoscience knowledge according to their spatiotemporal correlation.This model can efficiently represent geoscience knowledge,thereby avoiding the isolation of the spatiotemporal feature representation and improving the accuracy and efficiency of geoscience knowledge retrieval.It can further enable the alignment,transformation,computation,and reasoning of spatiotemporal information through a spatiotemporal ontology.