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
The main purpose of nonlinear time series analysis is based on the rebuilding theory of phase space, and to study how to transform the response signal to rebuilt phase space in order to extract dynamic feature informa...The main purpose of nonlinear time series analysis is based on the rebuilding theory of phase space, and to study how to transform the response signal to rebuilt phase space in order to extract dynamic feature information, and to provide effective approach for nonlinear signal analysis and fault diagnosis of nonlinear dynamic system. Now, it has already formed an important offset of nonlinear science. But, traditional method cannot extract chaos features automatically, and it needs man＇s participation in the whole process. A new method is put forward, which can implement auto-extracting of chaos features for nonlinear time series. Firstly, to confirm time delay r by autocorrelation method; Secondly, to compute embedded dimension m and correlation dimension D; Thirdly, to compute the maximum Lyapunov index λmax; Finally, to calculate the chaos degree Dch of Poincare map, and the non-circle degree Dnc and non-order degree Dno of quasi-phase orbit. Chaos features extracting has important meaning to fault diagnosis of nonlinear system based on nonlinear chaos features. Examples show validity of the proposed method.展开更多
Hematoxylin and Eosin(H&E)images,popularly used in the field of digital pathology,often pose challenges due to their limited color richness,hindering the differentiation of subtle cell features crucial for accurat...Hematoxylin and Eosin(H&E)images,popularly used in the field of digital pathology,often pose challenges due to their limited color richness,hindering the differentiation of subtle cell features crucial for accurate classification.Enhancing the visibility of these elusive cell features helps train robust deep-learning models.However,the selection and application of image processing techniques for such enhancement have not been systematically explored in the research community.To address this challenge,we introduce Salient Features Guided Augmentation(SFGA),an approach that strategically integrates machine learning and image processing.SFGA utilizes machine learning algorithms to identify crucial features within cell images,subsequently mapping these features to appropriate image processing techniques to enhance training images.By emphasizing salient features and aligning them with corresponding image processing methods,SFGA is designed to enhance the discriminating power of deep learning models in cell classification tasks.Our research undertakes a series of experiments,each exploring the performance of different datasets and data enhancement techniques in classifying cell types,highlighting the significance of data quality and enhancement in mitigating overfitting and distinguishing cell characteristics.Specifically,SFGA focuses on identifying tumor cells from tissue for extranodal extension detection,with the SFGA-enhanced dataset showing notable advantages in accuracy.We conducted a preliminary study of five experiments,among which the accuracy of the pleomorphism experiment improved significantly from 50.81%to 95.15%.The accuracy of the other four experiments also increased,with improvements ranging from 3 to 43 percentage points.Our preliminary study shows the possibilities to enhance the diagnostic accuracy of deep learning models and proposes a systematic approach that could enhance cancer diagnosis,contributing as a first step in using SFGA in medical image enhancement.展开更多
Advanced Persistent Threats(APTs)pose significant challenges to detect due to their“low-and-slow”attack patterns and frequent use of zero-day vulnerabilities.Within this task,the extraction of long-term features is ...Advanced Persistent Threats(APTs)pose significant challenges to detect due to their“low-and-slow”attack patterns and frequent use of zero-day vulnerabilities.Within this task,the extraction of long-term features is often crucial.In this work,we propose a novel end-to-end APT detection framework named Long-Term Feature Association Provenance Graph Detector(LT-ProveGD).Specifically,LT-ProveGD encodes contextual information of the dynamic provenance graph while preserving the topological information with space efficiency.To combat“low-and-slow”attacks,LT-ProveGD develops an autoencoder with an integrated multi-head attention mechanism to extract long-term dependencies within the encoded representations.Furthermore,to facilitate the detection of previously unknown attacks,we leverage Jenks’natural breaks methodology,enabling detection without relying on specific attack information.By conducting extensive experiments on five widely used datasets with state-of-the-art attack detection methods,we demonstrate the superior effectiveness of LT-ProveGD.展开更多
Considering that the algorithm accuracy of the traditional sparse representation models is not high under the influence of multiple complex environmental factors,this study focuses on the improvement of feature extrac...Considering that the algorithm accuracy of the traditional sparse representation models is not high under the influence of multiple complex environmental factors,this study focuses on the improvement of feature extraction and model construction.Firstly,the convolutional neural network(CNN)features of the face are extracted by the trained deep learning network.Next,the steady-state and dynamic classifiers for face recognition are constructed based on the CNN features and Haar features respectively,with two-stage sparse representation introduced in the process of constructing the steady-state classifier and the feature templates with high reliability are dynamically selected as alternative templates from the sparse representation template dictionary constructed using the CNN features.Finally,the results of face recognition are given based on the classification results of the steady-state classifier and the dynamic classifier together.Based on this,the feature weights of the steady-state classifier template are adjusted in real time and the dictionary set is dynamically updated to reduce the probability of irrelevant features entering the dictionary set.The average recognition accuracy of this method is 94.45%on the CMU PIE face database and 96.58%on the AR face database,which is significantly improved compared with that of the traditional face recognition methods.展开更多
In response to the challenges in highway pavement distress detection,such as multiple defect categories,difficulties in feature extraction for different damage types,and slow identification speeds,this paper proposes ...In response to the challenges in highway pavement distress detection,such as multiple defect categories,difficulties in feature extraction for different damage types,and slow identification speeds,this paper proposes an enhanced pavement crack detection model named Star-YOLO11.This improved algorithm modifies the YOLO11 architecture by substituting the original C3k2 backbone network with a Star-s50 feature extraction network.The enhanced structure adjusts the number of stacked layers in the StarBlock module to optimize detection accuracy and improve model efficiency.To enhance the accuracy of pavement crack detection and improve model efficiency,three key modifications to the YOLO11 architecture are proposed.Firstly,the original C3k2 backbone is replaced with a StarBlock-based structure,forming the Star-s50 feature extraction backbone network.This lightweight redesign reduces computational complexity while maintaining detection precision.Secondly,to address the inefficiency of the original Partial Self-attention(PSA)mechanism in capturing localized crack features,the convolutional prior-aware Channel Prior Convolutional Attention(CPCA)mechanism is integrated into the channel dimension,creating a hybrid CPC-C2PSA attention structure.Thirdly,the original neck structure is upgraded to a Star Multi-Branch Auxiliary Feature Pyramid Network(SMAFPN)based on the Multi-Branch Auxiliary Feature Pyramid Network architecture,which adaptively fuses high-level semantic and low-level spatial information through Star-s50 connections and C3k2 extraction blocks.Additionally,a composite dataset augmentation strategy combining traditional and advanced augmentation techniques is developed.This strategy is validated on a specialized pavement dataset containing five distinct crack categories for comprehensive training and evaluation.Experimental results indicate that the proposed Star-YOLO11 achieves an accuracy of 89.9%(3.5%higher than the baseline),a mean average precision(mAP)of 90.3%(+2.6%),and an F1-score of 85.8%(+0.5%),while reducing the model size by 18.8%and reaching a frame rate of 225.73 frames per second(FPS)for real-time detection.It shows potential for lightweight deployment in pavement crack detection tasks.展开更多
Parkinson’s disease(PD)is a debilitating neurological disorder affecting over 10 million people worldwide.PD classification models using voice signals as input are common in the literature.It is believed that using d...Parkinson’s disease(PD)is a debilitating neurological disorder affecting over 10 million people worldwide.PD classification models using voice signals as input are common in the literature.It is believed that using deep learning algorithms further enhances performance;nevertheless,it is challenging due to the nature of small-scale and imbalanced PD datasets.This paper proposed a convolutional neural network-based deep support vector machine(CNN-DSVM)to automate the feature extraction process using CNN and extend the conventional SVM to a DSVM for better classification performance in small-scale PD datasets.A customized kernel function reduces the impact of biased classification towards the majority class(healthy candidates in our consideration).An improved generative adversarial network(IGAN)was designed to generate additional training data to enhance the model’s performance.For performance evaluation,the proposed algorithm achieves a sensitivity of 97.6%and a specificity of 97.3%.The performance comparison is evaluated from five perspectives,including comparisons with different data generation algorithms,feature extraction techniques,kernel functions,and existing works.Results reveal the effectiveness of the IGAN algorithm,which improves the sensitivity and specificity by 4.05%–4.72%and 4.96%–5.86%,respectively;and the effectiveness of the CNN-DSVM algorithm,which improves the sensitivity by 1.24%–57.4%and specificity by 1.04%–163%and reduces biased detection towards the majority class.The ablation experiments confirm the effectiveness of individual components.Two future research directions have also been suggested.展开更多
Numerous vibration-based techniques are rarely used in diesel engines fault diagnosis in a direct way, due to the surface vibration signals of diesel engines with the complex non-stationary and nonlinear time-varying ...Numerous vibration-based techniques are rarely used in diesel engines fault diagnosis in a direct way, due to the surface vibration signals of diesel engines with the complex non-stationary and nonlinear time-varying fea- tures. To investigate the fault diagnosis of diesel engines, fractal correlation dimension, wavelet energy and entropy as features reflecting the diesel engine fault fractal and energy characteristics are extracted from the decomposed signals through analyzing vibration acceleration signals derived from the cylinder head in seven different states of valve train. An intelligent fault detector FastICA-SVM is applied for diesel engine fault diagnosis and classification. The results demonstrate that FastlCA-SVM achieves higher classification accuracy and makes better general- ization performance in small samples recognition. Besides, the fractal correlation dimension and wavelet energy and entropy as the special features of diesel engine vibration signal are considered as input vectors of classifier FastlCA- SVM and could produce the excellent classification results. The proposed methodology improves the accuracy of fea- ture extraction and the fault diagnosis of diesel engines.展开更多
A novel class of periodically changing features hidden in radar pulse sequence environment,named G features,is proposed.Combining fractal theory and Hilbert-Huang transform,the features are extracted using changing ch...A novel class of periodically changing features hidden in radar pulse sequence environment,named G features,is proposed.Combining fractal theory and Hilbert-Huang transform,the features are extracted using changing characteristics of pulse parameters in radar emitter signals.The features can be applied in modern complex electronic warfare environment to address the issue of signal sorting when radar emitter pulse signal parameters severely or even completely overlap.Experiment results show that the proposed feature class and feature extraction method can discriminate periodically changing pulse sequence signal sorting features from radar pulse signal flow with complex variant features,therefore provide a new methodology for signal sorting.展开更多
For measurement of component content in the extraction and separation process of praseodymium/neodymium(Pr/Nd), a soft measurement method was proposed based on modeling of ion color features, which is suitable for fas...For measurement of component content in the extraction and separation process of praseodymium/neodymium(Pr/Nd), a soft measurement method was proposed based on modeling of ion color features, which is suitable for fast estimation of component content in production field. Feature analysis on images of the solution is conducted,which are captured from Pr/Nd extraction/separation field. H/S components in the HSI color space are selected as model inputs, so as to establish the least squares support vector machine(LSSVM) model for Nd(Pr) content,while the model parameters are determined with the GA algorithm. To improve the adaptability of the model,the adaptive iteration algorithm is used to correct parameters of the LSSVM model, on the basis of model correction strategy and new sample data. Using the field data collected from rare earth extraction production, predictive methods for component content and comparisons are given. The results indicate that the proposed method presents good adaptability and high prediction precision, so it is applicable to the fast detection of element content in the rare earth extraction.展开更多
The rapid growth of multimedia content necessitates powerful technologies to filter, classify, index and retrieve video documents more efficiently. However, the essential bottleneck of image and video analysis is the ...The rapid growth of multimedia content necessitates powerful technologies to filter, classify, index and retrieve video documents more efficiently. However, the essential bottleneck of image and video analysis is the problem of semantic gap that low level features extracted by computers always fail to coincide with high-level concepts interpreted by humans. In this paper, we present a generic scheme for the detection video semantic concepts based on multiple visual features machine learning. Various global and local low-level visual features are systelrtically investigated, and kernelbased learning method equips the concept detection system to explore the potential of these features. Then we combine the different features and sub-systen on both classifier-level and kernel-level fusion that contribute to a more robust system Our proposed system is tested on the TRECVID dataset. The resulted Mean Average Precision （MAP） score is rmch better than the benchmark perforrmnce, which proves that our concepts detection engine develops a generic model and perforrrs well on both object and scene type concepts.展开更多
Feature information extraction is one of the key steps in prognostics and health management of rotating machinery.In the present study,an investigation about the feasibility of a methodology based on generalized S tra...Feature information extraction is one of the key steps in prognostics and health management of rotating machinery.In the present study,an investigation about the feasibility of a methodology based on generalized S transform(GST)and singular value decomposition(SVD)methods for feature extraction in rolling bearing,due to local damage under variable conditions,is conducted.The technique adopts the GST method,following the time-frequency analysis,to transform a raw fault signal of the rolling bearing into a two-dimensional complex matrix.And then,the SVD method is performed to decompose the matrix to obtain the feature vectors.By this procedure it is possible to obtain the fault feature information of rolling bearing under different speeds and different loads.In order to streamline the feature parameters of the feature vectors to train more uncomplicated models,the principal component analysis(PCA)subsequently performed.The particle swarm optimization-support vector machine(PSO-SVM)model is used to identify and classify the different fault states of rolling bearing.Furthermore,in order to highlight the superiority of the proposed method some comparisons are conducted with the conventional methods.The obtained results show that the proposed method can effectively extract fault features of the rolling bearing under variable conditions.展开更多
The rapid identification of pathogens is crucial in controlling the food quality and safety.The proposed system for the rapid and label-free identification of pathogens is based on the principle of laser scattering fr...The rapid identification of pathogens is crucial in controlling the food quality and safety.The proposed system for the rapid and label-free identification of pathogens is based on the principle of laser scattering from the bacterial microbes.The clinical prototype consists of three parts:the laser beam,photodetectors,and the data acquisition system.The bacterial testing sample was mixed with 10 mL distilled water and placed inside the machine chamber.When the bacterial microbes pass by the laser beam,the scattering of light occurs due to variation in size,shape,and morphology.Due to this reason,different types of pathogens show their unique light scattering patterns.The photo-detectors were arranged at the surroundings of the sample at different angles to collect the scattered light.The photodetectors convert the scattered light intensity into a voltage waveform.The waveform features were acquired by using the power spectral characteristics,and the dimensionality of extracted features was reduced by applying minimal-redundancy-maximal-relevance criterion(mRMR).A support vector machine(SVM)classifier was developed by training the selected power spectral features for the classification of three different bacterial microbes.The resulting average identification accuracies of E.faecalis,E.coli and S.aureus were 99%,87%,and 94%,respectively.The ove rall experimental results yield a higher accuracy of 93.6%,indicating that the proposed device has the potential for label-free identification of pathogens with simplicity,rapidity,and cost-effectiveness.展开更多
Collective improvement in the acceptable or desirable accuracy level of breast cancer image-related pattern recognition using various schemes remains challenging.Despite the combination of multiple schemes to achieve ...Collective improvement in the acceptable or desirable accuracy level of breast cancer image-related pattern recognition using various schemes remains challenging.Despite the combination of multiple schemes to achieve superior ultrasound image pattern recognition by reducing the speckle noise,an enhanced technique is not achieved.The purpose of this study is to introduce a features-based fusion scheme based on enhancement uniform-Local Binary Pattern(LBP)and filtered noise reduction.To surmount the above limitations and achieve the aim of the study,a new descriptor that enhances the LBP features based on the new threshold has been proposed.This paper proposes a multi-level fusion scheme for the auto-classification of the static ultrasound images of breast cancer,which was attained in two stages.First,several images were generated from a single image using the pre-processing method.Themedian andWiener filterswere utilized to lessen the speckle noise and enhance the ultrasound image texture.This strategy allowed the extraction of a powerful feature by reducing the overlap between the benign and malignant image classes.Second,the fusion mechanism allowed the production of diverse features from different filtered images.The feasibility of using the LBP-based texture feature to categorize the ultrasound images was demonstrated.The effectiveness of the proposed scheme is tested on 250 ultrasound images comprising 100 and 150 benign and malignant images,respectively.The proposed method achieved very high accuracy(98%),sensitivity(98%),and specificity(99%).As a result,the fusion process that can help achieve a powerful decision based on different features produced from different filtered images improved the results of the new descriptor of LBP features in terms of accuracy,sensitivity,and specificity.展开更多
The acoustic vibration signal of tank is disassembled into the sum of intrinsic mode function （IMF） by multi-resolution empirical mode decomposition （EMD） method. The instantaneous frequency is obtained, and featu...The acoustic vibration signal of tank is disassembled into the sum of intrinsic mode function （IMF） by multi-resolution empirical mode decomposition （EMD） method. The instantaneous frequency is obtained, and feature transformation matrix is figured out by class scatter matrix. Multi- dimensional scale energy vector is mapped into low-dimensional eigenvector, and classification extraction is realized. This method sufficiently separates of different sound target features. The test result indicates that it is effective.展开更多
Feature recognition aims at extracting manufacturing features with geometrical information from solid model and is considered to be an efficient way of changing the interactive NC machining programming mode.Existing r...Feature recognition aims at extracting manufacturing features with geometrical information from solid model and is considered to be an efficient way of changing the interactive NC machining programming mode.Existing recognition methods have some disadvantages in practical applications.They can essentially handle prismatic components with regular shapes and are difficult to recognize the intersecting features and curved surfaces.Besides,the robustness of them is not strong enough.A new feature recognition approach is proposed based on the analysis of aircraft integral panels＇ geometry and machining characteristics.In this approach,the aircraft integral panel is divided into a number of local machining domains.The machining domains are extracted and recognized first by finding the principal face of machining domain and extracting the sides around the principal face.Then the machining domains are divided into various features in terms of the face type.The main sections of the proposed method are presented including the definition,classification and structure of machining domain,the relationship between machining domain and principal face loop,the rules of machining domains recognition,and the algorithm of machining feature recognition.In addition,a robotic feature recognition module is developed for aircraft integral panels and tested with several panels.Test results show that the strategy presented is robust and valid.Features extracted can be post processed and linked to various downstream applications.The approach is able to solve the difficulties in recognizing the aircraft integral panel＇s features and automatic obtaining the machining zone in NC programming,and can be used to further develop the automatic programming of NC machining.展开更多
In this research, a content-based image retrieval （CBIR） system for high resolution satellite images has been developed by using texture features. The proposed approach uses the local binary pattern （LBP） texture ...In this research, a content-based image retrieval （CBIR） system for high resolution satellite images has been developed by using texture features. The proposed approach uses the local binary pattern （LBP） texture feature and a block based scheme. The query and database images are divided into equally sized blocks, from which LBP histograms are extracted. The block histograms are then compared by using the Chi-square distance. Experimental results show that the LBP representation provides a powerful tool for high resolution satellite images （HRSI） retrieval.展开更多
Five-electrode configurations were designed to simulate the distribution inhomogeneity of electric field intensities in the air-insulating medium, and the characteristic data waveforms of partial discharge generated b...Five-electrode configurations were designed to simulate the distribution inhomogeneity of electric field intensities in the air-insulating medium, and the characteristic data waveforms of partial discharge generated by different electrode configurations under the excitation of power frequency AC voltage were carefully collected in this paper. Furthermore, the feature vectors of the corresponding fingerprint, contained in partial discharge data, were extracted by rigorous mathematical algorithms, and the artificial neural network was employed to realize the pattern recognition of partial discharge caused by the inhomogeneity of electric field intensity with different electrode configurations. The results indicate that the J<sub>4</sub> value in the space of 7 feature quantities is 1905.6, and the recognition rate is 100% when the hidden layer neuron of the network is 19. However, the J<sub>5</sub> value of 9 feature quantities is 1589.9, and the purpose of recognition has been achieved when the number of hidden layer neurons of the network is 6. Increasing the number of hidden layer neurons will only waste computing resources. Of course, PD information collection mode, feature quantity selection, optimal feature space composition, network structure and classification algorithm are the key to realizing PD fault intelligence identification.展开更多
In recent years, the accuracy of speech recognition (SR) has been one of the most active areas of research. Despite that SR systems are working reasonably well in quiet conditions, they still suffer severe performance...In recent years, the accuracy of speech recognition (SR) has been one of the most active areas of research. Despite that SR systems are working reasonably well in quiet conditions, they still suffer severe performance degradation in noisy conditions or distorted channels. It is necessary to search for more robust feature extraction methods to gain better performance in adverse conditions. This paper investigates the performance of conventional and new hybrid speech feature extraction algorithms of Mel Frequency Cepstrum Coefficient (MFCC), Linear Prediction Coding Coefficient (LPCC), perceptual linear production (PLP), and RASTA-PLP in noisy conditions through using multivariate Hidden Markov Model (HMM) classifier. The behavior of the proposal system is evaluated using TIDIGIT human voice dataset corpora, recorded from 208 different adult speakers in both training and testing process. The theoretical basis for speech processing and classifier procedures were presented, and the recognition results were obtained based on word recognition rate.展开更多
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The main purpose of nonlinear time series analysis is based on the rebuilding theory of phase space, and to study how to transform the response signal to rebuilt phase space in order to extract dynamic feature information, and to provide effective approach for nonlinear signal analysis and fault diagnosis of nonlinear dynamic system. Now, it has already formed an important offset of nonlinear science. But, traditional method cannot extract chaos features automatically, and it needs man＇s participation in the whole process. A new method is put forward, which can implement auto-extracting of chaos features for nonlinear time series. Firstly, to confirm time delay r by autocorrelation method; Secondly, to compute embedded dimension m and correlation dimension D; Thirdly, to compute the maximum Lyapunov index λmax; Finally, to calculate the chaos degree Dch of Poincare map, and the non-circle degree Dnc and non-order degree Dno of quasi-phase orbit. Chaos features extracting has important meaning to fault diagnosis of nonlinear system based on nonlinear chaos features. Examples show validity of the proposed method.
基金supported by grants fromthe North China University of Technology Research Start-Up Fund(11005136024XN147-14)and(110051360024XN151-97)Guangzhou Development Zone Science and Technology Project(2023GH02)+4 种基金the National Key R&D Program of China(2021YFE0201100 and 2022YFA1103401 to Juntao Gao)National Natural Science Foundation of China(981890991 to Juntao Gao)Beijing Municipal Natural Science Foundation(Z200021 to Juntao Gao)CAS Interdisciplinary Innovation Team(JCTD-2020-04 to Juntao Gao)0032/2022/A,by Macao FDCT,and MYRG2022-00271-FST.
文摘Hematoxylin and Eosin(H&E)images,popularly used in the field of digital pathology,often pose challenges due to their limited color richness,hindering the differentiation of subtle cell features crucial for accurate classification.Enhancing the visibility of these elusive cell features helps train robust deep-learning models.However,the selection and application of image processing techniques for such enhancement have not been systematically explored in the research community.To address this challenge,we introduce Salient Features Guided Augmentation(SFGA),an approach that strategically integrates machine learning and image processing.SFGA utilizes machine learning algorithms to identify crucial features within cell images,subsequently mapping these features to appropriate image processing techniques to enhance training images.By emphasizing salient features and aligning them with corresponding image processing methods,SFGA is designed to enhance the discriminating power of deep learning models in cell classification tasks.Our research undertakes a series of experiments,each exploring the performance of different datasets and data enhancement techniques in classifying cell types,highlighting the significance of data quality and enhancement in mitigating overfitting and distinguishing cell characteristics.Specifically,SFGA focuses on identifying tumor cells from tissue for extranodal extension detection,with the SFGA-enhanced dataset showing notable advantages in accuracy.We conducted a preliminary study of five experiments,among which the accuracy of the pleomorphism experiment improved significantly from 50.81%to 95.15%.The accuracy of the other four experiments also increased,with improvements ranging from 3 to 43 percentage points.Our preliminary study shows the possibilities to enhance the diagnostic accuracy of deep learning models and proposes a systematic approach that could enhance cancer diagnosis,contributing as a first step in using SFGA in medical image enhancement.
基金supported in part by the Fundamental Research Funds for the Central Universities(2024JBMC031)the OpenFund of Advanced Cryptography and System Security Key Laboratory of Sichuan Province(No.SKLACSS-202312)+2 种基金the CCF-NSFOCUS Open Fund,the National Natural Science Foundation of China(Grant Nos.62202042,U20A6003,62076146,62021002,U19A2062,62127803,U1911401 and 6212780016)the Fundamental Research Funds for the Central Universities,JLU,the Industrial Technology Infrastructure Public Service Platform Project‘Public Service Platform for Urban Rail Transit Equipment Signal System Testing and Safety Evaluation’(No.2022-233-225)Ministry of Industry and Information Technology of China.
文摘Advanced Persistent Threats(APTs)pose significant challenges to detect due to their“low-and-slow”attack patterns and frequent use of zero-day vulnerabilities.Within this task,the extraction of long-term features is often crucial.In this work,we propose a novel end-to-end APT detection framework named Long-Term Feature Association Provenance Graph Detector(LT-ProveGD).Specifically,LT-ProveGD encodes contextual information of the dynamic provenance graph while preserving the topological information with space efficiency.To combat“low-and-slow”attacks,LT-ProveGD develops an autoencoder with an integrated multi-head attention mechanism to extract long-term dependencies within the encoded representations.Furthermore,to facilitate the detection of previously unknown attacks,we leverage Jenks’natural breaks methodology,enabling detection without relying on specific attack information.By conducting extensive experiments on five widely used datasets with state-of-the-art attack detection methods,we demonstrate the superior effectiveness of LT-ProveGD.
基金the financial support from Natural Science Foundation of Gansu Province(Nos.22JR5RA217,22JR5RA216)Lanzhou Science and Technology Program(No.2022-2-111)+1 种基金Lanzhou University of Arts and Sciences School Innovation Fund Project(No.XJ2022000103)Lanzhou College of Arts and Sciences 2023 Talent Cultivation Quality Improvement Project(No.2023-ZL-jxzz-03)。
文摘Considering that the algorithm accuracy of the traditional sparse representation models is not high under the influence of multiple complex environmental factors,this study focuses on the improvement of feature extraction and model construction.Firstly,the convolutional neural network(CNN)features of the face are extracted by the trained deep learning network.Next,the steady-state and dynamic classifiers for face recognition are constructed based on the CNN features and Haar features respectively,with two-stage sparse representation introduced in the process of constructing the steady-state classifier and the feature templates with high reliability are dynamically selected as alternative templates from the sparse representation template dictionary constructed using the CNN features.Finally,the results of face recognition are given based on the classification results of the steady-state classifier and the dynamic classifier together.Based on this,the feature weights of the steady-state classifier template are adjusted in real time and the dictionary set is dynamically updated to reduce the probability of irrelevant features entering the dictionary set.The average recognition accuracy of this method is 94.45%on the CMU PIE face database and 96.58%on the AR face database,which is significantly improved compared with that of the traditional face recognition methods.
基金funded by the Jiangxi SASAC Science and Technology Innovation Special Project and the Key Technology Research and Application Promotion of Highway Overload Digital Solution.
文摘In response to the challenges in highway pavement distress detection,such as multiple defect categories,difficulties in feature extraction for different damage types,and slow identification speeds,this paper proposes an enhanced pavement crack detection model named Star-YOLO11.This improved algorithm modifies the YOLO11 architecture by substituting the original C3k2 backbone network with a Star-s50 feature extraction network.The enhanced structure adjusts the number of stacked layers in the StarBlock module to optimize detection accuracy and improve model efficiency.To enhance the accuracy of pavement crack detection and improve model efficiency,three key modifications to the YOLO11 architecture are proposed.Firstly,the original C3k2 backbone is replaced with a StarBlock-based structure,forming the Star-s50 feature extraction backbone network.This lightweight redesign reduces computational complexity while maintaining detection precision.Secondly,to address the inefficiency of the original Partial Self-attention(PSA)mechanism in capturing localized crack features,the convolutional prior-aware Channel Prior Convolutional Attention(CPCA)mechanism is integrated into the channel dimension,creating a hybrid CPC-C2PSA attention structure.Thirdly,the original neck structure is upgraded to a Star Multi-Branch Auxiliary Feature Pyramid Network(SMAFPN)based on the Multi-Branch Auxiliary Feature Pyramid Network architecture,which adaptively fuses high-level semantic and low-level spatial information through Star-s50 connections and C3k2 extraction blocks.Additionally,a composite dataset augmentation strategy combining traditional and advanced augmentation techniques is developed.This strategy is validated on a specialized pavement dataset containing five distinct crack categories for comprehensive training and evaluation.Experimental results indicate that the proposed Star-YOLO11 achieves an accuracy of 89.9%(3.5%higher than the baseline),a mean average precision(mAP)of 90.3%(+2.6%),and an F1-score of 85.8%(+0.5%),while reducing the model size by 18.8%and reaching a frame rate of 225.73 frames per second(FPS)for real-time detection.It shows potential for lightweight deployment in pavement crack detection tasks.
基金The work described in this paper was fully supported by a grant from Hong Kong Metropolitan University(RIF/2021/05).
文摘Parkinson’s disease(PD)is a debilitating neurological disorder affecting over 10 million people worldwide.PD classification models using voice signals as input are common in the literature.It is believed that using deep learning algorithms further enhances performance;nevertheless,it is challenging due to the nature of small-scale and imbalanced PD datasets.This paper proposed a convolutional neural network-based deep support vector machine(CNN-DSVM)to automate the feature extraction process using CNN and extend the conventional SVM to a DSVM for better classification performance in small-scale PD datasets.A customized kernel function reduces the impact of biased classification towards the majority class(healthy candidates in our consideration).An improved generative adversarial network(IGAN)was designed to generate additional training data to enhance the model’s performance.For performance evaluation,the proposed algorithm achieves a sensitivity of 97.6%and a specificity of 97.3%.The performance comparison is evaluated from five perspectives,including comparisons with different data generation algorithms,feature extraction techniques,kernel functions,and existing works.Results reveal the effectiveness of the IGAN algorithm,which improves the sensitivity and specificity by 4.05%–4.72%and 4.96%–5.86%,respectively;and the effectiveness of the CNN-DSVM algorithm,which improves the sensitivity by 1.24%–57.4%and specificity by 1.04%–163%and reduces biased detection towards the majority class.The ablation experiments confirm the effectiveness of individual components.Two future research directions have also been suggested.
基金Supported by National Science and Technology Support Program of China(Grant No.2015BAF07B04)
文摘Numerous vibration-based techniques are rarely used in diesel engines fault diagnosis in a direct way, due to the surface vibration signals of diesel engines with the complex non-stationary and nonlinear time-varying fea- tures. To investigate the fault diagnosis of diesel engines, fractal correlation dimension, wavelet energy and entropy as features reflecting the diesel engine fault fractal and energy characteristics are extracted from the decomposed signals through analyzing vibration acceleration signals derived from the cylinder head in seven different states of valve train. An intelligent fault detector FastICA-SVM is applied for diesel engine fault diagnosis and classification. The results demonstrate that FastlCA-SVM achieves higher classification accuracy and makes better general- ization performance in small samples recognition. Besides, the fractal correlation dimension and wavelet energy and entropy as the special features of diesel engine vibration signal are considered as input vectors of classifier FastlCA- SVM and could produce the excellent classification results. The proposed methodology improves the accuracy of fea- ture extraction and the fault diagnosis of diesel engines.
基金supported by the National Natural Science Foundation of China (60872108)the Postdoctoral Science Foundation of China(200902411+3 种基金20080430903)Heilongjiang Postdoctoral Financial Assistance (LBH-Z08129)the Scientific and Technological Creative Talents Special Research Foundation of Harbin Municipality (2008RFQXG030)Central University Basic Research Professional Expenses Special Fund Project
文摘A novel class of periodically changing features hidden in radar pulse sequence environment,named G features,is proposed.Combining fractal theory and Hilbert-Huang transform,the features are extracted using changing characteristics of pulse parameters in radar emitter signals.The features can be applied in modern complex electronic warfare environment to address the issue of signal sorting when radar emitter pulse signal parameters severely or even completely overlap.Experiment results show that the proposed feature class and feature extraction method can discriminate periodically changing pulse sequence signal sorting features from radar pulse signal flow with complex variant features,therefore provide a new methodology for signal sorting.
基金supported by Tianjin Sci-tech Planning Projects (14RCGFGX00846)Natural Science Foundation of Hebei Province (F2015202239)+1 种基金Tianjin Sci-tech Planning Projects (15ZCZDNC00130)Science and Technology Research Project of Hebei Province (Z2015044)
基金Supported by the National Natural Science Foundation of China(51174091,61364013,61164013)Earlier Research Project of the State Key Development Program for Basic Research of China(2014CB360502)
文摘For measurement of component content in the extraction and separation process of praseodymium/neodymium(Pr/Nd), a soft measurement method was proposed based on modeling of ion color features, which is suitable for fast estimation of component content in production field. Feature analysis on images of the solution is conducted,which are captured from Pr/Nd extraction/separation field. H/S components in the HSI color space are selected as model inputs, so as to establish the least squares support vector machine(LSSVM) model for Nd(Pr) content,while the model parameters are determined with the GA algorithm. To improve the adaptability of the model,the adaptive iteration algorithm is used to correct parameters of the LSSVM model, on the basis of model correction strategy and new sample data. Using the field data collected from rare earth extraction production, predictive methods for component content and comparisons are given. The results indicate that the proposed method presents good adaptability and high prediction precision, so it is applicable to the fast detection of element content in the rare earth extraction.
基金Acknowledgements This paper was supported by the coUabomtive Research Project SEV under Cant No. 01100474 between Beijing University of Posts and Telecorrrcnications and France Telecom R＆D Beijing the National Natural Science Foundation of China under Cant No. 90920001 the Caduate Innovation Fund of SICE, BUPT, 2011.
文摘The rapid growth of multimedia content necessitates powerful technologies to filter, classify, index and retrieve video documents more efficiently. However, the essential bottleneck of image and video analysis is the problem of semantic gap that low level features extracted by computers always fail to coincide with high-level concepts interpreted by humans. In this paper, we present a generic scheme for the detection video semantic concepts based on multiple visual features machine learning. Various global and local low-level visual features are systelrtically investigated, and kernelbased learning method equips the concept detection system to explore the potential of these features. Then we combine the different features and sub-systen on both classifier-level and kernel-level fusion that contribute to a more robust system Our proposed system is tested on the TRECVID dataset. The resulted Mean Average Precision （MAP） score is rmch better than the benchmark perforrmnce, which proves that our concepts detection engine develops a generic model and perforrrs well on both object and scene type concepts.
基金Guangdong Provincial Natural Science Foundation of China(Grant No.2020B1515120006)Guangdong Innovation Team(Grant Nos.2020KCXTD015,2022KCXTD029)Guangdong Universities New Information Field(Grant No.2021ZDZX1057).
文摘Feature information extraction is one of the key steps in prognostics and health management of rotating machinery.In the present study,an investigation about the feasibility of a methodology based on generalized S transform(GST)and singular value decomposition(SVD)methods for feature extraction in rolling bearing,due to local damage under variable conditions,is conducted.The technique adopts the GST method,following the time-frequency analysis,to transform a raw fault signal of the rolling bearing into a two-dimensional complex matrix.And then,the SVD method is performed to decompose the matrix to obtain the feature vectors.By this procedure it is possible to obtain the fault feature information of rolling bearing under different speeds and different loads.In order to streamline the feature parameters of the feature vectors to train more uncomplicated models,the principal component analysis(PCA)subsequently performed.The particle swarm optimization-support vector machine(PSO-SVM)model is used to identify and classify the different fault states of rolling bearing.Furthermore,in order to highlight the superiority of the proposed method some comparisons are conducted with the conventional methods.The obtained results show that the proposed method can effectively extract fault features of the rolling bearing under variable conditions.
基金the National Key Special Science Program(No.2017YFA0205301)the National Natural Science Foundation of China(Nos.61527806,61971187,61901168,81902153,61971216 and 61401217)+3 种基金the Clinical Advanced Technology of Social Development Projects in Jiangsu Province(No.BE2018695)Natural Science Foundation of Jiangsu Province(No.BK20140900)key project supported by Medical Science and Technology Development Foundation,Nanjing Department of Health(Nos.ZKX18029 and ZKX18016)the joint fund of Southeast University and Nanjing Medical University。
文摘The rapid identification of pathogens is crucial in controlling the food quality and safety.The proposed system for the rapid and label-free identification of pathogens is based on the principle of laser scattering from the bacterial microbes.The clinical prototype consists of three parts:the laser beam,photodetectors,and the data acquisition system.The bacterial testing sample was mixed with 10 mL distilled water and placed inside the machine chamber.When the bacterial microbes pass by the laser beam,the scattering of light occurs due to variation in size,shape,and morphology.Due to this reason,different types of pathogens show their unique light scattering patterns.The photo-detectors were arranged at the surroundings of the sample at different angles to collect the scattered light.The photodetectors convert the scattered light intensity into a voltage waveform.The waveform features were acquired by using the power spectral characteristics,and the dimensionality of extracted features was reduced by applying minimal-redundancy-maximal-relevance criterion(mRMR).A support vector machine(SVM)classifier was developed by training the selected power spectral features for the classification of three different bacterial microbes.The resulting average identification accuracies of E.faecalis,E.coli and S.aureus were 99%,87%,and 94%,respectively.The ove rall experimental results yield a higher accuracy of 93.6%,indicating that the proposed device has the potential for label-free identification of pathogens with simplicity,rapidity,and cost-effectiveness.
基金This research received funding from Duhok Polytechnic University.
文摘Collective improvement in the acceptable or desirable accuracy level of breast cancer image-related pattern recognition using various schemes remains challenging.Despite the combination of multiple schemes to achieve superior ultrasound image pattern recognition by reducing the speckle noise,an enhanced technique is not achieved.The purpose of this study is to introduce a features-based fusion scheme based on enhancement uniform-Local Binary Pattern(LBP)and filtered noise reduction.To surmount the above limitations and achieve the aim of the study,a new descriptor that enhances the LBP features based on the new threshold has been proposed.This paper proposes a multi-level fusion scheme for the auto-classification of the static ultrasound images of breast cancer,which was attained in two stages.First,several images were generated from a single image using the pre-processing method.Themedian andWiener filterswere utilized to lessen the speckle noise and enhance the ultrasound image texture.This strategy allowed the extraction of a powerful feature by reducing the overlap between the benign and malignant image classes.Second,the fusion mechanism allowed the production of diverse features from different filtered images.The feasibility of using the LBP-based texture feature to categorize the ultrasound images was demonstrated.The effectiveness of the proposed scheme is tested on 250 ultrasound images comprising 100 and 150 benign and malignant images,respectively.The proposed method achieved very high accuracy(98%),sensitivity(98%),and specificity(99%).As a result,the fusion process that can help achieve a powerful decision based on different features produced from different filtered images improved the results of the new descriptor of LBP features in terms of accuracy,sensitivity,and specificity.
文摘The acoustic vibration signal of tank is disassembled into the sum of intrinsic mode function （IMF） by multi-resolution empirical mode decomposition （EMD） method. The instantaneous frequency is obtained, and feature transformation matrix is figured out by class scatter matrix. Multi- dimensional scale energy vector is mapped into low-dimensional eigenvector, and classification extraction is realized. This method sufficiently separates of different sound target features. The test result indicates that it is effective.
文摘Feature recognition aims at extracting manufacturing features with geometrical information from solid model and is considered to be an efficient way of changing the interactive NC machining programming mode.Existing recognition methods have some disadvantages in practical applications.They can essentially handle prismatic components with regular shapes and are difficult to recognize the intersecting features and curved surfaces.Besides,the robustness of them is not strong enough.A new feature recognition approach is proposed based on the analysis of aircraft integral panels＇ geometry and machining characteristics.In this approach,the aircraft integral panel is divided into a number of local machining domains.The machining domains are extracted and recognized first by finding the principal face of machining domain and extracting the sides around the principal face.Then the machining domains are divided into various features in terms of the face type.The main sections of the proposed method are presented including the definition,classification and structure of machining domain,the relationship between machining domain and principal face loop,the rules of machining domains recognition,and the algorithm of machining feature recognition.In addition,a robotic feature recognition module is developed for aircraft integral panels and tested with several panels.Test results show that the strategy presented is robust and valid.Features extracted can be post processed and linked to various downstream applications.The approach is able to solve the difficulties in recognizing the aircraft integral panel＇s features and automatic obtaining the machining zone in NC programming,and can be used to further develop the automatic programming of NC machining.
文摘In this research, a content-based image retrieval （CBIR） system for high resolution satellite images has been developed by using texture features. The proposed approach uses the local binary pattern （LBP） texture feature and a block based scheme. The query and database images are divided into equally sized blocks, from which LBP histograms are extracted. The block histograms are then compared by using the Chi-square distance. Experimental results show that the LBP representation provides a powerful tool for high resolution satellite images （HRSI） retrieval.
文摘Five-electrode configurations were designed to simulate the distribution inhomogeneity of electric field intensities in the air-insulating medium, and the characteristic data waveforms of partial discharge generated by different electrode configurations under the excitation of power frequency AC voltage were carefully collected in this paper. Furthermore, the feature vectors of the corresponding fingerprint, contained in partial discharge data, were extracted by rigorous mathematical algorithms, and the artificial neural network was employed to realize the pattern recognition of partial discharge caused by the inhomogeneity of electric field intensity with different electrode configurations. The results indicate that the J<sub>4</sub> value in the space of 7 feature quantities is 1905.6, and the recognition rate is 100% when the hidden layer neuron of the network is 19. However, the J<sub>5</sub> value of 9 feature quantities is 1589.9, and the purpose of recognition has been achieved when the number of hidden layer neurons of the network is 6. Increasing the number of hidden layer neurons will only waste computing resources. Of course, PD information collection mode, feature quantity selection, optimal feature space composition, network structure and classification algorithm are the key to realizing PD fault intelligence identification.
文摘In recent years, the accuracy of speech recognition (SR) has been one of the most active areas of research. Despite that SR systems are working reasonably well in quiet conditions, they still suffer severe performance degradation in noisy conditions or distorted channels. It is necessary to search for more robust feature extraction methods to gain better performance in adverse conditions. This paper investigates the performance of conventional and new hybrid speech feature extraction algorithms of Mel Frequency Cepstrum Coefficient (MFCC), Linear Prediction Coding Coefficient (LPCC), perceptual linear production (PLP), and RASTA-PLP in noisy conditions through using multivariate Hidden Markov Model (HMM) classifier. The behavior of the proposal system is evaluated using TIDIGIT human voice dataset corpora, recorded from 208 different adult speakers in both training and testing process. The theoretical basis for speech processing and classifier procedures were presented, and the recognition results were obtained based on word recognition rate.
