This paper extends the criterion of the misclassification ratio of discriminant model and presents a new selection method of discriminant model.For selecting the discriminant model,this method establishes the rule of ...This paper extends the criterion of the misclassification ratio of discriminant model and presents a new selection method of discriminant model.For selecting the discriminant model,this method establishes the rule of misclassification degree ratio through misclassification ratio of the discriminant model and misclassification degree of the samples.To test the effect of this method,this work uses seven UCI data sets.Numerical experiments on these examples indicate that this method has certain rationality and has a better effect to select a discriminant model.展开更多
Autonomous intelligence plays a significant role in aviation security.Since most aviation accidents occur in the take-off and landing stage,accurate tracking of moving object in airport apron will be a vital approach ...Autonomous intelligence plays a significant role in aviation security.Since most aviation accidents occur in the take-off and landing stage,accurate tracking of moving object in airport apron will be a vital approach to ensure the operation of the aircraft safely.In this study,an adaptive object tracking method based on a discriminant is proposed in multi-camera panorama surveillance of large-scale airport apron.Firstly,based on channels of color histogram,the pre-estimated object probability map is employed to reduce searching computation,and the optimization of the disturbance suppression options can make good resistance to similar areas around the object.Then the object score of probability map is obtained by the sliding window,and the candidate window with the highest probability map score is selected as the new object center.Thirdly,according to the new object location,the probability map is updated,the scale estimation function is adjusted to the size of real object.From qualitative and quantitative analysis,the comparison experiments are verified in representative video sequences,and our approach outperforms typical methods,such as distraction-aware online tracking,mean shift,variance ratio,and adaptive colour attributes.展开更多
Nuclear security usually requires the simultaneous detection of neutrons and gamma rays.With the development of crystalline materials in recent years,Cs2LiLaBr6(CLLB)dual-readout detectors have attracted extensive att...Nuclear security usually requires the simultaneous detection of neutrons and gamma rays.With the development of crystalline materials in recent years,Cs2LiLaBr6(CLLB)dual-readout detectors have attracted extensive attention from researchers,where real-time neutron/gamma pulse discrimination is the critical factor among detector performance parameters.This study investigated the discrimination performance of the charge comparison,amplitude comparison,time comparison,and pulse gradient_(m)ethods and the effects of a Sallen–Key filter on their performance.Experimental results show that the figure of merit(FOM)of all four methods is improved by proper filtering.Among them,the charge comparison method exhibits excellent noise resistance;moreover,it is the most_(s)uitable method of real-time discrimination for CLLB detectors.However,its discrimination performance depends on the parameters t_(s),t_(m),and t_(e).When t_(s)corresponds to the moment at which the pulse is at 10%of its peak value,t_(e)requires a delay of only 640–740 ns compared to t_(s),at which time the potentially optimal FOM of the charge comparison method at 3.1–3.3 MeV is greater than 1.46.The FOM obtained using the t_(m)value calculated by a proposed maximized discrimination difference model(MDDM)and the potentially optimal FOM differ by less than 3.9%,indicating that the model can provide good guidance for parameter selection in the charge comparison method.展开更多
Predictive models for assessing the risk of developing lung cancers can help identify high-risk individuals with the aim of recommending further screening and early intervention.To facilitate pre-hospital self-assessm...Predictive models for assessing the risk of developing lung cancers can help identify high-risk individuals with the aim of recommending further screening and early intervention.To facilitate pre-hospital self-assessments,some studies have exploited predictive models trained on non-clinical data(e.g.,smoking status and family history).The performance of these models is limited due to not considering clinical data(e.g.,blood test and medical imaging results).Deep learning has shown the potential in processing complex data that combine both clinical and non-clinical information.However,predicting lung cancers remains difficult due to the severe lack of positive samples among follow-ups.To tackle this problem,this paper presents a generative-discriminative framework for improving the ability of deep learning models to generalize.According to the proposed framework,two nonlinear generative models,one based on the generative adversarial network and another on the variational autoencoder,are used to synthesize auxiliary positive samples for the training set.Then,several discriminative models,including a deep neural network(DNN),are used to assess the lung cancer risk based on a comprehensive list of risk factors.The framework was evaluated on over 55000 subjects questioned between January 2014 and December 2017,with 699 subjects being clinically diagnosed with lung cancer between January 2014 and August 2019.According to the results,the best performing predictive model built using the proposed framework was based on DNN.It achieved an average sensitivity of 76.54%and an area under the curve of 69.24%in distinguishing between the cases of lung cancer and normal cases on test sets.展开更多
Near-infrared reflectance spectroscopy (NIRS) was applied to classify grape wines of different geographical origins (Changli, Huailai, and Yantai, China). Near infrared (NIR) spectra were collected in transmission mod...Near-infrared reflectance spectroscopy (NIRS) was applied to classify grape wines of different geographical origins (Changli, Huailai, and Yantai, China). Near infrared (NIR) spectra were collected in transmission mode in the wavelength range of 800-2500 nm. Wines (n=90) were randomly split into two sets, calibration set (n=54) and validation set (n=36). Discriminant analysis models were developed using BP neural network and discriminant partial least-squares discriminant analysis (PLS-DA). The prediction performance of calibration models in different wavelength range was also investigated. BP neural network models and PLS-DA models correctly classified 100% of the wines in calibration set. When used to predict wines in validation set, BP neural network models correctly classified 100%, 81.8%, and 90.9% of the wines from Changli, Huailai, and Yantai respectively, and PLS-DA models correctly classified 100% of all samples. The results demonstrated that NIRS could be used to discriminate Chinese grape wines as a rapid and reliable method.展开更多
This manuscript presents a stochastic model updating method, taking both uncertainties in models and variability in testing into account. The updated finite element(FE) models obtained through the proposed technique...This manuscript presents a stochastic model updating method, taking both uncertainties in models and variability in testing into account. The updated finite element(FE) models obtained through the proposed technique can aid in the analysis and design of structural systems. The authors developed a stochastic model updating method integrating distance discrimination analysis(DDA) and advanced Monte Carlo(MC) technique to(1) enable more efficient MC by using a response surface model,(2) calibrate parameters with an iterative test-analysis correlation based upon DDA, and(3) utilize and compare different distance functions as correlation metrics. Using DDA, the influence of distance functions on model updating results is analyzed. The proposed stochastic method makes it possible to obtain a precise model updating outcome with acceptable calculation cost. The stochastic method is demonstrated on a helicopter case study updated using both Euclidian and Mahalanobis distance metrics. It is observed that the selected distance function influences the iterative calibration process and thus, the calibration outcome, indicating that an integration of different metrics might yield improved results.展开更多
Traditional hand-crafted features for representing local image patches are evolving into current data-driven and learning-based image feature, but learning a robust and discriminative descriptor which is capable of co...Traditional hand-crafted features for representing local image patches are evolving into current data-driven and learning-based image feature, but learning a robust and discriminative descriptor which is capable of controlling various patch-level computer vision tasks is still an open problem. In this work, we propose a novel deep convolutional neural network(CNN) to learn local feature descriptors. We utilize the quadruplets with positive and negative training samples, together with a constraint to restrict the intra-class variance, to learn good discriminative CNN representations. Compared with previous works, our model reduces the overlap in feature space between corresponding and non-corresponding patch pairs, and mitigates margin varying problem caused by commonly used triplet loss. We demonstrate that our method achieves better embedding result than some latest works, like PN-Net and TN-TG, on benchmark dataset.展开更多
Background: Differentiating intracerebral hemorrhage (ICH) from cerebral infarction as early as possible is vital tbr the timely initiation of different treatments. This study developed an applicable model for the ...Background: Differentiating intracerebral hemorrhage (ICH) from cerebral infarction as early as possible is vital tbr the timely initiation of different treatments. This study developed an applicable model for the ambulance system to differentiate stroke subtypes. Methods: From 26,163 patients initially screened over 4 years, this study comprised 1989 consecutive patients with potential first-ever acute stroke with sudden onset of the focal neurological deficit, conscious or not, and given ambulance transport for admission to two county hospitals in Yutian County of Hebei Province. All the patients underwent cranial computed tomography (CT) or magnetic resonance imaging to confirm the final diagnosis based on stroke criteria. Correlation with stroke subtype clinical features was calculated and Bayes' discriminant model was applied to discriminate stroke subtypes. Results: Among the 1989 patients, 797,689, 109, and 394 received diagnoses of cerebral infarction, ICH, subarachnoid hemorrhage, and other forms of nonstroke, respectively. A history of atrial fibrillation, vomiting, and diabetes mellitus were associated with cerebral infarction, while vomiting, systolic blood pressure _〉180 mmHg, and age 〈65 years were more typical of ICH. For noncomatose stroke patients, Bayes' discriminant model for stroke subtype yielded a combination of multiple items that provided 72.3% agreement in the test model and 79.3% in the validation model; for comatose patients, corresponding agreement rates were 75.4% and 73.5%. Conclusions: The model herein presented, with multiple parameters, can predict stroke subtypes with acceptable sensitivity and specificity before CT scanning, either in alert or comatose patients. This may facilitate prehospital management for patients with stroke.展开更多
Place of origin has an important influence on walnut quality and commercial value,which results in the requirement of rapid geographical traceability method.Thus,a method for geographical origin identification of waln...Place of origin has an important influence on walnut quality and commercial value,which results in the requirement of rapid geographical traceability method.Thus,a method for geographical origin identification of walnuts on the basis of nutritional quality of walnuts from China was conducted.The concentrations of 43 phytochemical components were analyzed in walnut samples from five different walnut-producing regions of China.Based on 14 chemical markers selected by the Random Forest method from these phytochemical components,a new discriminant model for geographical origin was built,with the corresponding correct classification rate of 99.3%.In addition,the quantitative quality differences of walnuts from five regions were analyzed,with values of 0.17–1.43.Moreover,the top three chemical markers for the geographical origin discriminant analysis were Mo,V,and stearic acid,with contribution rates of 26.8%,18.9%,and 10.9%,respectively.This study provides a potentially viable method for application in food authentication.展开更多
With the rapid growth of power systems measurements in terms of size and complexity,discovering statistical patterns for a large variety of real-world applications such as renewable energy prediction,demand response,e...With the rapid growth of power systems measurements in terms of size and complexity,discovering statistical patterns for a large variety of real-world applications such as renewable energy prediction,demand response,energy disaggregation,and state estimation is considered a crucial challenge.In recent years,deep learning has emerged as a novel class of machine learning algorithms that represents power systems data via a large hypothesis space that leads to the state-of-the-art performance compared to most recent data-driven algorithms.This study explores the theoretical advantages of deep representation learning in power systems research.We review deep learning methodologies presented and applied in a wide range of supervised,unsupervised,and semi-supervised applications as well as reinforcement learning tasks.We discuss various settings of problems solved by discriminative deep models including stacked autoencoders and convolutional neural networks as well as generative deep architectures such as deep belief networks and variational autoencoders.The theoretical and experimental analysis of deep neural networks in this study motivates longterm research on optimizing this cutting-edge class of models to achieve significant improvements in the future power systems research.展开更多
An important component of a spoken term detection (STD) system involves estimating confidence measures of hypothesised detections.A potential problem of the widely used lattice-based confidence estimation,however,is...An important component of a spoken term detection (STD) system involves estimating confidence measures of hypothesised detections.A potential problem of the widely used lattice-based confidence estimation,however,is that the confidence scores are treated uniformly for all search terms,regardless of how much they may differ in terms of phonetic or linguistic properties.This problem is particularly evident for out-of-vocabulary (OOV) terms which tend to exhibit high intra-term diversity.To address the impact of term diversity on confidence measures,we propose in this work a term-dependent normalisation technique which compensates for term diversity in confidence estimation.We first derive an evaluation-metric-oriented normalisation that optimises the evaluation metric by compensating for the diverse occurrence rates among terms,and then propose a linear bias compensation and a discriminative compensation to deal with the bias problem that is inherent in lattice-based confidence measurement and from which the Term Specific Threshold (TST) approach suffers.We tested the proposed technique on speech data from the multi-party meeting domain with two state-ofthe-art STD systems based on phonemes and words respectively.The experimental results demonstrate that the confidence normalisation approach leads to a significant performance improvement in STD,particularly for OOV terms with phonemebased systems.展开更多
The performance of scene classification of satellite images strongly relies on the discriminative power of the low-level and mid-level feature representation. This paper presents a novel approach, named multi-level ma...The performance of scene classification of satellite images strongly relies on the discriminative power of the low-level and mid-level feature representation. This paper presents a novel approach, named multi-level max-margin analysis (M 3 DA) for semantic classification for high-resolution satellite images. In our M 3 DA model, the maximum entropy discrimination latent Dirichlet allocation (MedLDA) model is applied to learn the topic-level features first, and then based on a bag-of-words repre- sentation of low-level local image features, the large margin nearest neighbor (LMNN) classifier is used to optimize a multiple soft label composed of word-level features (generated by SVM classifier) and topic-level features. The categorization performances on 21-class land-use dataset have demonstrated that the proposed model in multi-level max-margin scheme can distinguish different categories of land-use scenes reasonably.展开更多
基金Supported by the National Natural Science Foundation of China(52070119)Key Laboratory of Financial Mathematics of Fujian Province University(Putian University)(JR201801).
文摘This paper extends the criterion of the misclassification ratio of discriminant model and presents a new selection method of discriminant model.For selecting the discriminant model,this method establishes the rule of misclassification degree ratio through misclassification ratio of the discriminant model and misclassification degree of the samples.To test the effect of this method,this work uses seven UCI data sets.Numerical experiments on these examples indicate that this method has certain rationality and has a better effect to select a discriminant model.
基金This work was supported in part by the National Natural Science Foundation of China under Grant Nos.61806028,61672437 and 61702428Sichuan Sci-ence and Technology Program under Grant Nos.2018GZ0245,21ZDYF2484,18ZDYF3269,2021YFN0104,2021YFN0104,21GJHZ0061,21ZDYF3629,2021YFG0295,2021YFG0133,21ZDYF2907,21ZDYF0418,21YYJC1827,21ZDYF3537,21ZDYF3598,2019YJ0356the Chinese Scholarship Council under Grant Nos.202008510036,201908515022。
文摘Autonomous intelligence plays a significant role in aviation security.Since most aviation accidents occur in the take-off and landing stage,accurate tracking of moving object in airport apron will be a vital approach to ensure the operation of the aircraft safely.In this study,an adaptive object tracking method based on a discriminant is proposed in multi-camera panorama surveillance of large-scale airport apron.Firstly,based on channels of color histogram,the pre-estimated object probability map is employed to reduce searching computation,and the optimization of the disturbance suppression options can make good resistance to similar areas around the object.Then the object score of probability map is obtained by the sliding window,and the candidate window with the highest probability map score is selected as the new object center.Thirdly,according to the new object location,the probability map is updated,the scale estimation function is adjusted to the size of real object.From qualitative and quantitative analysis,the comparison experiments are verified in representative video sequences,and our approach outperforms typical methods,such as distraction-aware online tracking,mean shift,variance ratio,and adaptive colour attributes.
基金supported by cooperation projects between an enterprise(CNPE)and a research institute(ASIPP)(Y15HX16706).
文摘Nuclear security usually requires the simultaneous detection of neutrons and gamma rays.With the development of crystalline materials in recent years,Cs2LiLaBr6(CLLB)dual-readout detectors have attracted extensive attention from researchers,where real-time neutron/gamma pulse discrimination is the critical factor among detector performance parameters.This study investigated the discrimination performance of the charge comparison,amplitude comparison,time comparison,and pulse gradient_(m)ethods and the effects of a Sallen–Key filter on their performance.Experimental results show that the figure of merit(FOM)of all four methods is improved by proper filtering.Among them,the charge comparison method exhibits excellent noise resistance;moreover,it is the most_(s)uitable method of real-time discrimination for CLLB detectors.However,its discrimination performance depends on the parameters t_(s),t_(m),and t_(e).When t_(s)corresponds to the moment at which the pulse is at 10%of its peak value,t_(e)requires a delay of only 640–740 ns compared to t_(s),at which time the potentially optimal FOM of the charge comparison method at 3.1–3.3 MeV is greater than 1.46.The FOM obtained using the t_(m)value calculated by a proposed maximized discrimination difference model(MDDM)and the potentially optimal FOM differ by less than 3.9%,indicating that the model can provide good guidance for parameter selection in the charge comparison method.
基金supported in part by Zhejiang Provincial Natural Science Foundation of China(LQ20F030013)Research Foundation of Hwa Mei Hospital,University of Chinese Academy of Sciences(2020HMZD22)+1 种基金Ningbo Public Service Technology Foundation(202002N3181)Medical Scientific Research Foundation of Zhejiang Province(2021431314)。
文摘Predictive models for assessing the risk of developing lung cancers can help identify high-risk individuals with the aim of recommending further screening and early intervention.To facilitate pre-hospital self-assessments,some studies have exploited predictive models trained on non-clinical data(e.g.,smoking status and family history).The performance of these models is limited due to not considering clinical data(e.g.,blood test and medical imaging results).Deep learning has shown the potential in processing complex data that combine both clinical and non-clinical information.However,predicting lung cancers remains difficult due to the severe lack of positive samples among follow-ups.To tackle this problem,this paper presents a generative-discriminative framework for improving the ability of deep learning models to generalize.According to the proposed framework,two nonlinear generative models,one based on the generative adversarial network and another on the variational autoencoder,are used to synthesize auxiliary positive samples for the training set.Then,several discriminative models,including a deep neural network(DNN),are used to assess the lung cancer risk based on a comprehensive list of risk factors.The framework was evaluated on over 55000 subjects questioned between January 2014 and December 2017,with 699 subjects being clinically diagnosed with lung cancer between January 2014 and August 2019.According to the results,the best performing predictive model built using the proposed framework was based on DNN.It achieved an average sensitivity of 76.54%and an area under the curve of 69.24%in distinguishing between the cases of lung cancer and normal cases on test sets.
文摘Near-infrared reflectance spectroscopy (NIRS) was applied to classify grape wines of different geographical origins (Changli, Huailai, and Yantai, China). Near infrared (NIR) spectra were collected in transmission mode in the wavelength range of 800-2500 nm. Wines (n=90) were randomly split into two sets, calibration set (n=54) and validation set (n=36). Discriminant analysis models were developed using BP neural network and discriminant partial least-squares discriminant analysis (PLS-DA). The prediction performance of calibration models in different wavelength range was also investigated. BP neural network models and PLS-DA models correctly classified 100% of the wines in calibration set. When used to predict wines in validation set, BP neural network models correctly classified 100%, 81.8%, and 90.9% of the wines from Changli, Huailai, and Yantai respectively, and PLS-DA models correctly classified 100% of all samples. The results demonstrated that NIRS could be used to discriminate Chinese grape wines as a rapid and reliable method.
基金supported by the National Natural Science Foundation of China (No. 10972019)the Innovation Foundation of BUAA for Ph.D. Graduates of China, and the China Scholarship Council
文摘This manuscript presents a stochastic model updating method, taking both uncertainties in models and variability in testing into account. The updated finite element(FE) models obtained through the proposed technique can aid in the analysis and design of structural systems. The authors developed a stochastic model updating method integrating distance discrimination analysis(DDA) and advanced Monte Carlo(MC) technique to(1) enable more efficient MC by using a response surface model,(2) calibrate parameters with an iterative test-analysis correlation based upon DDA, and(3) utilize and compare different distance functions as correlation metrics. Using DDA, the influence of distance functions on model updating results is analyzed. The proposed stochastic method makes it possible to obtain a precise model updating outcome with acceptable calculation cost. The stochastic method is demonstrated on a helicopter case study updated using both Euclidian and Mahalanobis distance metrics. It is observed that the selected distance function influences the iterative calibration process and thus, the calibration outcome, indicating that an integration of different metrics might yield improved results.
基金supported by the Natural Science Foundation of Zhejiang Province(No.Y16F020023)
文摘Traditional hand-crafted features for representing local image patches are evolving into current data-driven and learning-based image feature, but learning a robust and discriminative descriptor which is capable of controlling various patch-level computer vision tasks is still an open problem. In this work, we propose a novel deep convolutional neural network(CNN) to learn local feature descriptors. We utilize the quadruplets with positive and negative training samples, together with a constraint to restrict the intra-class variance, to learn good discriminative CNN representations. Compared with previous works, our model reduces the overlap in feature space between corresponding and non-corresponding patch pairs, and mitigates margin varying problem caused by commonly used triplet loss. We demonstrate that our method achieves better embedding result than some latest works, like PN-Net and TN-TG, on benchmark dataset.
文摘Background: Differentiating intracerebral hemorrhage (ICH) from cerebral infarction as early as possible is vital tbr the timely initiation of different treatments. This study developed an applicable model for the ambulance system to differentiate stroke subtypes. Methods: From 26,163 patients initially screened over 4 years, this study comprised 1989 consecutive patients with potential first-ever acute stroke with sudden onset of the focal neurological deficit, conscious or not, and given ambulance transport for admission to two county hospitals in Yutian County of Hebei Province. All the patients underwent cranial computed tomography (CT) or magnetic resonance imaging to confirm the final diagnosis based on stroke criteria. Correlation with stroke subtype clinical features was calculated and Bayes' discriminant model was applied to discriminate stroke subtypes. Results: Among the 1989 patients, 797,689, 109, and 394 received diagnoses of cerebral infarction, ICH, subarachnoid hemorrhage, and other forms of nonstroke, respectively. A history of atrial fibrillation, vomiting, and diabetes mellitus were associated with cerebral infarction, while vomiting, systolic blood pressure _〉180 mmHg, and age 〈65 years were more typical of ICH. For noncomatose stroke patients, Bayes' discriminant model for stroke subtype yielded a combination of multiple items that provided 72.3% agreement in the test model and 79.3% in the validation model; for comatose patients, corresponding agreement rates were 75.4% and 73.5%. Conclusions: The model herein presented, with multiple parameters, can predict stroke subtypes with acceptable sensitivity and specificity before CT scanning, either in alert or comatose patients. This may facilitate prehospital management for patients with stroke.
基金supported by the Fundamental Research Funds of Chinese Academy of Forestry(CAF)(CAFYBB2017QC002,CAFYBB2019QD002)。
文摘Place of origin has an important influence on walnut quality and commercial value,which results in the requirement of rapid geographical traceability method.Thus,a method for geographical origin identification of walnuts on the basis of nutritional quality of walnuts from China was conducted.The concentrations of 43 phytochemical components were analyzed in walnut samples from five different walnut-producing regions of China.Based on 14 chemical markers selected by the Random Forest method from these phytochemical components,a new discriminant model for geographical origin was built,with the corresponding correct classification rate of 99.3%.In addition,the quantitative quality differences of walnuts from five regions were analyzed,with values of 0.17–1.43.Moreover,the top three chemical markers for the geographical origin discriminant analysis were Mo,V,and stearic acid,with contribution rates of 26.8%,18.9%,and 10.9%,respectively.This study provides a potentially viable method for application in food authentication.
基金supported by the Science and Technology Project of State Grid Corporation of China(No.5455HJ180018).
文摘With the rapid growth of power systems measurements in terms of size and complexity,discovering statistical patterns for a large variety of real-world applications such as renewable energy prediction,demand response,energy disaggregation,and state estimation is considered a crucial challenge.In recent years,deep learning has emerged as a novel class of machine learning algorithms that represents power systems data via a large hypothesis space that leads to the state-of-the-art performance compared to most recent data-driven algorithms.This study explores the theoretical advantages of deep representation learning in power systems research.We review deep learning methodologies presented and applied in a wide range of supervised,unsupervised,and semi-supervised applications as well as reinforcement learning tasks.We discuss various settings of problems solved by discriminative deep models including stacked autoencoders and convolutional neural networks as well as generative deep architectures such as deep belief networks and variational autoencoders.The theoretical and experimental analysis of deep neural networks in this study motivates longterm research on optimizing this cutting-edge class of models to achieve significant improvements in the future power systems research.
文摘An important component of a spoken term detection (STD) system involves estimating confidence measures of hypothesised detections.A potential problem of the widely used lattice-based confidence estimation,however,is that the confidence scores are treated uniformly for all search terms,regardless of how much they may differ in terms of phonetic or linguistic properties.This problem is particularly evident for out-of-vocabulary (OOV) terms which tend to exhibit high intra-term diversity.To address the impact of term diversity on confidence measures,we propose in this work a term-dependent normalisation technique which compensates for term diversity in confidence estimation.We first derive an evaluation-metric-oriented normalisation that optimises the evaluation metric by compensating for the diverse occurrence rates among terms,and then propose a linear bias compensation and a discriminative compensation to deal with the bias problem that is inherent in lattice-based confidence measurement and from which the Term Specific Threshold (TST) approach suffers.We tested the proposed technique on speech data from the multi-party meeting domain with two state-ofthe-art STD systems based on phonemes and words respectively.The experimental results demonstrate that the confidence normalisation approach leads to a significant performance improvement in STD,particularly for OOV terms with phonemebased systems.
基金Supported by the Open Projects Program of National Laboratory of Pattern Recognition and the National Natural Science Foundation of China(91338113)
文摘The performance of scene classification of satellite images strongly relies on the discriminative power of the low-level and mid-level feature representation. This paper presents a novel approach, named multi-level max-margin analysis (M 3 DA) for semantic classification for high-resolution satellite images. In our M 3 DA model, the maximum entropy discrimination latent Dirichlet allocation (MedLDA) model is applied to learn the topic-level features first, and then based on a bag-of-words repre- sentation of low-level local image features, the large margin nearest neighbor (LMNN) classifier is used to optimize a multiple soft label composed of word-level features (generated by SVM classifier) and topic-level features. The categorization performances on 21-class land-use dataset have demonstrated that the proposed model in multi-level max-margin scheme can distinguish different categories of land-use scenes reasonably.