In order to analyze changes in human settlement in Xuzhou city during the past 20 years, changes in land cover and vegetation were investigated based on multi-temporal remote sensing Landsat TM images. We developed a ...In order to analyze changes in human settlement in Xuzhou city during the past 20 years, changes in land cover and vegetation were investigated based on multi-temporal remote sensing Landsat TM images. We developed a hierarchical classifier system that uses different feature inputs for specific classes and conducted a classification post-processing approach to improve its accuracy. From our statistical analysis of changes in urban land cover from 1987 to 2007, we conclude that built-up land areas have obviously increased, while farmland has seen in a continuous loss due to urban growth and human activities. A NDVI difference approach was used to extract information on changes in vegetation. A false change information elimination approach was developed based on prior knowledge and statistical analysis. The areas of vegetation cover have been in continuous decline over the past 20 years, although some measures have been adopted to protect and maintain urban vegetation. Given the stability of underground coal exploitation since 1990s, urban growth has become the major driving force in vegetation loss, which is different from the vegetation change driven by coal exploitation mainly before 1990.展开更多
准确,高效的业务流识别与分类是保障多媒体通信端到端Qo S(Quality of Service),执行相关网络操作的前提。如今数据规模的剧烈增加为业务流的分类提出了挑战,而特征选择能够尽可能地减少特征维数,去除冗余特征,为大数据时代下的业务流...准确,高效的业务流识别与分类是保障多媒体通信端到端Qo S(Quality of Service),执行相关网络操作的前提。如今数据规模的剧烈增加为业务流的分类提出了挑战,而特征选择能够尽可能地减少特征维数,去除冗余特征,为大数据时代下的业务流分类提供解决办法。对现有的特征选择方法分成Filter、Wrapper、Embedded三类,分析了各类算法的性能原理。采用最新数据集对不同特征选择算法性能对比,从算法的运行时间、特征压缩率、准确率三个方面评估了特征选择算法的性能。另外,针对现有数据集分类情况进行分级分类以达到视频流的细分类,从而提高分类的准确率。展开更多
Automatic Speech Emotion Recognition(SER)is used to recognize emotion from speech automatically.Speech Emotion recognition is working well in a laboratory environment but real-time emotion recognition has been influen...Automatic Speech Emotion Recognition(SER)is used to recognize emotion from speech automatically.Speech Emotion recognition is working well in a laboratory environment but real-time emotion recognition has been influenced by the variations in gender,age,the cultural and acoustical background of the speaker.The acoustical resemblance between emotional expressions further increases the complexity of recognition.Many recent research works are concentrated to address these effects individually.Instead of addressing every influencing attribute individually,we would like to design a system,which reduces the effect that arises on any factor.We propose a two-level Hierarchical classifier named Interpreter of responses(IR).Thefirst level of IR has been realized using Support Vector Machine(SVM)and Gaussian Mixer Model(GMM)classifiers.In the second level of IR,a discriminative SVM classifier has been trained and tested with meta information offirst-level classifiers along with the input acoustical feature vector which is used in primary classifiers.To train the system with a corpus of versatile nature,an integrated emotion corpus has been composed using emotion samples of 5 speech corpora,namely;EMO-DB,IITKGP-SESC,SAVEE Corpus,Spanish emotion corpus,CMU's Woogle corpus.The hierarchical classifier has been trained and tested using MFCC and Low-Level Descriptors(LLD).The empirical analysis shows that the proposed classifier outperforms the traditional classifiers.The proposed ensemble design is very generic and can be adapted even when the number and nature of features change.Thefirst-level classifiers GMM or SVM may be replaced with any other learning algorithm.展开更多
基金supported by the National High Technology Research and Developmemt Program of China (No2007AA12Z162)the Program for New Century Excellent Talents in University, Ministry of Education (NoNCET-06-0476)the Jiangsu Provincial 333 Engineering for High Level Talents(No.BK2006505)
文摘In order to analyze changes in human settlement in Xuzhou city during the past 20 years, changes in land cover and vegetation were investigated based on multi-temporal remote sensing Landsat TM images. We developed a hierarchical classifier system that uses different feature inputs for specific classes and conducted a classification post-processing approach to improve its accuracy. From our statistical analysis of changes in urban land cover from 1987 to 2007, we conclude that built-up land areas have obviously increased, while farmland has seen in a continuous loss due to urban growth and human activities. A NDVI difference approach was used to extract information on changes in vegetation. A false change information elimination approach was developed based on prior knowledge and statistical analysis. The areas of vegetation cover have been in continuous decline over the past 20 years, although some measures have been adopted to protect and maintain urban vegetation. Given the stability of underground coal exploitation since 1990s, urban growth has become the major driving force in vegetation loss, which is different from the vegetation change driven by coal exploitation mainly before 1990.
文摘准确,高效的业务流识别与分类是保障多媒体通信端到端Qo S(Quality of Service),执行相关网络操作的前提。如今数据规模的剧烈增加为业务流的分类提出了挑战,而特征选择能够尽可能地减少特征维数,去除冗余特征,为大数据时代下的业务流分类提供解决办法。对现有的特征选择方法分成Filter、Wrapper、Embedded三类,分析了各类算法的性能原理。采用最新数据集对不同特征选择算法性能对比,从算法的运行时间、特征压缩率、准确率三个方面评估了特征选择算法的性能。另外,针对现有数据集分类情况进行分级分类以达到视频流的细分类,从而提高分类的准确率。
文摘Automatic Speech Emotion Recognition(SER)is used to recognize emotion from speech automatically.Speech Emotion recognition is working well in a laboratory environment but real-time emotion recognition has been influenced by the variations in gender,age,the cultural and acoustical background of the speaker.The acoustical resemblance between emotional expressions further increases the complexity of recognition.Many recent research works are concentrated to address these effects individually.Instead of addressing every influencing attribute individually,we would like to design a system,which reduces the effect that arises on any factor.We propose a two-level Hierarchical classifier named Interpreter of responses(IR).Thefirst level of IR has been realized using Support Vector Machine(SVM)and Gaussian Mixer Model(GMM)classifiers.In the second level of IR,a discriminative SVM classifier has been trained and tested with meta information offirst-level classifiers along with the input acoustical feature vector which is used in primary classifiers.To train the system with a corpus of versatile nature,an integrated emotion corpus has been composed using emotion samples of 5 speech corpora,namely;EMO-DB,IITKGP-SESC,SAVEE Corpus,Spanish emotion corpus,CMU's Woogle corpus.The hierarchical classifier has been trained and tested using MFCC and Low-Level Descriptors(LLD).The empirical analysis shows that the proposed classifier outperforms the traditional classifiers.The proposed ensemble design is very generic and can be adapted even when the number and nature of features change.Thefirst-level classifiers GMM or SVM may be replaced with any other learning algorithm.
