To enhance classification performance by making use of easily available unlabelled data to overcome the scarcity of labelled data, this paper proposes an Embedded Co-Adaboost algorithm that integrates multi-view learn...
详细信息
Active ultrasonic sensors for target tracking application may suffer from inter-sensor-interference if these highly dense deployed sensors are not scheduled, which can degrade the tracking performance. In this paper, ...
详细信息
In view of the complexity of function mapping coding and the uncertainty of function, this paper proposes that function equation should be separated from the algorithm to design an input program, and then multiple sof...
详细信息
An object tracking algorithm based on SURF is presented in this paper. Interest points are detected by SURF detector in reference region which is located in the first frame manually. In the following sequences, the SU...
详细信息
An object tracking algorithm based on SURF is presented in this paper. Interest points are detected by SURF detector in reference region which is located in the first frame manually. In the following sequences, the SURF feature points are extracted in a larger window which is selected as test region. In the matching stage, we calculate the Euclidean distance between the descriptor vectors of interest points in the test image and ones in the reference image. The proposed algorithm is implemented on an embedded platform with TI's DM6446 high performance processor. The experimental results show that our system implements a real-time tracking with robustness against appearance variations, scale change, and cluttered scenes.
An improved included angle dividing method is proposed for Hammerstein systems. In this dividing method, nonlinearity of the system is taken into account to distribute gridding points, so distribution is more reasonab...
详细信息
An improved included angle dividing method is proposed for Hammerstein systems. In this dividing method, nonlinearity of the system is taken into account to distribute gridding points, so distribution is more reasonable and computation load in dividing is reduced largely. Based on the dividing result, a multimodel predictive controller is designed to overcome the drawbacks of the common nonlinearity inversion control method for Hammerstein systems.
This paper is concerned with identification of linear parameter varying (LPV) systems in an input-output setting with Box-Jenkins (BJ) model structure. Classical linear time invariant prediction error method (PEM) is ...
详细信息
Aimed to solve the limitation of abundant data to constructing classification modeling in data mining, the paper proposed a novel effective preprocessing algorithm based on rough sets. Firstly, we construct the relati...
详细信息
Aimed to solve the limitation of abundant data to constructing classification modeling in data mining, the paper proposed a novel effective preprocessing algorithm based on rough sets. Firstly, we construct the relation Information System using original data sets. Secondly, make use of attribute reduction theory of Rough sets to produce the Core of Information System. Core is the most important and necessary information which cannot reduce in original Information System. So it can get a same effect as original data sets to data analysis, and can construct classification modeling using it. Thirdly, construct indiscernibility matrix using reduced Information System, and finally, get the classification of original data sets. Compared to existing techniques, the developed algorithm enjoy following advantages:(1) avoiding the abundant data in follow-up data processing, and(2) avoiding large amount of computation in whole data mining process.(3) The results become more effective because of introducing the attributes reducing theory of Rough Sets.
In this paper, a hierarchical video summarization representation algorithm was proposed for video analysis in compressed domain. In particular, Rough Sets(RS) theory is introduced for video analysis to inrease the eff...
详细信息
In this paper, a hierarchical video summarization representation algorithm was proposed for video analysis in compressed domain. In particular, Rough Sets(RS) theory is introduced for video analysis to inrease the efficiency of algorithm. Firstly, DCT coefficients and DC coefficients are extracted from video image sequences, so an Information System can construct with DC coefficient. Then, Information System is reduced by ruduction theory of RS, the representation of the video frame is obtained by reduced DC coefficients. Finaly, we can obtain the reduced Information System, i.e. the Core of Information System. Since the Core contained all the information in video sequences, and at the same time it banished redundant video frame, so it can be viewed as the efficient summarization representation. As our experimental results indicate that the algorithm can efficiently generate a set of video summarization representative of compressed domain videos sequences. Compared to conventional algorithm, the algorithm enjoys following advantages. (1)only a subset of video frames considered during video analysis, so it can avoid the computational complexity. (2) the video summarization representation becomes more scientific and efficient than previous methods.
Aimed to solve the limitation of abundant data to constructing classification modeling in data mining, the paper proposed a novel effective preprocessing algorithm based on rough sets. Firstly, we construct the relati...
详细信息
Aimed to solve the limitation of abundant data to constructing classification modeling in data mining, the paper proposed a novel effective preprocessing algorithm based on rough sets. Firstly, we construct the relation Information System using original data sets. Secondly, make use of attribute reduction theory of Rough sets to produce the Core of Information System. Core is the most important and necessary information which cannot reduce in original Information System. So it can get a same effect as original data sets to data analysis, and can construct classification modeling using it. Thirdly, construct indiscernibility matrix using reduced Information System, and finally, get the classification of original data sets. Compared to existing techniques, the developed algorithm enjoy following advantages: (1) avoiding the abundant data in follow-up data processing, and (2) avoiding large amount of computation in whole data mining process. (3) The results become more effective because of introducing the attributes reducing theory of Rough Sets.
Over the decade, due to the fact that the global energy resources are in deadly shortage, much emphasis is put on the energy consumption in thermal power throughout the world. Following developed is Circulating Fluidi...
详细信息
Over the decade, due to the fact that the global energy resources are in deadly shortage, much emphasis is put on the energy consumption in thermal power throughout the world. Following developed is Circulating Fluidized Bed Boiler (CFBB) in recent years, a kind of combustion boiler that can clean and desulfurize the coal efficiently in the combustion process. Circulating Fluidized Bed Boiler (CFBB) is a control object with features of time varying parameters, large delay, and multivariable control tightly coupled. Noticeably, many factors influence the combustion process. This paper designs a three level ART2-BP-BP of data fusion--fusion cluster control system based on methods of multi-sensor information fusion and cluster analysis. It completes data fusion from the data level, the feature level to the decision level. Especially, the concept of Situation Threat Space aiming at the potential threats in CFBB is presented. Results of simulation show that the control system in this paper is feasible and effective, in particular, the control system still has more satisfactory control effect in the case of a variety of sensor failures.
暂无评论