To identify some special formation lithology with imbalanced logging data, a framework of Multi-layer lithology identification method is proposed. In this framewoke, some special lithology is divided into one class in...
详细信息
To identify some special formation lithology with imbalanced logging data, a framework of Multi-layer lithology identification method is proposed. In this framewoke, some special lithology is divided into one class in the first layer, and each lithology is separated in the second layer. A novel algorithm of Ada Cost2-support vector machine(AdaC2-SVM) is put forward using logging data of actual well located in Karamay for training, and the support vector machine-recursive feature elimination(SVM-RFE) is adopted to select attribute, and logging data from another well nearby is used for testing. Experiment result shows the G-mean and accuracy of our method is up to 95.3% and 94.4%, which has better performance than random forest(RF)algorithm, particle swarm optimization-support vector machine(PSO-SVM) algorithm and improved PSO-SVM(IPSO-SVM)algorithm. In the future, the proposed method have a good prospect and give a valuable result for geology research.
To realize drilling visualization, an effective interpolation method is essential to construct three-dimensional *** is an effective interpolation method commonly used by geologists. However, the variogram model param...
详细信息
To realize drilling visualization, an effective interpolation method is essential to construct three-dimensional *** is an effective interpolation method commonly used by geologists. However, the variogram model parameters in traditional Kriging have a certain subjectivity, which will influence the accuracy of interpolation. Quantum Genetic Algorithm(QGA)is introduced to optimize the variogram model parameters selection in this paper. Elevation values of boreholes are used as data sets for simulation. The results show that the proposed improved Kriging has a better prediction accuracy.
A new method for localization of epileptic seizure onset zones(SOZs) is proposed, which uses the Shannon-entropybased complex Morlet wavelet transform to extract a satisfactory time-frequency feature of high-frequen...
详细信息
A new method for localization of epileptic seizure onset zones(SOZs) is proposed, which uses the Shannon-entropybased complex Morlet wavelet transform to extract a satisfactory time-frequency feature of high-frequency oscillations(Hfos).The singular value decomposition and the K-medoids clustering algorithm are employed to extract effective features from the redundant matrix of wavelet coefficients. A distinctive feature is to use the singular values to detect Hfos with the consideration that the singular values of Hfos are generally significantly higher than those of normal case. Based on the half-maximum method,the localization of SOZs are achieved by using the characteristics of Hfos. Comparisons show that our method provides a higher sensitivity and specificity than two existing methods do.
Facial expression recognition(FER) plays an important role in human-machine interaction. An assistant robot having a close interaction with human being should be able to recognize human facial expression. FER is a non...
详细信息
Facial expression recognition(FER) plays an important role in human-machine interaction. An assistant robot having a close interaction with human being should be able to recognize human facial expression. FER is a non-trivial problem because each individual has his own way to reveal his emotion and the facial expressions of two different persons may not be totally identical. Hence,facial expression recognition is still a challenging problem in computer vision. In this work, we propose a simple solution for facial expression recognition that uses a combination of Convolutional Neural Network and specific image pre-processing *** experiments employed to evaluate our technique were carried out using two largely used public databases(CK+, JAFFE).A study of the impact of each image pre-processing operation in the accuracy rate is presented. The proposed method: achieves competitive results when compared with other facial expression recognition methods-97.85% of accuracy in the CK+ database-it is fast to train,and it allows for real time facial expression recognition with standard computers.
This paper investigates the stability of neural networks with a time-varying *** on the good effectiveness of the augmented Lyapunov-Krasovskii functional(LKF),some useful integral vectors are summarized and used to c...
详细信息
This paper investigates the stability of neural networks with a time-varying *** on the good effectiveness of the augmented Lyapunov-Krasovskii functional(LKF),some useful integral vectors are summarized and used to construct single integral terms with augmented quadratic integrand so as to develop a novel augmented LKF *** an extended reciprocally convex matrix inequality and an auxiliary function-based inequality are utilized to estimate the derivative of the *** a result,an improved stability criterion is ***,the advantage of proposed method is demonstrated by a numerical example.
To improve the accuracy of Electroencephalogram(EEG) emotion recognition,a stacking emotion classification model is proposed,in which different classification models such as XGBoost,LightGBM and Random forest are inte...
详细信息
To improve the accuracy of Electroencephalogram(EEG) emotion recognition,a stacking emotion classification model is proposed,in which different classification models such as XGBoost,LightGBM and Random forest are integrated to learn the *** addition,the Renyi entropy of 32 channels' EEG signals are extracted as the feature and Linear discriminant analysis(LDA) is employed to reduce the dimension of the feature *** proposal is tested on the DEAP dataset,and the EEG emotional states are accessed in Arousal-Valence emotion space,in which HA/LA and HV/LV are classified,*** result shows that the average recognition accuracies of 77.19%for HA/LA and 79.06%for HV/LV are obtained,which demonstrates that the proposal is feasible in EEG emotion recognition.
The sintering process is one of the most energy-consuming processes in steelmaking, its carbon fuel consumption accounts for 8% to 10% in the steel production *** find ways of reducing the energy consumption, it is ne...
详细信息
The sintering process is one of the most energy-consuming processes in steelmaking, its carbon fuel consumption accounts for 8% to 10% in the steel production *** find ways of reducing the energy consumption, it is necessary to predict the carbon efficiency. The value of CO/CO in the carbon emission can reflect the utilization of carbon combustion in sintering process. In this study, the CO/CO is taken to be a measure of carbon efficiency and a hierarchical model is built to predict it. Firstly, the physical and chemical reactions and the carbon flow mechanism in the sintering process are analyzed, and the process parameters that affect the CO/CO are determined. Then, the gray relational analysis method is used to analyze the influence factors to determine the relationship between the parameters, and a hierarchical predictive model for CO/CO is established based on the relationship between the parameters. The hierarchical predictive model is divided into two parts: the predictive models for the thermal state parameters and the predictive model for CO/CO. The inputs of the predictive models for the thermal state parameters are the raw material parameters and the operating parameters, and the inputs of the predictive model for CO/CO are the predicted values of the predictive models for the thermal state parameters. Finally, the simulation results verify the effectiveness of the proposed modeling method. This method can provide a theoretical basis for the optimization and control of carbon efficiency in the sintering process.
With the application of magnetic thin films becoming more and more widespread,people pay more and more attention to the performance *** order to obtain a magnetic film with a specific performance,it is very important ...
详细信息
With the application of magnetic thin films becoming more and more widespread,people pay more and more attention to the performance *** order to obtain a magnetic film with a specific performance,it is very important to judge the quality of the magnetic film and measure the magnetic properties of the ***,with the increase of the film preparation process,the thickness of the prepared film is getting thinner and the magnetic moment signal contained therein is also *** brings a certain degree of difficulty to the traditional measurement *** example,the VSM system that obtains the hysteresis loop by measuring the magnetic moment signal has become somewhat inadequate for the measurement of ultra-thin *** order to solve this issue,a new method based on anomalous Hall effect is introduced in this *** test system of this system adopts the four-probe measuring method,a constant current is applied across the surface of the film sample,and the abnormal Hall voltage is measured at the other two *** R-H curve of the sample can be obtained through *** compared to VSM measurement,this method is simpler and stable,more accurate,which can greatly reduce the anomalous Hall-effect device R-H characteristic measurement cost.
This paper investigates the problem of finite-time H∞ state estimation for discrete-time stochastic switched genetic regulatory networks(GRNs) with time-varying delays and exogenous disturbances. A new discrete tim...
详细信息
This paper investigates the problem of finite-time H∞ state estimation for discrete-time stochastic switched genetic regulatory networks(GRNs) with time-varying delays and exogenous disturbances. A new discrete time-delayed stochastic switched GRN model with uncertain sojourn probabilities is devised, which is more general than the switched GRNs model with completely known sojourn probabilities. The sufficient conditions which guarantee the stochastic finite-time boundedness of the estimation error dynamics with a prescribed H∞ disturbance attenuation level are derived. By solving several matrix inequalities,the state estimator parameters can be obtained. A numerical example is given to illustrate the effectiveness of our results.
Considering the difficulties in estimating depth from single image, in this paper, we propose a method to obtain the absolute scale depth map by combining the convolution neural network and depth filter. We compute re...
详细信息
Considering the difficulties in estimating depth from single image, in this paper, we propose a method to obtain the absolute scale depth map by combining the convolution neural network and depth filter. We compute relative transformation between consecutive frames by direct tracking features, which are extracted from RGB images and whose depthes are predicted by deep network, and then optimize relative motion by searching for a better feature alignment in epipolar line, and finally update every pixel depth of the reference frame by depth filter. We evaluate the proposed method on the open dataset comparison against the state of the art in depth estimation to evaluate our method.
暂无评论