Automatic recognition of both printed and handwritten characters is the most progressive research area since last few periods. The printed character recognition rate still desires concentration of researchers because ...
详细信息
Automatic recognition of both printed and handwritten characters is the most progressive research area since last few periods. The printed character recognition rate still desires concentration of researchers because of variations in shape, size and design of characters. Odia is the native language of Odisha. Less works have been accounted in Odia script classification. Complex printed character recognition constantly demands a better feature extraction method for such diversified characters. Purpose of this manuscript is to apply autoencoder technique considering a suitable classifier for Odia printed numeral recognition. autoencoder proved to be a better method in performance for dimensionality reduction as well as classification of Odia numerals. (C) 2019 The Authors. Published by Elsevier Ltd.
Ischemic heart disease is one of the most common causes of death in the world. Recent studies show that it may be caused by sleep apnea, or disordered breathing during sleep;however, sleep apnea lacks subjective sympt...
详细信息
ISBN:
(纸本)9781728116341
Ischemic heart disease is one of the most common causes of death in the world. Recent studies show that it may be caused by sleep apnea, or disordered breathing during sleep;however, sleep apnea lacks subjective symptoms. Therefore, an effective method is required to monitor sleep apnea states while a subject is at home. In this context, this study proposes a method to estimate apnea conditions using an unconstrained sensing system placed under a mattress. The system provides the probability of the user being in an apnea state for given units of time. To reduce the effects from each individual user and their sleeping position, we applied a stacked autoencoder to obtain feature vectors. These are given to a feed-forward neural network with a two-dimensional output layer. The two-dimensional output is converted to probability values by a softmax function, and labels with a higher probability are estimated as an apnea state.
in this paper, we consider a problem of a biomechanical distraction of a driver;mostly it is related with hands secondary action while driving. Dealing with this type of distraction is crucial because the main causes ...
详细信息
ISBN:
(纸本)9781728151021
in this paper, we consider a problem of a biomechanical distraction of a driver;mostly it is related with hands secondary action while driving. Dealing with this type of distraction is crucial because the main causes of car accidents are due to negligent operation of in-vehicle technology, operating mobile phone while driving, and chatting with passenger. The effects of illumination conditions and driver's hands skin color make biomechanical distraction recognition very challenging. Accordingly, we propose a deep recognition model which can handle both effects, and its computational cost is relatively less when compared with the current state of the art. Our model is a sequential integration of two sub module: hand and face localizer and different types of biomechanical distraction recognizer. We conduct several simulations to investigate the effect of hand and face localization, and to measure the recognition performance of the proposed model, and its recognition performance, based on our experiment, our model achieves 98.68 % validation accuracy which is better than the current state of the art.
Artificial neural network method is a common method for short-term electricity price forecasting. However, when the amount of input and output data is large, the training speed will be slow, and it is easy to fall int...
详细信息
ISBN:
(纸本)9781728126586
Artificial neural network method is a common method for short-term electricity price forecasting. However, when the amount of input and output data is large, the training speed will be slow, and it is easy to fall into local extreme values or even the result is difficult to converge. In view of this, the paper proposes a deep learning model based on stacked autoencoder (SAE) to predict electricity price. This paper analyzes the factors affecting electricity price, proposes an algorithm based on stacked autoencoder model, and uses MATLAB tools to predict electricity price in PJM power market. The comparison between SAE algorithm and BP algorithm is carried out in the example. The results show that the prediction results based on SAE model are more accurate. The deep learning model has better ability to express the objective function than the shallow model, and can effectively solve the problem of traditional neural network training difficulties.
Fisher discriminant analysis (FDA) is an effective fault classification method for complicated industrial processes. However, it builds the classifier only based on the labeled training data and ignores vast amounts o...
详细信息
ISBN:
(纸本)9781728140940
Fisher discriminant analysis (FDA) is an effective fault classification method for complicated industrial processes. However, it builds the classifier only based on the labeled training data and ignores vast amounts of unlabeled data which can provide important data distribution information. To well utilize the unlabeled data for better fault classification performance, this paper presents a semisupervised FDA (SeFDA) method which integrates the stacked autoencoder (SAE) for intrinsic feature extraction and applies the majority voting based ensemble modeling strategy for a more precise solution. Different to the traditional FDA method, the proposed method builds the classifier by considering the labeled and unlabeled samples simultaneously. SeFDA first applies the SAE on the whole data samples to extract the data features. Then the data features corresponding to the labeled samples are used to train a fault discriminant model by FDA. Considering that the single SAE network training is easily affected by the initial weight values, a majority voting based ensemble modeling strategy is applied to combine the results from multiple classifiers. The case studies on the simulated Tennessee Eastman process and the real sucker rod pumping system show that the modified SeFDA method is superior to the basic FDA in terms of fault classification performance.
In the area of data-driven bearing prognostic, the construction of health indictor from condition monitoring data is important. This paper presents a novel bearing health indicator construction method based on ensembl...
详细信息
ISBN:
(纸本)9781538683576
In the area of data-driven bearing prognostic, the construction of health indictor from condition monitoring data is important. This paper presents a novel bearing health indicator construction method based on ensemble stacked autoencoders. Firstly, the proposed ensemble stacked autoencoders extract features directly from the FFT results of raw vibration signals. Then, a deep neural network which serves as a non-linear transformation is trained to map the multi-dimensional learned features to a one-dimensional health indicator. Finally, the proposed method is validated using the IEEE PHM2012 Challenge dataset. To show the superiority of the proposed method, its performance is evaluated and compared with other methods. The results demonstrate that the proposed method can automatically and effectively build high-quality health indictor from raw data without any signal processing and manual feature engineering.
Fault monitoring can find out-of-control conditions of equipment operation in a timely manner, which is essential for eliminating faults and for stable operation of industrial systems in batch processes. Many conventi...
详细信息
Fault monitoring can find out-of-control conditions of equipment operation in a timely manner, which is essential for eliminating faults and for stable operation of industrial systems in batch processes. Many conventional data-driven fault detection methods focus less on the non-Gaussian and Multi-stage characteristics of batch process data, which may result in degradation of monitoring performance. In this paper, a Multi-stage Fourth Order Moment Staked autoencoder(M-FOM-SAE) is designed to solve the above problems. The proposed method firstly automatically determines the number of clusters and divides the batch process into multiple stages. After that, the FOM-SAE model is established in each sub-stage, which can not only effectively learn the nonlinear features of process data, but also extract the non-Gaussian information. The proposed strategy is applied to real-world industrial processes. Experimental results indicate that it can better capture the non-Gaussian and Multi-stage characteristics of process data, and improve the ability to monitor abnormalities.
The visible light communication (VLC) system can provide data communication function as well as lighting. It is widely used in automated and sustainable smart farming for its energy efficiency, reliability, security w...
详细信息
ISBN:
(数字)9781728160672
ISBN:
(纸本)9781728160689
The visible light communication (VLC) system can provide data communication function as well as lighting. It is widely used in automated and sustainable smart farming for its energy efficiency, reliability, security without degrading the crops growth compared to radio frequency communication. A novel peak to average power ratio (PAPR) reduction method is applied based on stacked autoencoder network in VLC systems. The deep learning network is trained by a combined loss function. Simulation results reveal that the VLC system achieves a distinct PAPR reduction and shows an improved bit error ratio performance. The system can be easily implemented and controlled in the existing infrastructure. This research provides a basis for the feasibility of deep learning theory in the design of optical communication system.
Automatic recognition of both printed and handwritten characters is the most progressive research area since last few periods. The printed character recognition rate still desires concentration of researchers because ...
详细信息
Automatic recognition of both printed and handwritten characters is the most progressive research area since last few periods. The printed character recognition rate still desires concentration of researchers because of variations in shape, size and design of characters. Odia is the native language of Odisha. Less works have been accounted in Odia script classification. Complex printed character recognition constantly demands a better feature extraction method for such diversified characters. Purpose of this manuscript is to apply autoencoder technique considering a suitable classifier for Odia printed numeral recognition. autoencoder proved to be a better method in performance for dimensionality reduction as well as classification of Odia numerals.
Gear and its transmission are widely used in different transmission systems, and its complicated and changeable condition brings a series of problems to the fault feature extraction and diagnosis. In recent years, dee...
详细信息
Gear and its transmission are widely used in different transmission systems, and its complicated and changeable condition brings a series of problems to the fault feature extraction and diagnosis. In recent years, deep learning techniques have been gradually applied to feature extraction and pattern recognition, and the features of feature extraction and fault diagnosis in complex working environments have shown certain advantages. This study is based on stacked autoencoder under deep learning model, and improve training network performance by modified activation function. Through the network training before and after the experiment done, and to extract the fault feature data comparison in testing, improving network after activation function to extract fault features showed a greater advantage, can be a very good application in practical fault feature extraction.
暂无评论