• Journal of Broadcast Engineering
• /
• v.25 no.5
• /
• pp.742-749
• /
• 2020

This paper proposes a method for recognizing both phoneme and emotion using a common neural network and a multi-task training method for the common neural network. The common neural network performs the same function for both recognition tasks, which corresponds to the structure of multi-information recognition of human using a single auditory system. The multi-task training conducts a feature modeling that is commonly applicable to multiple information and provides generalized training, which enables to improve the performance by reducing an overfitting occurred in the conventional individual training for each information. A method for increasing phoneme recognition performance is also proposed that applies weight to the phoneme in the multi-task training. When using the same feature vector and neural network, it is confirmed that the proposed common neural network with multi-task training provides higher performance than the individual one trained for each task.

• Proceedings of the KAIS Fall Conference
• /
• 2010.11a
• /
• pp.344-348
• /
• 2010

The intent of this paper is to describe a neural network structure called multi dynamic neural network(MDNN), and examine how it can be used in developing a learning scheme for computing robot inverse kinematic transformations. The architecture and learning algorithm of the proposed dynamic neural network structure, the MDNN, are described. Computer simulations are demonstrate the effectiveness of the proposed learning using the MDNN.

• Proceedings of the KAIS Fall Conference
• /
• 2010.11a
• /
• pp.332-336
• /
• 2010

The intent of this paper is to describe a neural network structure called multi dynamic neural network(MDNN), and examine how it can be used in developing a learning scheme for computing robot inverse kinematic transformations. The architecture and learning algorithm of the proposed dynamic neural network structure, the MDNN, are described. Computer simulations are demonstrate the effectiveness of the proposed learning using the MDNN.

• Proceedings of the KAIS Fall Conference
• /
• 2009.05a
• /
• pp.454-457
• /
• 2009

The intent of this paper is to describe a neural network structure called multi dynamic neural network(MDNN), and examine how it can be used in developing a learning scheme for computing robot inverse kinematic transformations. The architecture and learning algorithm of the proposed dynamic neural network structure, the MDNN, are described. Computer simulations are demonstrate the effectiveness of the proposed learning using the MDNN.

• Proceedings of the IEEK Conference
• /
• 2001.06e
• /
• pp.115-118
• /
• 2001

In this paper we propose an FDI(fault detection and isolation) algorithm using neural network-based multi-fault models to detect and isolate single faults in nonlinear systems. When a change in the system occurs, the errors between the system output and the neural network nominal system output cross a threshold, and once a fault in the system is detected, the fault classifier statistically isolates the fault by using the error between each neural network-based fault model output and the system output.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.29B no.1
• /
• pp.58-66
• /
• 1992

This paper proposes a neuro-controller for position control of one-link flexible robot arm. Basically the controller consists of a multi-layer neural network and a conventional PD controller. Two controller are parallelly connected. Neural network is traind by the conventional error back propagation learning rules. During learning period, the weights of neural network are adjusted to minimize the position error between the desired hub angle and the actual one. Finally the effectiveness of the proposed approach will be demonstrated by computer simulation.

• Proceedings of the KAIS Fall Conference
• /
• 2006.11a
• /
• pp.122-128
• /
• 2006

The intent of this paper is to describe a neural network structure called multi dynamic neural network(MDNN), and examine how it can be used in developing a learning scheme for computing robot inverse kinematic transformations. The architecture and learning algorithm of the proposed dynamic neural network structure, the MDNN, are described. Computer simulations are demonstrate the effectiveness of the proposed learning using the MDNN.

• Journal of Information Processing Systems
• /
• v.17 no.3
• /
• pp.556-570
• /
• 2021

The existing video expression recognition methods mainly focus on the spatial feature extraction of video expression images, but tend to ignore the dynamic features of video sequences. To solve this problem, a multi-mode convolution neural network method is proposed to effectively improve the performance of facial expression recognition in video. Firstly, OpenFace 2.0 is used to detect face images in video, and two deep convolution neural networks are used to extract spatiotemporal expression features. Furthermore, spatial convolution neural network is used to extract the spatial information features of each static expression image, and the dynamic information feature is extracted from the optical flow information of multiple expression images based on temporal convolution neural network. Then, the spatiotemporal features learned by the two deep convolution neural networks are fused by multiplication. Finally, the fused features are input into support vector machine to realize the facial expression classification. Experimental results show that the recognition accuracy of the proposed method can reach 64.57% and 60.89%, respectively on RML and Baum-ls datasets. It is better than that of other contrast methods.

• Journal of the Korea Society of Computer and Information
• /
• v.8 no.3
• /
• pp.113-123
• /
• 2003

Many researches on the application of neural networks for making financial index prediction have proven their advantages over statistical and other methods. In this paper, a neural network model is proposed for the Buying, Holding or Selling timing prediction in stocks by the price index of stocks by inputting the closing price and volume of dealing in stocks and the technical indexes(MACD, Psychological Line). This model has an equalized multi-layer arithmetic function as well as the time series prediction function of backpropagation neural network algorithm. In the case that the numbers of learning data are unbalanced among the three categories (Buying, Holding or Selling), the neural network with conventional method has the problem that it tries to improve only the prediction accuracy of the most dominant category. Therefore, this paper, after describing the structure, working and learning algorithm of the neural network, shows the equalized multi-layer arithmetic method controlling the numbers of learning data by using information about the importance of each category for improving prediction accuracy of other category. Experimental results show that the financial index prediction using the equalized multi-layer arithmetic neural network has much higher correctness rate than the other conventional models.

• Journal of Power System Engineering
• /
• v.5 no.2
• /
• pp.93-99
• /
• 2001

This paper presents the automatic recognition system for car number plate. In our country, two types of number plate pattern is used. The one is old type of number plate, the other is new type of number plate. To recognize both new and old type number plates, the system must have flexibility. Therefore, in this paper, automatic recognition system is developed by use of the neural network for good adaptation, good generalization, and modulation. And because the number plate is made of three codes, the multi neural network consists of three networks. Neural network is teamed by GDR(Generalized Delta learning Rule) and it is verified the effectiveness of the method through experimental results.