• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.7 no.4
• /
• pp.212-222
• /
• 2014

In the current study, the authors propose a method for extracting license plate regions by means of a neural network trained to output the plate's center of gravity. The method is shown to be effective. Since the learning pattern presentation positions are defined by random numbers, a different pattern is submitted to the neural network for learning each time, which enables it to form a neural network with high universality of coverage. The article discusses issues of the optimal learning surface for a license plate covered by the learning pattern, the effect of suppression learning of the number and pattern enlargement/reduction and of concentration value conversion. Results of evaluation tests based on pictures of 595 vehicles taken at an under-ground parking garage demonstrated detection rates of 98.5%, 98.7%, and 100%, respectively.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.1 no.3
• /
• pp.105-109
• /
• 2008

In the current study, the authors propose a method for extracting license plate regions by means of a neural network trained to output the plate.s center of gravity. The method is shown to be effective. Since the learning pattern presentation positions are defined by random numbers, a different pattern is submitted to the neural network for learning each time, which enables it to form a neural network with high universality of coverage. The article discusses issues of the optimal learning surface for a license plate covered by the learning pattern, the effect of suppression learning of the number and headlight sections, as well as the effect of learning pattern enlargement/reduction and of concentration value conversion. Results of evaluation tests based on pictures of 595 vehicles taken at an underground parking garage demonstrated detection rates of 98.5%.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.1
• /
• pp.334-349
• /
• 2022

The effectiveness of Wi-Fi networks is greatly influenced by the optimization of contention window (CW) parameters. Unfortunately, the conventional approach employed by IEEE 802.11 wireless networks is not scalable enough to sustain consistent performance for the increasing number of stations. Yet, it is still the default when accessing channels for single-users of 802.11 transmissions. Recently, there has been a spike in attempts to enhance network performance using a machine learning (ML) technique known as reinforcement learning (RL). Its advantage is interacting with the surrounding environment and making decisions based on its own experience. Deep RL (DRL) uses deep neural networks (DNN) to deal with more complex environments (such as continuous state spaces or actions spaces) and to get optimum rewards. As a result, we present a new approach of CW control mechanism, which is termed as contention window threshold (CW_Threshold). It uses the DRL principle to define the threshold value and learn optimal settings under various network scenarios. We demonstrate our proposed method, known as a smart exponential-threshold-linear backoff algorithm with a deep Q-learning network (SETL-DQN). The simulation results show that our proposed SETL-DQN algorithm can effectively improve the throughput and reduce the collision rates.

• Journal of Korea Artificial Intelligence Association
• /
• v.2 no.1
• /
• pp.25-30
• /
• 2024

Kodály hand signs are symbols that intuitively represent pitch and note names based on the shape and height of the hand. They are an excellent tool that can be easily expressed using the human body, making them highly engaging for children who are new to music. Traditional hand signs help beginners easily understand pitch and significantly aid in music learning and performance. However, Kodály hand signs have distinctive features, such as the ability to indicate key changes or chords using both hands and to clearly represent accidentals. These features enable the effective use of Kodály hand signs. In this paper, we aim to investigate the changes in recognition rates according to the complexity of scales by creating a device for learning Kodály hand signs, teaching simple Do-Re-Mi scales, and then gradually increasing the complexity of the scales and teaching complex scales and children's songs (such as "May Had A Little Lamb"). The learning device utilizes accelerometer and bending sensors. The accelerometer detects the tilt of the hand, while the bending sensor detects the degree of bending in the fingers. The utilized accelerometer is a 6-axis accelerometer that can also measure angular velocity, ensuring accurate data collection. The learning and performance evaluation of the Kodály learning device were conducted using Python.

• International Journal of Internet, Broadcasting and Communication
• /
• v.16 no.3
• /
• pp.262-273
• /
• 2024

We provide a detailed analysis of the data processing and model training process for vulnerability classification using Transformer-based language models, especially sentence text-to-text transformers (ST5)-XXL and XLNet. The main purpose of this study is to compare the performance of the two models, identify the strengths and weaknesses of each, and determine the optimal learning rate to increase the efficiency and stability of model training. We performed data preprocessing, constructed and trained models, and evaluated performance based on data sets with various characteristics. We confirmed that the XLNet model showed excellent performance at learning rates of 1e-05 and 1e-04 and had a significantly lower loss value than the ST5-XXL model. This indicates that XLNet is more efficient for learning. Additionally, we confirmed in our study that learning rate has a significant impact on model performance. The results of the study highlight the usefulness of ST5-XXL and XLNet models in the task of classifying security vulnerabilities and highlight the importance of setting an appropriate learning rate. Future research should include more comprehensive analyzes using diverse data sets and additional models.

• Proceedings of the IEEK Conference
• /
• 1998.10a
• /
• pp.1273-1276
• /
• 1998

The HMM-Net is an architecture for a neural network that implements a hidden Markov model (HMM). The architecture is developed for the purpose of combining the discriminant power of neural networks with the time-domain modeling capability of HMMs. Criteria used for learning HMM-Net classifiers are maximum likelihood (ML), maximum mutual information (MMI), and minimization of mean squared error(MMSE). In this paper we propose an efficient learning method of HMM-Net classifiers using hybrid criteria, ML/MMSE and MMI/MMSE, and report the results of an experimental study comparing the performance of HMM-Net classifiers trained by the gradient descent algorithm with the above criteria. Experimental results for the isolated numeric digits from /0/ to /9/ show that the performance of the proposed method is better than the others in the respects of learning and recognition rates.

• Proceedings of the IEEK Conference
• /
• 1998.06a
• /
• pp.933-935
• /
• 1998

The HMM-Net is an architecture for a neural network that implements a hidden markov model (HMM). The architecture is developed for the purpose of combining the discriminant power of neural networks with the time-domain modeling capability of HMMs. Criteria used for learning HMM-Net classifiers are maximum likelihood(ML) and minimization of mean squared error(MMSE). In this paper we propose an efficient learning method of HMM_Net classifiers using a ML-MMSE hybrid criterion and report the results of an experimental study comparing the performance of HMM_Net classifiers trained by the gradient descent algorithm with the above criteria. Experimental results for the isolated numeric digits from /0/ to /9/ show that the performance of the proposed method is better than the others in the repects of learning and recognition rates.

• Proceedings of the KIEE Conference
• /
• 1992.07a
• /
• pp.480-483
• /
• 1992

Although the back-propagation(BP) algorithm is widely used for its simple structure and easy learning method, it has a drawback of slow convergence rate. In this paper, we propose an algorithm to improve this problem by manipulating the slope parameter of the activation function. The steepest descent method is used in learning the slope parameter, as in the case of weight. The simulation shows that the learning rates of the proposed algorithm is faster than the conventional BP algorithm.

• Proceedings of the Korean Nuclear Society Conference
• /
• 1995.10a
• /
• pp.293-299
• /
• 1995

A methodology is applied to identify tile learning trend related to the safety and availability of U.S. commercial nuclear power plants. The application is intended to aid in reducing likelihood of human errors. To assure that tile methodology ran be easily adapted to various types of classification schemes of operation data, a data bank classified by the Transient Analysis Classification and Evaluation(TRACE) scheme is selected for the methodology. The significance criteria for human-initiated events affecting tile systems and for events caused by human deficiencies were used. Clustering analysis was used to identify the learning trend in multi-dimensional histograms. A computer rode is developed based on tile K-Means algorithm and applied to find the learning period in which error rates are monotonously decreasing with plant age.

• Proceedings of the KOSOMBE Conference
• /
• v.1997 no.11
• /
• pp.44-47
• /
• 1997

This paper is a study on the hand gesture recognition using Instance-based teaming, Symbolic learning algorithms and Power Glove which supplies information on finger position, hand position and orientation. The data were carefully examined, and a few features of the data that would serve as good discriminants between signs when used with the learning algorithms were extracted. The hand gesture data collected from 5 people were applied to the teaming algorithms. In spite of the noise and accuracy constraints of the equipment used, some accuracy rates were achieved.