• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2003.10a
• /
• pp.797-800
• /
• 2003

In the current study, the authors propose a method for extracting license plate regions by means of a neural network trained to output the plates 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 revered by the learning pattern, the effort 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%.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.34 no.2
• /
• pp.111-120
• /
• 2016

Prior knowledge about the SNS (Social Network Services) datasets is often required to conduct valuable analysis using social media data. Understanding the characteristics of the information extracted from SNS datasets leaves much to be desired in many ways. This paper purposes on analyzing the detail of the target social network services, Twitter, Instagram, and YouTube to establish the spatial informatization process to integrate social media information with existing spatial datasets. In this study, valuable information in SNS datasets have been selected and total 12,938 data have been collected in Seoul via Open API. The dataset has been geo-coded and turned into the point form. We also removed the overlapped values of the dataset to conduct spatial integration with the existing building layers. The resultant of this spatial integration process will be utilized in various industries and become a fundamental resource to further studies related to geospatial integration using social media datasets.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37A no.9
• /
• pp.780-785
• /
• 2012

In the current study, the authors propose a method for extracting license plate regions by means of a neural network trained to output the plates 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%.

• Convergence Security Journal
• /
• v.8 no.2
• /
• pp.71-78
• /
• 2008

This research is on the study of integrating internet and intranet that was built for their own enterprise into one network system that promises a more economic and secure use. Compared to this system, the traditional network conversion device not only has its mechanical limits, but also requires a separate network set up. This raises both interoperability and security problems and results in a higher cost. Therefore, I propose dual defence system based on the inverse multiplexing of ATM Forum and IPSec to improve network ability and deliver enhanced system reliability while reducing cost. Furthermore, I also addressed some of the weaknesses of the Scanning Attacking Method, SNMP and Spooler Port and proposed counter measures that will deal with these weakness at the dual defence system.

• 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%.

• Journal of KIISE:Computing Practices and Letters
• /
• v.5 no.5
• /
• pp.619-627
• /
• 1999

To support multimedia services between network domains with different environments, it is required to map the functionalities in many aspects. In this paper, we implemented interworking gateway which provides protocol conversion and QoS(Quality of Service) adaptation to interwork DAVIC services based on ATM(Asynchronous TRansfer Model )network and Internet AV services. The interworking gateway converts RTSP(Real-Time Streaming Protocol ) message into DSM-CC(Digital Storage Media Command & Control) messages to control the stream that is served in ATM network, and transmits data stream by using RTP(Real-Time Transport Protocol) The interworking gateway provides QoS adaptation functionalities by QoS monitoring and MPEG filtering to meet the variation of network bandwidth.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.19 no.11
• /
• pp.2584-2589
• /
• 2015

In this paper, we want to able to reduce the time effectively about the analysis of protocol and development of protocol converter in industrial network. For this purpose, we developed the user protocol converter that can be to analysing the communication protocol in the industrial network. and also that is able to converting to user protocol and transmitting to another network. The developed user protocol converter was consist of S/W and H/W, and is supporting the Modbus, the NMEA0183 and the ethernet. The hardware main processor was used the AVR128, and placed the serial communication part and the ethernet part in converter board. The experiment results, we acquired very high transmission rate and conversion rate.

• Journal of Broadcast Engineering
• /
• v.27 no.6
• /
• pp.885-896
• /
• 2022

Motor failure in manufacturing plays an important role in future A/S and reliability. Motor failure is detected by measuring sound, current, and vibration. For the data used in this paper, the sound of the car's side mirror motor gear box was used. Motor sound consists of three classes. Sound data is input to the network model through a conversion process through MelSpectrogram. In this paper, various methods were applied, such as data augmentation to improve the performance of classifying fault motors and various methods according to class imbalance were applied resampling, reweighting adjustment, change of loss function and representation learning and classification into two stages. In addition, the curriculum learning method and self-space learning method were compared through a total of five network models such as Bidirectional LSTM Attention, Convolutional Recurrent Neural Network, Multi-Head Attention, Bidirectional Temporal Convolution Network, and Convolution Neural Network, and the optimal configuration was found for motor sound classification.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.24 no.2
• /
• pp.69-80
• /
• 1999

A genetic algorithm to determine the optimal signal sequence and double cycle pattern is described. The signal sequence and double cycle pattern are used as the input for TRANSYT to find optimal signal timing at each junction in the area traffic networks, In the genetic process, the partially matched crossover and simple crossover operators are used for evolution of signal sequence and double cycle pattern respectively. A special conversion algorithm is devised to convert the signal sequence into the link-stage assignment for TRANSYT. Results from tests using data from an area traffic network in Leicester region R are given.

• Proceedings of the IEEK Conference
• /
• 2003.11c
• /
• pp.47-50
• /
• 2003

This paper describes the design and implementation of a HNCP-UPnP bridge. HNCP networked devices targeted to white goods are allowed to be controlled and monitored by UPnP Control Point. Bridge provides automated creation of UPnP proxy device and message conversion. The developed system enables integration of heterogeneous home network.