• The KIPS Transactions:PartB
• /
• v.11B no.4
• /
• pp.491-500
• /
• 2004

This paper presents an ontology construction method using various computational language resources, and an ontology-based word sense disambiguation method. In order to acquire a reasonably practical ontology the Kadokawa thesaurus is extended by inserting additional semantic relations into its hierarchy, which are classified as case relations and other semantic relations. To apply the ontology to disambiguate word senses, we apply the previously-secured dictionary information to select the correct senses of some ambiguous words with high precision, and then use the ontology to disambiguate the remaining ambiguous words. The mutual information between concepts in the ontology was calculated before using the ontology as knowledge for disambiguating word senses. If mutual information is regarded as a weight between ontology concepts, the ontology can be treated as a graph with weighted edges, and then we locate the weighted path from one concept to the other concept. In our practical machine translation system, our word sense disambiguation method achieved a 9% improvement over methods which do not use ontology for Korean translation.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.26 no.2
• /
• pp.405-413
• /
• 2016

In this paper, we propose a proactive inference method of finding suspicious domains. Our method detects potential malicious domains from the seed domain information extracted from the TLD Zone files and WHOIS information. The inference process follows the three steps: searching the candidate domains, machine learning, and generating a suspicious domain pool. In the first step, we search the TLD Zone files and build a candidate domain set which has the same name server information with the seed domain. The next step clusters the candidate domains by the similarity of the WHOIS information. The final step in the inference process finds the seed domain's cluster, and make the cluster as a suspicious domain set. In experiments, we used .COM and .NET TLD Zone files, and tested 10 seed domains selected by our analysts. The experimental results show that our proposed method finds 55 suspicious domains and 52 true positives. F1 scores 0.91, and precision is 0.95 We hope our proposal will contribute to the further proactive malicious domain blacklisting research.

• Journal of Biosystems Engineering
• /
• v.31 no.1 s.114
• /
• pp.59-64
• /
• 2006

Application of herbicide for rice cropping is inevitable but notorious for its side effect of environmental pollution. Precision fanning will be one of important tools for the least input and sustainable fanning and could be achieved by implementation of the variable rating technology. If a device to detect weeds in rice field is available, herbicide could be applied only to the places where it is needed by the manner of the variable rating technology. The study was carried out to develop an algorithm of image processing to detect weeds in rice field using a machine vision system of multi-spectral digital images. A series of multi-spectral rice field picture of 560, 680 and 800 nm of center wavelengths were acquired from the 27th day to the 39th day after transplanting in the ineffective tillering stage of a rice growing period. A discrimination model to distinguish pixels of weeds from those of rice plant and weed image was developed. The model was proved as having accuracies of 83.6% and 58.9% for identifying the rice plant and the weed, respectively. The model was used in the algorithm to differentiate weed images from mingled images of rice plant and weed in a frame of rice field picture. The developed algorithm was tested with the acquired rice field pictures and resulted that 82.7%, 11.9% and 5.4% of weeds in the pictures were noted as the correctly detected, the undetected and the misclassified as rice, respectively, and 81.9% and 18.0% of rice plants in the pictures were marked as the correctly detected and the misclassified as weed, respectively.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.3
• /
• pp.621-629
• /
• 2016

This paper presents a decision method of middle ear disease which is developed in children and adults. In the proposed method, features are extracted from the middle ear disease images and normal images using HoG (histogram of oriented gradient) descriptor and the extracted features are learned by SVM (support vector machine) classifier. To obtain an input vector into SVM, an input image is resized to a predefined size and then the resized image is partitioned into 16 blocks each of which is partitioned into 4 sub-blocks (namely cell). Finally, the feature vector with 576 components is given by using HoG with 9 bins and it is used as SVM learning and classification. Input images are classified by SVM classifier based on the model of learning features. Experimental results show that the proposed method yields the precision of over 90% in decision.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.649-652
• /
• 2005

The camera cellular phone has a large portion of cellular phone market in recent year. The variety of a customer demand makes a fast model change and the spatial resolution is changed from VGA to multi-mega pixel. The 1.3 mega pixel (MP) camera cellular phone was first released into the Korean market in October 2003. The major cellular phone companies released a 2MP camera cellular phone that supports zoom function and a 2MP camera cellular phone is settled down with the Korea cellular phone market. It makes a keen competition in price and demands automation for phone camera module. There is an increasing requirement for the automatic assembly to correspond to a fast model change. The hard automation techniques that rely on dedicated manufacturing system are too inflexible to meet this requirement. Therefore in this study, this system is designed with the flexibility concept in order to cope with phone camera module change. The system has a same platform that has X-Y-Z motion or X-Z motion with ${\mu}m$order accuracy. It has a special gripper according to the type of a component to be put together. If the camera model changes, the gripper may be updated to fit for the camera module. The controller of this system acquires the data sets that have the information about the assembly part by the tray. This information is obtained ahead of an inspection step. The controller excludes an inferior part to be assembled by using this information to diminish the inferior goods. The assembly jig used in this system has a function of self adjustment that reduces the tact time and also diminish the inferior goods. Finally, the intelligent assembly system for phone camera module will be designed to get a flexibility to meet model change and a high productivity with a high reliability.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.6 no.2
• /
• pp.364-370
• /
• 2002

A banknote counter is a machine that automates counting the money in some agencies to treat much banknotes as well as general banking agencies. The banknote counter materialized in this paper is the machine that adds the function of banknote sorting, detecting plural banknote and detecting counterfeit banknote to an existing banknote counter. The technique of sensor signal processing are used for banknote sorting. The technique of sensor application and data processing are used for detecting counterfeit banknote. The technique of precision equipment design and microprocessor application are used for high speed count. Software improved in debugging and difficulties to link with additional hardware. It was materialized through effective control algorithm and real-time signal processing with C-language on the basis of RTOS(real-time operating system) Photodiode, its applications and a magnetic resistance sensor are used as a sensor device with regard to hardware cost -cutting and process velocity. PCF80C552-24 of Philips using Intel I8051 core is used as a control microprocessor. As the results so far achieved, counterfeit banknotes made by the use of a color duplicator and a color Printer, are distinguished from real banknotes through mixing an optical with a magnetic sensor. and, in case that there are some different banknotes while counting, it is prevented for them to be counted without discriminating from the same kind of banknotes in addition to the fu notion of banknote sorting.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.14 no.3
• /
• pp.15-20
• /
• 2015

After the US government declared 3D printing technology a next-generation manufacturing technology, there have been many practical studies conducted to expand 3D printing technology to manufacturing technologies, called AMERICA MAKES. In particular, the Keck Center, located at the University of Texas at El Paso, has studied techniques for easily combing the 3D stacking process with space mobility and expanded these techniques to simultaneous staking techniques for multiple materials. Additionally, it developed convergence manufacturing techniques, such as direct inking techniques, in order to produce a module structure that combines electronic circuits and components, such as CUBESET. However, in these studies, it is impossible to develop a unified system using traditional independent through simple sequencing connections. This is because there are many problems in the integration between the stacking modeling of 3D printers and post-machining, such as thermal deformations, the precision accuracy of 3D printers, and independently driven coordinate problems among process systems. Therefore, in this paper, the integration method is suggested, which combines these 3D printers and subsequent machining process systems through an Internet-based cloud. Additionally, the sequential integrated system of a 3D printer, an NC milling machine, machine vision, and direct inking are realized.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.12
• /
• pp.1-7
• /
• 2019

With the in-depth development of machine learning, the deep learning method has made great progress, especially with the Convolution Neural Network(CNN). Compared with traditional text sentiment classification methods, deep learning based CNNs have made great progress in text classification and processing of complex multi-label and multi-classification experiments. However, there are also problems with the neural network for text sentiment classification. In this paper, we propose a fusion model based on Long-Short Term Memory networks(LSTM) and CNN deep learning methods, and applied to multi-category news datasets, and achieved good results. Experiments show that the fusion model based on deep learning has greatly improved the precision and accuracy of text sentiment classification. This method will become an important way to optimize the model and improve the performance of the model.

• Journal of Advanced Navigation Technology
• /
• v.16 no.1
• /
• pp.151-157
• /
• 2012

Recently, owing to the technology advances in computer science and image processing, age of face classification have become prevalent topics. It is difficult to estimate age of facial shape with statistical figures because facial shape of the person should change due to not only biological gene but also personal habits. In this paper, we proposed a robust age of face classification method by using Gabor feature and fuzzy support vector machine(SVM). Gabor wavelet function is used for extracting facial feature vector and in order to solve the intrinsic age ambiguity problem, a fuzzy support vector machine(FSVM) is introduced. By utilizing the FSVM age membership functions is defined. Some experiments have conducted to testify the proposed approach and experimental results showed that the proposed method can achieve better age of face classification precision.

• Korean Journal of Applied Biomechanics
• /
• v.19 no.3
• /
• pp.509-517
• /
• 2009

The purpose of this study was to investigate the effective posture for air pistol shooting. Participants were 3 male athletes of shooting with at least five years of experience and another group of 3 males athletes with less than three years of experience. For the purpose, the shooting motion was analysed using three dimensional image technology. Data from each event for the two groups, competent and less competent ones, were compared to see the differences from the kinematical point of view. Time of period in competent group was longer than less competent group during the shooting posture. Displacement of center of mass and pistol about medial/lateral and antero/posterior in competent group was little than less competent group from aim to shooting. And these result were effect to the velocity. Distance and time in competent group within coaching machine were smaller than less competent group. To the result, it was appear that precision of aim in competent group was higher than less competent group.