• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.17 no.9
• /
• pp.292-301
• /
• 2016

Because damages arising from the occurrence of foot-and-mouth disease (FMD) are very great, it is essential to make a preemptive diagnosis to cope with it in order to minimize those damages. The main symptoms of foot-and-mouth disease are body temperature increase, loss of appetite, formation of blisters in the mouth, on hooves and breasts, etc. in a cow or a bull, among which the body temperature check is the easiest and quickest way to detect the disease. In this paper, an algorithm to detect FMD from the hooves of cattle was developed and implemented for preemptive coping with foot-and-mouth disease, and a hoof check test is conducted after the installation of a high-resolution camera module, a thermo-graphic camera, and a temperature/humidity module in the cattle shed. Through the algorithm and system developed in this study, it is possible to cope with an early-stage situation in which cattle are suspected as suffering from foot-and-mouth disease, creating an optimized growth environment for cattle. In particular, in this study, the system to cope with FMD does not use a portable thermo-graphic camera, but a fixed camera attached to the cattle shed. It does not need additional personnel, has a function to measure the temperature of cattle hooves automatically through an image algorithm, and includes an automated alarm for a smart phone. This system enables the prediction of a possible occurrence of foot-and-mouth disease on a real-time basis, and also enables initial-stage disinfection to be performed to cope with the disease without needing extra personnel.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.12
• /
• pp.29-40
• /
• 2022

In this paper, we studied a system that detects and analyzes the pathological features of diabetic retinopathy using Mask R-CNN and a Random Forest classifier. Those are one of the deep learning techniques and automatically diagnoses diabetic retinopathy. Diabetic retinopathy can be diagnosed through fundus images taken with special equipment. Brightness, color tone, and contrast may vary depending on the device. Research and development of an automatic diagnosis system using artificial intelligence to help ophthalmologists make medical judgments possible. This system detects pathological features such as microvascular perfusion and retinal hemorrhage using the Mask R-CNN technique. It also diagnoses normal and abnormal conditions of the eye by using a Random Forest classifier after pre-processing. In order to improve the detection performance of the Mask R-CNN algorithm, image augmentation was performed and learning procedure was conducted. Dice similarity coefficients and mean accuracy were used as evaluation indicators to measure detection accuracy. The Faster R-CNN method was used as a control group, and the detection performance of the Mask R-CNN method through this study showed an average of 90% accuracy through Dice coefficients. In the case of mean accuracy it showed 91% accuracy. When diabetic retinopathy was diagnosed by learning a Random Forest classifier based on the detected pathological symptoms, the accuracy was 99%.

• Proceedings of the Korean Information Science Society Conference
• /
• 2007.06a
• /
• pp.17-18
• /
• 2007

데이터베이스 튜닝은 일반적으로 데이터베이스 어플리케이션을 "좀 더 빠르게" 실행하게 하는 일련의 활동을 뜻한다[1]. 데이터베이스 관리자가 튜닝에 필요한 주먹구구식 룰(Rule of thumb)들을 모두 파악 하고 상황에 맞추어 적용하는 것은 비싼 비용과 오랜 시간을 요구한다. 그렇게 때문에 서로 다른 어플 리케이션들이 맞물려 있는 복잡한 서비스는 필수적으로 자동화된 데이터베이스 성능 관리와 튜닝을 필 요로 한다. 본 논문에서는 이를 해결하기 위하여 지식 도매인(Knowledge Domain)을 기초로 한 자동화 된 데이터베이스 튜닝 원칙(Tuning Principle)을 제시하는 시스템을 제안한다. 각각의 데이터베이스 튜닝 이론들은 지식 도매인의 지식으로 활용되며, 성능에 영향을 미치는 요소들을 개체(Object)와 콘셉트 (Concept)로 구성하고 추론 시스템을 통해 튜닝 원칙을 추론하여 쉽고 빠르게 현재 상황에 맞는 튜닝 방법론을 적용시킬 수 있다. 자동화된 데이터베이스 튜닝에 대해 여러 분야에 걸쳐 학문적인 연구가 이루어지고 있다. 그 예로써 Microsoft의 AutoAdmin Project[2], Oracle의 SQL 튜닝 아키텍처[3], COLT[4], DBA Companion[5], SQUASH[6] 등을 들 수 있다. 이러한 최적화 기법들을 각각의 기능적인 방법론에 따라 다시 분류하면 크게 Design Tuning, Logical Structure Tuning, Sentence Tuning, SQL Tuning, Server Tuning, System/Network Tuning으로 나누어 볼 수 있다. 이 중 SQL Tuning 등은 수치적으로 결정되어 이미 존재하는 정보를 이용하기 때문에 구조화된 모델로 표현하기 쉽고 사용자의 다양한 요구에 의해 변화하는 조건들을 수용하기 쉽기 때문에 이에 중점을 두고 성능 문제를 해결하는 데 초점을 맞추었다. 데이터베이스 시스템의 일련의 처리 과정에 따라 DBMS를 구성하는 개체들과 속성, 그리고 연관 관계들이 모델링된다. 데이터베이스 시스템은 Application / Query / DBMS Level의 3개 레벨에 따라 구조화되며, 본 논문에서는 개체, 속성, 연관 관계 및 데이터베이스 튜닝에 사용되는 Rule of thumb들을 분석하여 튜닝 원칙을 포함한 지식의 형태로 변환하였다. 튜닝 원칙은 데이터베이스 시스템에서 발생하는 문제를 해결할 수 있게 하는 일종의 황금률로써 지식 도매인의 바탕이 되는 사실(Fact)과 룰(Rule) 로써 표현된다. Fact는 모델링된 시스템을 지식 도매인의 하나의 지식 개체로 표현하는 방식이고, Rule 은 Fact에 기반을 두어 튜닝 원칙을 지식의 형태로 표현한 것이다. Rule은 다시 시스템 모델링을 통해 사전에 정의되는 Rule와 튜닝 원칙을 추론하기 위해 사용되는 Rule의 두 가지 타업으로 나뉘며, 대부분의 Rule은 입력되는 값에 따라 다른 솔루션을 취하게 하는 분기의 역할을 수행한다. 사용자는 제한적으로 자동 생성된 Fact와 Rule을 통해 튜닝 원칙을 추론하여 데이터베이스 시스템에 적용할 수 있으며, 요구나 필요에 따라 GUI를 통해 상황에 맞는 Fact와 Rule을 수동으로 추가할 수도 었다. 지식 도매인에서 튜닝 원칙을 추론하기 위해 JAVA 기반의 추론 엔진인 JESS가 사용된다. JESS는 스크립트 언어를 사용하는 전문가 시스템[7]으로 선언적 룰(Declarative Rule)을 이용하여 지식을 표현 하고 추론을 수행하는 추론 엔진의 한 종류이다. JESS의 지식 표현 방식은 튜닝 원칙을 쉽게 표현하고 수용할 수 있는 구조를 가지고 있으며 작은 크기와 빠른 추론 성능을 가지기 때문에 실시간으로 처리 되는 어플리케이션 튜닝에 적합하다. 지식 기반 모률의 가장 큰 역할은 주어진 데이터베이스 시스템의 모델을 통하여 필요한 새로운 지식을 생성하고 저장하는 것이다. 이를 위하여 Fact와 Rule은 지식 표현 의 기본 단위인 트리플(Triple)의 형태로 표현된다, 트리플은 Subject, Property, Object의 3가지 요소로 구성되며, 대부분의 Fact와 Rule들은 트리플의 기본 형태 또는 트리플의 조합으로 이루어진 C Condition과 Action의 두 부분의 결합으로 구성된다. 이와 같이 데이터베이스 시스템 모델의 개체들과 속성, 그리고 연관 관계들을 표현함으로써 지식들이 추론 엔진의 Fact와 Rule로 기능할 수 있다. 본 시스템에서는 이를 구현 및 실험하기 위하여 웹 기반 서버-클라이언트 시스템을 가정하였다. 서버는 Process Controller, Parser, Rule Database, JESS Reasoning Engine으로 구성 되 어 있으며, 클라이 언트는 Rule Manager Interface와 Result Viewer로 구성되어 었다. 실험을 통해 얻어지는 튜닝 원칙 적용 전후의 실행 시간 측정 등 데이터베이스 시스템 성능 척도를 비교함으로써 시스템의 효용을 판단하였으며, 실험 결과 적용 전에 비하여 튜닝 원칙을 적용한 경우 최대 1초 미만의 전처리에 따른 부하 시간 추가와 최소 약 1.5배에서 최대 약 3배까지의 처리 시간 개선을 확인하였다. 본 논문에서 제안하는 시스템은 튜닝 원칙을 자동으로 생성하고 지식 형태로 변형시킴으로써 새로운 튜닝 원칙을 파생하여 제공하고, 성능에 영향을 미치는 요소와 함께 직접 Fact과 Rule을 추가함으로써 커스터마이정된 튜닝을 수행할 수 있게 하는 장점을 가진다. 추후 쿼리 자체의 튜닝 및 인텍스 최적화 등의 프로세스 자동화와 Rule을 효율적으로 정의하고 추가하는 방법 그리고 시스템 모델링을 효과적으로 구성하는 방법에 대한 연구를 통해 본 연구를 더욱 개선시킬 수 있을 것이다.

• Journal of Institute of Control, Robotics and Systems
• /
• v.21 no.12
• /
• pp.1130-1135
• /
• 2015

The bond pull test is the most widely used technique for the evaluation and control of wire bond quality. The wire being tested is pulled upward until the wire or bond to the die or substrate breaks. The inspector test strength of wire by manually and it takes around 3 minutes to perform the test. In this paper, we develop a 3D vision system to measure 3D position of wire. It gives 3D position data of wire to move a hook into wires. The 3D measurement method to use here is a confocal imaging system. The conventional confocal imaging system is a spot scanning method which has a high resolution and good illumination efficiency. However, a conventional confocal systems has a disadvantage to perform XY axis scanning in order to achieve 3D data in given FOV (Field of View) through spot scanning. We propose a method to improve a parallel mode confocal system using a micro-lens and pin-hole array to remove XY scan. 2D imaging system can detect 2D location of wire and it can reduce time to measure 3D position of wire. In the experimental results, the proposed system can measure 3D position of wire with reasonable accuracy.

• Journal of KIISE
• /
• v.43 no.12
• /
• pp.1325-1333
• /
• 2016

Software classification has several applications such as copyright infringement detection, malware classification, and software automatic categorization in software repositories. It can be also employed by software filtering systems to prevent the transmission of illegal software. If illegal software is identified by measuring software similarity in software filtering systems, the average number of comparisons can be reduced by shrinking the search space. In this study, we focused on the classification of Windows executables using API call information and machine learning. We evaluated the classification performance of machine learning-based classifier according to the refinement method for API information and machine learning algorithm. The results showed that the classification success rate of SVM (Support Vector Machine) with PolyKernel was higher than other algorithms. Since the API call information can be extracted from binary executables and machine learning-based classifier can identify tampered executables, API call information and machine learning-based software classifiers are suitable for software filtering systems.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2017.11a
• /
• pp.206-208
• /
• 2017

In recent years, we have been working on the improvement project of mercury 3-lecture type at the Geummundo lighthouse for more than 100 years. We have installed toxic mercury removal, LED light source, and dualization system for nonstop. It can be seen as an example. Circular restoration of the entire frame is performed, and the real-time status can be grasped through the installation of the double-ball bearing and the remote control system. After installation, optical measurement through Hanvit secured the optical performance of 2,000,000 cd of floating brightness of 300cd, proving the efficient possibility of introducing the domestic rotary type light source using LED light source.

• Journal of Navigation and Port Research
• /
• v.41 no.3
• /
• pp.87-92
• /
• 2017

To enhance the reliability and/or resilience of the PNT service included in the e-Navigation strategy of the IMO, the evaluation of the AIS-TWR method for the asynchronous R-mode for maritime service, which is available even in the absence of the GNSS, is described. For the AIS-TWR, which is capable of ranging through message exchange even without high precision synchronization, the operation scenario and the error factors according to the AIS system specifications are proposed and analyzed. Cramer-Rao Lower Bound is presented for the performance evaluation of the AIS-TWR algorithm. A simulation by AIS-TWR method of two AIS systems in a 3 km static environment shows estimation error of about 41m compared to the real value..

• The KIPS Transactions:PartB
• /
• v.14B no.7
• /
• pp.481-492
• /
• 2007

In this paper, a method for automatic indexing of contents to reduce effort for inputting paper information and constructing index is considered. Existing document analysis methods can't analyse various table of contents of journal paper formats efficiently because they have many exceptions. In this paper, various contents formats for journals, which have different features from those for general documents, are analysed and described. The principal elements that we want to represent are titles, authors, and pages for each papers. Thus, the three principal elements are modeled according to the order of their arrangement, and their features are extracted. And a table of content recognition system of journal is implemented, based on the proposed modeling method. The accuracy of exact extraction ratio of 91.5% on title, author, and page type on 660 published papers of various journals is obtained.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.26 no.9
• /
• pp.821-827
• /
• 2015

Electromagnetic waves which are emitted from the Sun due to solar flare explosion can cause failures in HF radio communications in the day-side area of the Earth, that is so-call as Radio Blackouts. The international scale representing the severity of the Radio Blackouts is determined by the solar X-ray flux which is measured by United States Geostationary Operational Environmental Satellite. However, the scale is not always applicable to HF communication users in the different area on the Earth, because the HF communication effects depend not only on the X-ray strength but also on the subsolar point location. To solve this problem, we developed a HF radio spectrum monitoring system utilizing a spectrum analyzer. This system conducts a real-time measure of the HF spectrum, and automatically calculates signal to noise ratios and the occurrences of the HF blackouts as comparing with the interference level which is described from the ITU recommendation.

• Korean Journal of Cognitive Science
• /
• v.13 no.3
• /
• pp.11-21
• /
• 2002

Parke and Waters' model［1］ of muscle-based face deformation was used to develop a system that can synthesize facial expressions when the pleasure-displeasure and arousal-sleep coordinate values of internal states are specified. Facial expressions sampled from a database developed by Chung, Oh, Lee and Byun ［2］ and its underlying model of internal states were used to find rules for face deformation. The internal - state model included dimensional and categorical values of the sampled facial expressions. To find out deformation rules for each of the expressions, changes in the lengths of 21 facial muscles were measured. Then, a set of multiple regression analyses was performed to find out the relationship between the muscle lengths and internal states. The deformation rules obtained from the process turned out to produce natural-looking expressions when the internal states were specified by the pleasure-displeasure and arousal-sleep coordinate values. Such a result implies that the rules derived from a large scale database and regression analyses capturing the variations of individual muscles can be served as a useful and powerful tool for synthesizing facial expressions.