• Journal of Broadcast Engineering
• /
• v.27 no.1
• /
• pp.133-145
• /
• 2022

Compared to the continuously increasing dog population and industry size in Korea, systematic analysis of related data and research on breed classification methods are very insufficient. In this paper, an automatic breed classification method is proposed using deep learning technology for 14 major dog breeds domestically raised. To do this, dog images are collected for deep learning training and a dataset is built, and a breed classification algorithm is created by performing transfer learning based on VGG-16 and Resnet-34 as backbone networks. In order to check the transfer learning effect of the two models on dog images, we compared the use of pre-trained weights and the experiment of updating the weights. When fine tuning was performed based on VGG-16 backbone network, in the final model, the accuracy of Top 1 was about 89% and that of Top 3 was about 94%, respectively. The domestic dog breed classification method and data construction proposed in this paper have the potential to be used for various application purposes, such as classification of abandoned and lost dog breeds in animal protection centers or utilization in pet-feed industry.

• Journal of the Korea Computer Industry Society
• /
• v.2 no.11
• /
• pp.1427-1438
• /
• 2001

We design the evaluation factors to represent the relevance distribution information between a query and sources and proposes the scheme to get relevance distribution information based on evaluation factors. Then it is developed that the organization is able to classify the best source toward query, and shown the algorithms that is able to select the best source toward users query, it is developed algorithms that is decided ranking and mering these, after choose the best source to evaluate a query, Finally, it merges the result from the source, and present them to the user to the issued query. This paper also develops the scheme to classify the best sources for query and presents the selection algorithm of the best information sources. Finally the ranking and merging Federated Retrieval System is presented.

• Journal of the Korean Institute of Gas
• /
• v.6 no.4 s.18
• /
• pp.1-7
• /
• 2002

Human errors can take place in all levels that include the design, production, construction, operation and maintenance of plant facilities. It was found that the causes were concerned with the effects of human error. This study verified characteristics of the on-site operators and error mechanism, and used the classifying sheet to analyze human error that occurred in process. Also, by applying the ASEP(Accident Sequence Evaluation Program) HRA(Human Reliability Analysis) procedure, the algorithm to estimate the HEP and the ASEP HEP program to analyze human error in the plant were developed. If it is built in on-site, possible human error incident will be prevented and the systematic human error prevention strategy will be devised.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2010.07a
• /
• pp.293-296
• /
• 2010

현재 지능형 서비스로봇이 활성화되고 있으며 지능형 로봇 분야에서 청소로봇은 개인생활의 본질적인 운동성을 보조하며 인간과 공생의 형태를 지원하는 로봇으로 성장하고 있다. 청소로봇은 가정에서 다양한 서비스를 제공하기 위하여 가정환경과 적절한 상호작용은 필수적이다. 따라서 로봇이 외부의 서버와 협동을 통한 원격제어는 필수적이라 할 수 있다. 현재 지도 작성이 가능한 청소로봇은 바닥에 센서를 부착하거나 작업환경에 카메라 및 기타 많은 장치를 구축하므로 고가의 장비로 분류되어 일반적으로 활용지 못하고 있다. 본 연구에서는 청소로봇의 효율적인 제어를 위하여 작업환경에서 로봇의 궤적과 센서정보를 원격으로 서버에 전달하고 서버는 이러한 정보를 바탕으로 지도를 작성하고 최적의 작업을 수행할 수 있는 알고리즘을 연산하여 로봇에 전달할 수 있는 방법을 연구하였다. 이는 로봇이 지도 작성을 위한 연산 과정을 줄이므로 수행 시간을 단축하고 로봇에 주어지는 부하를 줄일 수 있다. 또한 원격으로 제어를 수행하므로 무선통신이 가능한 각종 디바이스와 상호 연동할 수 있는 체계를 만들어 유비쿼터스 환경을 구축할 수 있는 계기를 만들므로 청소로봇의 성능이 향상될 수 있음을 보인다. 본 연구에서는 청소로봇을 제작하고 시뮬레이션과 실험을 통하여 전체적으로 주어진 작업을 효율적으로 수행할 수 있음을 보여준다.

• Journal of Korean Society of Transportation
• /
• v.19 no.6
• /
• pp.161-170
• /
• 2001

Vehicle detection is the basic of traffic monitoring. Video based systems have several apparent advantages compared with other kinds of systems. However, In video based systems, shadows make troubles for vehicle detection. especially active shadows resulted from moving vehicles. In this paper a new method that combines background subtraction and edge detection is proposed for vehicle detection and shadow rejection. The method is effective and the correct rate of vehicle detection is higher than 98(%) in experiments, during which the passive shadows resulted from roadside buildings grew considerably. Based on the proposed vehicle detection method, vehicle tracking, counting, classification and speed estimation are achieved so that traffic information concerning traffic flow is obtained to describe the load of each lane.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.8 no.1
• /
• pp.9-21
• /
• 2009

This study presents a useful heuristic algorithm to classify vehicle classes using vehicle length information, which is extracted from inductive loop vehicle signatures. A high-speed scanning equipment was used to extract more detailed change of inductance magnitude for individual vehicles. Vehicle detection time and individual vehicle speeds were used to derive vehicle length information that is an input of the proposed algorithm. The spatial and temporal transferability tests were further conducted to evaluate algorithm. The spatial and temporal transferability tests were further conducted to evaluate algorithm performance more systematically. It is expected that the proposed method would be useful for obtaining vehicle classification information from wide-spread existing loop infrastructure.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.19 no.4
• /
• pp.13-29
• /
• 2020

Automobiles have long been an essential part of daily life, but the social costs of car traffic accidents exceed 9% of the national budget of Korea. Hence, it is necessary to establish prevention and response system for car traffic accidents. In order to present a model that can classify and predict the degree of injury in car traffic accidents, we used big data analysis techniques of K-nearest neighbor, logistic regression analysis, naive bayes classifier, decision tree, and ensemble algorithm. The performances of the models were analyzed by using the data on the nationwide traffic accidents over the past three years. In particular, considering the difference in the number of data among the respective injury severity levels, we used down-sampling methods for the group with a large number of samples to enhance the accuracy of the classification of the models and then verified the statistical significance of the models using ANOVA.

• Journal of Korean Society of Transportation
• /
• v.21 no.2
• /
• pp.63-70
• /
• 2003

Real-time information on vehicle queue at intersections is essential for optimal traffic signal control, which is substantial part of Intelligent Transport Systems (ITS). Computer vision is also potentially an important element in the foundation of integrated traffic surveillance and control systems. The objective of this research is to propose a method for detecting an exact queue lengths at signalized intersections using image processing techniques and a neural network model Fuzzy ARTMAP, which is a supervised and self-organizing system and claimed to be more powerful than many expert systems, genetic algorithms. and other neural network models like Backpropagation, is used for recognizing different patterns that come from complicated real scenes of a car park. The experiments have been done with the traffic scene images at intersections and the results show that the method proposed in the paper could be efficient for the noise, shadow, partial occlusion and perspective problems which are inevitable in the real world images.

• Journal of Digital Convergence
• /
• v.13 no.9
• /
• pp.489-495
• /
• 2015

Image availability evaluated by the degree of agreement and sensitive using the process improve visualization according to the Algorithm modification in Image Post-Processing. Reliability measured by the Breast Imaging Reporting and Data System. 172 patients visit same period divided by BI-RADS, category five stages, and contents of breast parenchyma into Calcification, Nodule and Mass. Evaluated the TE/PV image reliability, visualization sensitive, agreement of diagnosis. Convergence analysis was an in various fields. According to the result of this research, PV has higher sensitive and accuracy about lesions than TE visual and there is a difference insensitive by contents of breast parenchyma. Therefore, practical use of Algorithm Modification(Tissue Equalization: TE, Premium View: PV) is expected to improve more accurate, useful diagnosis, which has not been easy until now.

• Management & Information Systems Review
• /
• v.33 no.4
• /
• pp.39-48
• /
• 2014

In order that a company may reduce expense, according to companies' income statement, the largest amount of money is set to total sales amount, and the next is the sales cost. The method of reducing the sales cost is the common and it is important to take down a unit price at the time of purchase, or to reduce inventory cost. In this research, the cost reduction model for logistics cost reduction is built using the real data of P company, and in order to verify the validity of a model, logistics cost is calculated through a simulation. A real logistics cost is compared with the logistics cost through logistics route rearrange of the proposal model. It will become an opportunity which analyzes the logistics expense of P company through this research, and can be solved in search of the problem of logistics system.