Search

There is no doubt that CNN has made remarkable technological developments as the core technology of computer vision, but the pooling technique used for CNN has its own issues. This study set out to solve the issues of the pooling technique by proposing conditional min pooling and a restructured convolutional neural network that improved the pooling structure to ensure efficient use of the conditional min pooling. Some Caltech 101 and crawling data were used to test the performance of the conditional min pooling and restructured convolutional neural network. The pooling performance test based on Caltech 101 increased in accuracy by 0.16~0.52% and decreased in loss by 19.98~28.71% compared with the old pooling technique. The restructured convolutional neural network did not have a big improvement in performance compared to the old algorithm, but it provided significant outcomes with similar performance results to the algorithm. This paper presents the results that the loss rate was reduced rather than the accuracy rate, and this result was achieved without the improvement of convolution.

This paper set out to revise and improve existing autonomous driving models using reinforcement learning, thus proposing a reinforced autonomous driving prediction model. The paper conducted training for a reinforcement learning model using DQN, a reinforcement learning algorithm. The main aim of this paper was to reduce the time spent on training and improve self-driving performance. Rewards for reinforcement learning agents were developed to mimic human driving behavior as much as possible. High rewards were given for greater distance travelled within lanes and higher speed. Negative rewards were given when a vehicle crossed into other lanes or had a collision. Performance evaluation was carried out in urban environments without pedestrians. The performance test results show that the model with the collision prevention model exhibited faster performance improvement within the same time compared to when the model was not applied. However, vulnerabilities to factors such as pedestrians and vehicles approaching from the side were not addressed, and the lack of stability in the definition of compensation functions and limitations with respect to the excessive use of memory were shown.

요 약석탄과 석유를 대신할 새로운 대체 에너지의 연구 및 개발은 지난 수년간 계속해서 연구되고 있다. 최근엔 유한 에너지 가격 상승으로 인하여 무공해 신재생 에너지의 관심이 증가하고 있다. 특히 기존에 활발히 연구되고 상용화된 신재생 에너지 분야 중 태양에너지는 태양광 에너지에 주된 연구가 이루어져 있다. 그러나 태양광 에너지의 효율성은 이미 포화상태에 도달하였다. 이에 본 논문에서는 태양광이 아닌 태양열 에너지를 활용할 수 있는 태양광 추적형 파라볼릭 집열장치를 제안한다. 제안하는 태양열 집열장치는 효과적으로 태양열을 집열하기 위하여 파라볼릭 형태로 제안하고, 계절 및 시기에 따른 태양의 위치변화를 감지할 수 있도록 5축 센서 방식의 태양광 추적형 트래킹 기술을 결합한 태양열 집열장치의 프로토타입을 제안한다. 또한 집열장치의 효율적인 관리를 위하여 관리자 인터페이스를 설계 및 구현한다.ABSTRACT

This study set out to invent an Information and Communication Technologies (ICT)-based smart Acer mono sap collection electric device to make efficient use of the labor force by reducing inefficient activities of old manual work to record sap exudation and state information. Based on the assumption that environmental information would have close connections with Acer mono sap exudation to reinforce the competitive edge of production in forest products, the study analyzed correlations between Acer mono sap exudation and environmental information and predicted Acer mono exudation. A smart collection of electric devices would gather data about Acer mono sap exudation per hour on outdoor temperature, humidity, conductivity, and wind direction and velocity, and was installed in four areas in the Republic of Korea, including Sancheong, Gwangyang, Geoje, and Inje. Collected data were used to analyze correlations between environmental information and Acer mono sap exudation using four different algorithms, including linear regression, Support Vector Machine (SVM), Artificial Neural Network (ANN), and random forest, to predict Acer mono sap exudation. Remarkable outcomes were obtained across all the algorithms except for linear regression, demonstrating close connections between environmental information and Acer mono sap exudation. The random forest model, which showed the most outstanding performance, was used to make a mobile app capable of providing predicted Acer mono sap exudation and collected environmental information.

There is a shortage of mobile application systems readily applicable to the field of crop cultivation in relation to diseases and insect pests directly connected to the quality of crops. Most of system have been devoted to diseases and insect pests that would offer full predictions and basic information about diseases and insect pests currently. But for lack of the instant diagnostic functions seriously and the field of crop cultivation, we design and implement a crop cultivation field-oriented smart diseases and insect pests information retrieval system based on mobile for u-Farm. The proposed system had such advantages as providing information about diseases and insect pests in the field of crop cultivation and allowing the users to check the information with their smart-phones real-time based on the Lucene, a search library useful for the specialized analysis of images, and JSON data structure. And it was designed based on object-oriented modeling to increase its expandability and reusability. It was capable of search based on such image characteristic information as colors as well as the meta-information of crops and meta-information-based texts. The system was full of great merits including the implementation of u-Farm, the real-time check, and management of crop yields and diseases and insect pests by both the farmers and cultivation field managers.

Recently, it is being held festival according to regional and seasonally cultural feature in nationwide. Especially, local festivals are being provided festival information services a variety of forms for tourists depending on development of information device. However, systems to provide notification information of local festivals are not persistent but mostly short term service. In this paper, we propose a mobile commerce and information notification system through festival information alerts and festival surrounding information using local based on service. In addition, we implement a festival information notification system that applies to beacon based on bluetooth in order to inform them on diverse festival. Finally, we take into account in terms of design of system using object-oriented design modeling to improve reusability of proposed system.

All data created in BigData times is included potentially meaning and correlation in data. A variety of data during a day in all society sectors has become created and stored. Research areas in analysis and grasp meaning between data is proceeding briskly. Especially, accuracy of meaning prediction and data imbalance problem between data for analysis is part in course of something important in data analysis field. In this paper, we proposed data prediction model based on data weights and neural network using R for meaning analysis between data. Proposed data prediction model is composed of classification model and analysis model. Classification model is working as weights application of normal distribution and optimum independent variable selection of multiple regression analysis. Analysis model role is increased prediction accuracy of output variable through neural network. Performance evaluation result, we were confirmed superiority of prediction model so that performance of result prediction through primitive data was measured 87.475% by proposed data prediction model.

A lot of people were trapped in elevators without power supply when BLACK-OUT situation occurred in 2011. The telephone network of control room connected to the elevators had problem operating poorly. In this paper we propose an ICT based elevator emergency call device prototype system and evaluate the performance of the system. The proposed system quickly responds in emergency situation to guarantee passenger safety. For the goal, firstly the system tries to connect to a control room. If it fails the system attempts to call numbers for emergency contact and a rescue team sequentially. The system is designed to quickly support emergency contact as well. Finally, the information of elevator failure is rapidly transferred to the failure process device by the proposed system.

Datum point for displaying location and altitude of point has being advantage usefully in various ∙제1저자 : 정세훈 ∙교신저자 : 심춘보 ∙투고일 : 2014. 11. 12, 심사일 : 2014. 12. 15, 게재확정일 : 2015. 1. 6. * 순천대학교 멀티미디어공학과 박사수료(Dept. of Multimedia Engineering, Sunchon National University) ** 순천대학교 멀티미디어공학과 부교수(Dept. of Multimedia Engineering, Sunchon National University) ※ 본 논문은 중소기업청에서 지원하는 2012년도 산학연공동기술개발사업(C0036798)의 연구수행으로 인한 결과물임을 밝힙니다. 100 Journal of The Korea Society of Computer and Information February 2015 measurement parts. However, datum point has been increasing loss cases owing to weather changes and stratum changes and neglecting meaninglessly. In this paper, we design and implement a multi electronic measurement system monitoring software with functions such as include maximize utilization of existing measurement datum system as well as collected various environment data and detection stratum changes of surround area. Proposed software is implemented to support that reusability and extensibility of software using object oriented modeling method. Our software supports a GUI for electronic measurement datum point administrator as well as for web user and mobile user. Our system can support a graph GUI for various data analysis and reposition in realtime to database that measured location information and various sensing information to prevent loss of electronic measurement datum point and to detected stratum changes. In addition, we include a QR code and RFID recognition function. Finally, we suggest performance evaluation result to confirm stratum changes detection and GPS location error rate. ▸

Recently, information providing service using Big Data is being expanded. Big Data processing technology is actively being academic research to an important issue in the IT industry. In this paper, we analyze a skilled pattern of welder through Big Data analysis or extraction of welding based on R programming. We are going to reduce cost on welding work including weld quality, weld operation time by providing analyzed results non-skilled welder. Welding has a problem that should be invested long time to be a skilled welder. For solving these issues, we apply connection rules algorithms and regression method to much pattern variable for welding pattern analysis of skilled welder. We analyze a pattern of skilled welder according to variable of analyzed rules by analyzing top N rules. In this paper, we confirmed the pattern structure of power consumption rate and wire consumption length through experimental results of analyzed welding pattern analysis. 키워드

A number of crime and disease relief situation show a clear increasing trend in domestic and overseas country. In particular, we desperately need a fast location information sharing with high accuracy for dealing with this situation efficiently. In this paper, we propose a DGPS framework and performance evaluation based on Web service enable increase location accuracy using low-cost GPS or AGPS. The framework consists of mobile station to realize the final service by receiving the correction values that are created in the base station closest to the current position from DGPS server, base station system to transmit the correct values to DPGS server by installing at base point and DGPS framework working at DGPS server. Our framework is fundamentally designed to respond based on XML to service request. In addition, for performance evaluation, we make use of 3 sites of Level 2 city base-station among the base stations being operated in Suncheon-si. Through performance evaluation, we show that our framework outperforms about 10%~15% in terms of error improvement rate, compared with the existing schemes. And we have the advantage that various services can be expanded owing to receiving the correction values through mobile device such as, smart phone, smart pad, net-book by using XML based Web services.

In this paper, We proposes multipurpose function electronic measurement reference station prototype system based on LED and Gyro Sensor. This system have developed to possible with real time data collection and landscape illumination and surrounding diastrophism monitoring for that analyze problems of existing planted reference station. Hardware for the proposed system consists of a gyro sensor module to detect diastrophism, GPS module to process location information, environment sensor module to process surrounding environmental data and CDMA wireless data communication to send the collected data to server. In addition, this paper intends to enhance system management and future usability in a way that applies LED, QR-code, RFID and Solar cell module to outer side of electronic measurement reference station to improve usability of H/W for electronic measurement reference station. Lastly, this paper conducted current amount evaluation to supply stabilized electricity with its various functions in the proposed multipurpose function electronic measurement reference station and it`s proved that this system can be stably operated with its electricity loss factor of 2.29% loss factor. And this paper conducted a qualitative comparison with existing electronic measurement reference station system in order to evaluate superiority of the proposed electronic measurement reference station.

Currently, registered number of imported vehicles is increasing rapidly over the years. Accordingly, environment improvements of vehicle maintenance company for maintenance of luxury vehicle such as imported vehicle are continuously being made. In this paper, we propose a key frame extraction method based on HSV color model for smart vehicle management system implementation to offer for customer reliability of maintenance vehicle. After automatically recognize the license plates of the vehicle using vehicle license plate recognition system when the vehicle come in the car center, we check the repair history and request of the vehicle based on it. We implement mobile services which provide extracted key frame images to the user after extract key frames from vehicle repair video. In addition, we verify the superiority of key frame extraction method by applying a smart vehicle management system. Finally, we convert the RGB color to HSV color to improve the performance of proposed key frame extraction scheme. As a result, we confirmed that our scheme is more excellence about 30% in terms of recall than RGB color model from the performance evaluations.

A variety of location-based services applications, such as missing children search, emergency rescue requests and so on that requiring high-precision location information are increasing. Precision of GPS that can be used in most systems, however, is still low. In this paper, we design and propose a low cost differential global positioning system(DGPS) based on Web services using object-oriented modeling technique which can offer useable location service, variety device and safe service in wireless environment. The proposed system is designed with UML based on object-oriented modeling to maximize system recyclability and system scalability. In addition, we would like to improve the precision of the GPS in accordance with mobile station location when build low cost mobile station, location differential framework and server. We implement a communication interface based on web-service which is available in the form of a variety of services and can offer stable according to mobile environments. Finally, as performance evaluation results, we can obtain precision location within 1 ~ 2m through proposed system and 88.5% probability of less than 2m.

The context aware service is the service to provide useful information to the users by recognizing surroundings around people who receive the service via computer based on computing and communication, and by conducting self-decision. But CAS(Context Awareness System) shows the weak point of small-scale context awareness processing capacity due to restricted mobile function under the current mobile environment, memory space, and inference cost increment. In this paper, we propose a mobile cloud context system with using Google App Engine based on PaaS(Platform as a Service) in order to get context service in various mobile devices without any subordination to any specific platform. Inference design method of the proposed system makes use of knowledge-based framework with semantic inference that is presented by SWRL rule and OWL ontology and Jess with rule-based inference engine. As well as, it is intended to shorten the context service reasoning time with mapping the regular reasoning of SWRL to Jess reasoning engine by connecting the values such as Class, Property and Individual which are regular information in the form of SWRL to Jess reasoning engine via JessTab plug-in in order to overcome the demerit of queries reasoning method of SparQL in semantic search which is a previous reasoning method.

Although underground water resource has relatively less pollution rate compared with surface water, its recovery faces many difficulties due to poor management. Our country monitors underground water to manage it effectively through auxiliary observation network for underground water. In this paper, we suggest water-well auto measure system based on Geosensor for business efficiency increase of water-well management and realtime monitering. In this system is consist of user GUI(Graphic User Interface) composed with water-well information and movement sub-observation network prototype system composed with GPS(Global Positioning System) and wireless sensor node such as water temperature, water level, electrical conductivity. In this system is using the light of the sun for self-power, variety water-well information collected wireless sensor node was a wireless transmitting/receiving a using CDMA(Code Division Multiple Access) module. Also, for promote with user ease, user GUI express that water-well collected in GIS(Geographic Information System) map. For performance evaluation of the proposed system, we perform experiment using sensing information through designed sub-observation network. And we was proved superiority of the proposed system through qualitative evaluation with other paper.

Due to recent development in wireless communication technologies and mobile information devices, the services on ubiquitous computing technology without time and place restriction have been spotlighted. Moreover, the interest of the Wireless Sensor Networks (WSNs) and context-awareness technologies have largely been escalating and their technologies utilization is active in the various applications such as healthcare, farm management and so on. However, the direct adaption of the existing context-awareness technique to the WSN technology has several drawbacks as follows. First, such systems waste precious energy of sensor nodes, due to unnecessary data transmissions. Secondly, since the existing work was designed to support only specific applications, it is required to implement a new context-awareness application for a specific purpose. Therefore, we, in this paper, propose a new real-time monitoring system based on context-awareness in WSNs. Our system not only enhances energy efficiency by reducing data transmissions by doing context-awareness on a sensor node, but also is scalable in terms of supporting new context-awareness functionalities through modularization.

Recently, fire monitoring application systems have been an active research area due to the safety of industries, historical monuments and so on. The fire monitoring application systems can reduce the damage of properties by providing earlier warning for possible fire situation. However, the existing systems have a drawback that they detect fire with delay due to their uniform epoch in fire detection algorithm. Moreover, they do not provide user-friendly graphical user interfaces in their fire monitoring systems. To resolve the problems, First, we propose a new fire detection algorithm (Early Fire Detection Fire Algorithm) which uses the distribution of sensing data for early fire detection. Our fire detection algorithm is better in terms of fire detection time than the existing work because it can set a start time of fire detection epoch dynamically based on data distribution. Second, we develop a fire monitoring application system which provides users with both a user-friendly graphical user interface and a fire alarm message when fire occurs. Finally, we show from our experiment that our developed system is effectively used for early fire warning in a variety of fire situations.

One of the main purpose of multiple-view image processing technology is support realistic 3D image to device user by using multiple viewpoint display devices and compressed data restoration devices. This paper proposes a multiple view image synthesis scheme based on RAM disk which makes possible to browse 3D images generated by applying effective composing method to real time input stereo images. The proposed scheme first converts input images to binary image. We applies edge detection algorithm such as Sobel algorithm and Prewiit algorithm to find edges used to evaluate disparities from images of 4 multi-cameras. In addition, we make use of time interval between hardware trigger and software trigger to solve the synchronization problem which has stated ambiguously in related studies. We use a unique identifier on each snapshot of images for distributed environment. With respect of performance results, the proposed scheme takes 0.67 sec in each binary array. to transfer entire images which contains left and right side with disparity information for high quality 3D image browsing. We conclude that the proposed scheme is suitable for real time 3D applications.

Various types of derivative information have been increasing exponentially, based on mobile devices and social networking sites (SNSs), and the information technologies utilizing them have also been developing rapidly. Technologies to classify and analyze such information are as important as data generation. This study concentrates on data clustering through principal component analysis and K-means algorithms to analyze and classify user data efficiently. We propose a technique of changing the cluster choice before cluster processing in the existing K-means practice into a variable cluster choice through principal component analysis, and expanding the scope of data clustering. The technique also applies an artificial neural network learning model for user recommendation and prediction from the clustered data. The proposed processing model for predicted data generated results that improved the existing artificial neural network–based data clustering and learning model by approximately 9.25%.