• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.29 no.7
• /
• pp.550-559
• /
• 2018

The effectiveness of deep neural networks (DNNs) for detection and classification of micro-Doppler signals generated by human walking and background noise sources is investigated. Previous research included a complex process for extracting meaningful features that directly affect classifier performance, and this feature extraction is based on experiences and statistical analysis. However, because a DNN gradually reconstructs and generates features through a process of passing layers in a network, the preprocess for feature extraction is not required. Therefore, binary classifiers and multiclass classifiers were designed and analyzed in which multilayer perceptrons (MLPs) and DNNs were applied, and the effectiveness of DNNs for recognizing micro-Doppler signals was demonstrated. Experimental results showed that, in the case of MLPs, the classification accuracies of the binary classifier and the multiclass classifier were 90.3% and 86.1%, respectively, for the test dataset. In the case of DNNs, the classification accuracies of the binary classifier and the multiclass classifier were 97.3% and 96.1%, respectively, for the test dataset.

• The KIPS Transactions:PartC
• /
• v.17C no.1
• /
• pp.27-36
• /
• 2010

In Web registration pages and online polling applications, CAPTCHA(Completely Automated Public Turing Test To Tell Computers and Human Apart) is used for distinguishing human users from automated programs. Text-based CAPTCHAs have been widely used in many popular Web sites in which distorted text is used. However, because the advanced optical character recognition techniques can recognize the distorted texts, the reliability becomes low. Image-based CAPTCHAs have been proposed to improve the reliability of the text-based CAPTCHAs. However, these systems also are known as having some drawbacks. First, some image-based CAPTCHA systems with small number of image files in their image dictionary is not so reliable since attacker can recognize images by repeated executions of machine learning programs. Second, users may feel uncomfortable since they have to try CAPTCHA tests repeatedly when they fail to input a correct keyword. Third, some image-base CAPTCHAs require high communication cost since they should send several image files for one CAPTCHA. To solve these problems of image-based CAPTCHA, this paper proposes a new CAPTCHA based on both image and text. In this system, an image and keywords are integrated into one CAPTCHA image to give user a hint for the answer keyword. The proposed CAPTCHA can help users to input easily the answer keyword with the hint in the fused image. Also, the proposed system can reduce the communication costs since it uses only a fused image file for one CAPTCHA. To improve the reliability of the image-text fusion CAPTCHA, we also propose a dynamic building method of large image dictionary from gathering huge amount of images from theinternet with filtering phase for preserving the correctness of CAPTCHA images. In this paper, we proved that the proposed image-text fusion CAPTCHA provides users more convenience and high reliability than the image-based CAPTCHA through experiments.

• Restorative Dentistry and Endodontics
• /
• v.12 no.2
• /
• pp.45-53
• /
• 1987

The purpose of this study was to evaluate the tensile bond strength between composite resin and the human enamel. Three composite resin systems, two chemical (Clearfil Posterior, and Clearfil Posterior-3) and one light cure (Photo Clearfil-A), used with and without an intermediate resin (clearfil bonding agent), were evaluated under different amounts of load (10g, 200g and 200g for a moment) for in vitro tensile bond strength to acid-eched human enamel. Clinically intact buccal or lingual surfaces of 144 freshly extracted human permanent molars, embedded in acrylic were flattened with No #600 carborundum discs. Samples were randomly assigned to the different materials and treatments using a table of random numbers. Eight samples were thus prepared for each group(Table 2) these surfaces were etched with an acid etchant (Kurarey Co. Japan) in a mode of etching for 30 seconds, washing for 15 seconds, and drying for 30-seconds. During the polymerization of composite resin on the acid-etched enamel surfaces with and without bonding agent 10-gram, 200 gram and temporary 200 gram of load were applied. The specimens were stored in 50% relation humidity at $37^{\circ}C$ for 24 hours before testing. An universal Testing machine (Intesco model No. 2010, Tokyo, Japan) was used to apply tensile loads in the vertical directed (fig 5), and the force required for separation was recorded with a cross head speed of 0.25 mm/min and 20 kg in full scale. The results were as follow: 1. The tensile bond strength was much greater in applying a bonding agent than in not doing that. 2. The tensile bond strength of chemical cure composite resin was higher than that of light cure composite resin with applying on bonding agent on the acid-etched enamel. 3. In case of not applying a bonding agents on the acid-etching enamel, the highest tensile bond strength under 200 gram of load was measured in light cure composite resin. 4. The tensile bond strength under 200-gram of load has no relation with applying the bonding agent. 5. Under the load of 10-gram, There was significant difference in tensile bond strength as applying the bonding agent.

• Management & Information Systems Review
• /
• v.36 no.4
• /
• pp.129-148
• /
• 2017

Conventional exhibition management information systems recommend tourist attractions that are close to the place in which an exhibition is held. Some recommended attractions by the location-based recommendation could be meaningless when nothing is related to the exhibition's topic. Our goal is to recommend attractions that are related to the content presented in the exhibition, which can be coined as content-based recommendation. Even though human exhibition curators can do this, the quality is limited to their manual task and knowledge. We propose an automatic way of discovering attractions relevant to an exhibition of interests. Language resources are incorporated to discover attractions that are more meaningful. Because a typical single machine is unable to deal with such large-scale language resources efficiently, we implemented the algorithm on top of Apache Spark, which is a well-known distributed computing framework. As a user interface prototype, a web-based system is implemented that provides users with a list of relevant attractions when users are browsing exhibition information, available at http://bike.snu.ac.kr/WARP. We carried out a case study based on Korean Tourism Organization Linked Open Data with Korean Wikipedia as a language resource. Experimental results are demonstrated to show the efficiency and effectiveness of the proposed system. The effectiveness was evaluated against well-known exhibitions. It is expected that the proposed approach will contribute to the development of both exhibition and tourist industries by motivating exhibition visitors to become active tourists.

• Journal of the Ergonomics Society of Korea
• /
• v.31 no.5
• /
• pp.637-646
• /
• 2012

Objective: The aim of this study was to develop the adaptive device for severe physical disabilities using smart device in the driving simulator and its performance evaluation. Development of appropriate driving adaptive device for the people with serious physical limitation could contribute to maintain their community mobility. Background: There is lack of adaptive driving devices for the people with disabilities in Korea. However, if smart device systems like iPod and iPhone are used for driving a car, the people with serious physical limitations can improve their community mobility. Method: Both gyroscope and accelerometer from iPod were used to measure the tilted angle of the smart device for driving. Customized Labview program was also used to control three axis motors for steering wheel, accelerator and brake pedals. Thirteen subjects were involved in the experiment for performance evaluation of smart device in simulator. Five subjects had driver licenses. Another four subjects did not have driver licenses. Others were people with disabilities. Results: Average driving score of the normal group with driver license in the simulator increased 46.6% compared with the normal group without driver license and increased 30.4% compared with the disabled group(p<0.01). There was no significant difference in the average driving score between normal group without driver license and disabled group(p>0.05). Conclusion: The normal group with driver license showed significantly higher driving score than other groups. The normal group without driver license and disabled group could improve their driving skills with training in simulator. Application: If follow-up studies would be continued and applied in adapted vehicle for on road environment, many people with more severe disabilities could drive and improve the quality of life.

• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.307-311
• /
• 2003

2-mass system is generally used as controller of the variable-speed to transfer electromotion power to mechanical load such as industrial robot, driving parts of electric vehicle, rolling machine system of steel plant and driving parts of elevator. In this case, PI controller is often used as a velocity controller because of simplicity of system. But PI control algorithm is not enough for obtaining the control characteristics required for this system. To solve this problem, 2-mass system based on the PID controller derives the optimum PID parameters by pole assignment and estimation of the ITAE performance index. In this case, the system have tenacious properties about disturbance, but it causes extreme overshoot and vibration because of rapidly output of controller in early transient response about desired value. And if speed control system is applied by 2-DOF parameter ${\alpha}$, a temporary value, we must induce most suitable parameter by complicate pole assignment and estimation of the ITAE performance index whenever ${\alpha}$ changes. In this paper, to solve this problem we suggest control algorithm to followed exactly value of ${\alpha}$ as 2-DOF parameter by using fuzzy algorithm . So, intelligence algorithm modeled by human knowledge, experience, teachability and judgment follow exact ${\alpha}$ value and it can compose the efficient 2-DOF PID controller to improve following performance, overshoot decrease.

• Journal of Intelligence and Information Systems
• /
• v.13 no.3
• /
• pp.141-160
• /
• 2007

User interface has been regarded as a crucial issue to increase the acceptance of mobile services. In special, even though to what extent the machine as speaker communicates with human as listener in a timely and polite manner is important, fundamental studies to come up with these issues have been very rare. Hence, the purpose of this paper is to propose a methodology of estimating politeness level in a certain context-aware setting and then to design a context-aware system for polite mobile interface. We will focus on Korean language for the polite level estimation simply because the polite interface would highly depend on cultural and linguistic characteristics. Nested Minkowski aggregation model, which amends Minkowski aggregation model, is adopted as a privacy-preserving similarity evaluation for case retrieval under distributed computing environment such as ubiquitous computing environment. To show the feasibility of the methodology proposed in this paper, simulation-based experiment with drama cases has performed to show the performance of the methodology proposed in this paper.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.1
• /
• pp.288-307
• /
• 2016

Image distortions are typically characterized by degradations of structures. Dictionaries learned from natural images can capture the underlying structures in images, which are important for image quality assessment (IQA). This paper presents a general-purpose no-reference image quality metric using a GRadient-Induced Dictionary (GRID). A dictionary is first constructed based on gradients of natural images using K-means clustering. Then image features are extracted using the dictionary based on Euclidean-norm coding and max-pooling. A distortion classification model and several distortion-specific quality regression models are trained using the support vector machine (SVM) by combining image features with distortion types and subjective scores, respectively. To evaluate the quality of a test image, the distortion classification model is used to determine the probabilities that the image belongs to different kinds of distortions, while the regression models are used to predict the corresponding distortion-specific quality scores. Finally, an overall quality score is computed as the probability-weighted distortion-specific quality scores. The proposed metric can evaluate image quality accurately and efficiently using a small dictionary. The performance of the proposed method is verified on public image quality databases. Experimental results demonstrate that the proposed metric can generate quality scores highly consistent with human perception, and it outperforms the state-of-the-arts.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.1612-1616
• /
• 2005

In this paper, a Home Mess-Cleanup Robot(HMR), which has a practical function of the automatic mess-cleanup, is developed. The vacuum-cleaner had made the burden of house chore lighten but the operation labour of a vacuum-cleaner had been so severe. Recently, the cleaning robot was producted to perfectly solve the cleaning labour of a house but it also was not successful because it still had a problem of mess-cleaning, which was the clean-up of big trash and the arrangement of newspapers, clothes, etc. The cleaning robot is to just vacuum dust and small trash but has no function to arrange and take away before the automatic vacuum-cleaning. For this reason, the market for the cleaning robot is not yet built up. So, we need a design method and technological algorithm of new automatic machine to solve the problem of mess-cleanup in house. It needs functions of agile automatic navigation, novel manipulation system for mess-cleanup. The automatic navigation system has to be controlled for the full scanning of living room, to recognize the absolute position and orientation of the self, the precise tracking of the desired path, and to distinguish the mess object to clean-up from obstacle object to just avoid. The manipulator, which is not needed in the vacuum-cleaning robot, must have the functions, how to distinguish big trash to clean from mess objects to arrange, how to grasp in according to the form of mess objects, how to move to the destination in according to mess objects and arrange them. We use the RFID system to solve the problems in this paper and propose the reading algorithm of RFID tags installed in indoor objects and environments. Then, it should be an intelligent system so that the mess cleaning task can be autonomously performed in a wide variety of situations and environments. It needs to also has the entertainment functions for the good communication between the human and HMR. Finally, the good performance of the designed HMR is confirmed through the results of the mess clean-up and arrangement.

• Journal of Contemporary Eastern Asia
• /
• v.18 no.2
• /
• pp.16-32
• /
• 2019

Media manipulation of breaking news through article selection, ranking and tweaking of social media data and comment streams is a growing concern for society. We argue that the combination of human and machine curation on media portals marks a new period for news media and journalism. Although intermediary platforms routinely claim that they are merely the neutral technological platform which facilitates news and information flows, rejecting any criticisms that they are operating as de facto media organisations; instead, we argue for an alternative, more active interpretation of their roles. In this article we provide a contemporary account of the South Korean ('Korean') online news media ecology as an exemplar of how contemporary media technologies, and in particular portals and algorithmic recommender systems, perform a powerful role in shaping the kind of news and information that citizens access. By highlighting the key stakeholders and their positions within the production, publication and distribution of news media, we argue that the overall impact of the major portal platforms of Naver and Kakao is far more consequential than simply providing an entertaining media diet for consumers. These portals are central in designing how and which news is sourced, produced and then accessed by Korean citizens. From a regulatory perspective the provision of news on the portals can be a somewhat ambiguous and moving target, subject to soft and harder regulatory measures. While we investigate a specific case study of the South Korean experience, we also trace out connections with the larger global media ecology. We have relied on policy documents, stakeholder interviews and portal user 'walk throughs' to understand the changing role of news and its surfacing on a distinctive breed of media platforms.