• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.49 no.3
• /
• pp.10-17
• /
• 2012

In general, image-based 3D scenes can now be found in many popular vision systems, computer games and virtual reality tours. In this paper, we propose a method for creating 3D virtual scenes based on 2D image that is completely automatic and requires only a single scene as input data. The proposed method is similar to the creation of a pop-up illustration in a children's book. In particular, to estimate geometric structure information for 3D scene from a single outdoor image, we apply the tensor voting to an image segmentation. The tensor voting is used based on the fact that homogeneous region in an image is usually close together on a smooth region and therefore the tokens corresponding to centers of these regions have high saliency values. And then, our algorithm labels regions of the input image into coarse categories: "ground", "sky", and "vertical". These labels are then used to "cut and fold" the image into a pop-up model using a set of simple assumptions. The experimental results show that our method successfully segments coarse regions in many complex natural scene images and can create a 3D pop-up model to infer the structure information based on the segmented region information.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.43 no.4 s.310
• /
• pp.50-57
• /
• 2006

In Electronic Commerce, the latest most of the personalized recommender systems have applied to the collaborative filtering technique. This method calculates the weight of similarity among users who have a similar preference degree in order to predict and recommend the item which hits to propensity of users. In this case, we commonly use Pearson Correlation Coefficient. However, this method is feasible to calculate a correlation if only there are the items that two users evaluated a preference degree in common. Accordingly, the accuracy of prediction falls. The weight of similarity can affect not only the case which predicts the item which hits to propensity of users, but also the performance of the personalized recommender system. In this study, we verify the improvement of the prediction accuracy through an experiment after observing the rule of the weight of similarity applying Vector similarity, Entropy, Inverse user frequency, and Default voting of Information Retrieval field. The result shows that the method combining the weight of similarity using the Entropy with Default voting got the most efficient performance.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.43 no.5 s.311
• /
• pp.38-47
• /
• 2006

Conventional classification method uses a single classifier based on shape or motion feature. However this method exhibits a weakness if naively used since the classification performance is highly sensitive to the accuracy of moving region to be detected. The detection accuracy, in turn, depends on the condition of the image background. In this paper, we propose to resolve the drawback and thus strengthen the classification reliability by employing a Bayesian decision fusion and by optimally combining the decisions of three classifiers. The first classifier is based on shape information obtained from Fourier descriptors while the second is based on the shape information obtained from image gradients. The third classifier uses motion information. Our experimental results on the classification Performance of human and vehicle with a static camera in various directions confirm a significant improvement and indicate the superiority of the proposed decision fusion method compared to the conventional Majority Voting and Weight Average Score approaches.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.42 no.6
• /
• pp.63-74
• /
• 2005

In Electronic Commerce, the latest most of the personalized recommender systems have applied to the collaborative filtering technique. This method calculates the weight of similarity among users who have a similar preference degree in order to predict and recommend the item which hits to propensity of users. In this case, we commonly use Pearson Correlation Coefficient. However, this method is feasible to calculate a correlation if only there are the items that two users evaluated a preference degree in common. Accordingly, the accuracy of prediction falls. The weight of similarity can affect not only the case which predicts the item which hits to propensity of users, but also the performance of the personalized recommender system. In this study, we verify the improvement of the prediction accuracy through an experiment after observing the rule of the weight of similarity applying Vector similarity, Entropy, Inverse user frequency, and Default voting of Information Retrieval field. The result shows that the method combining the weight of similarity using the Entropy with Default voting got the most efficient performance.

• Asia-pacific Journal of Multimedia Services Convergent with Art, Humanities, and Sociology
• /
• v.8 no.3
• /
• pp.895-905
• /
• 2018

Public opinion polls have a strong influence on modern society as a means of examining the tendency of social groups on specific issues. As the influence of the polls increases, the problem of forgery and falsification of the results becomes an important issue. So, to guarantee the reliability of the results, our society needs novel mechanisms. As one of such mechanisms, the Ethereum blockchain is an environment for developing decentralized applications with the reliable blockchain technology. Ethereum decentralized applications can utilize smart contracts to provide services for users in transparent and reliable ways. In this paper, we propose a polling method that guarantees reliability using the blockchain technology, which is a distributed ledger technique that makes forgery or falsification actually impossible. The proposed method provides a robust verification function on the results of the associated polls for individual voters and verification organizations. Also, we present a distributed opinion polling application running on our private Ethereum blockchain network, showing the effectiveness of the proposed method.

• Journal of the HCI Society of Korea
• /
• v.10 no.1
• /
• pp.5-17
• /
• 2015

Nowadays various public data have been serviced to the public. Through the opening of public data, the transparency and effectiveness of public policy developed by governments are increased and users can lead to the growth of industry related to public data. Since end-users of using public data are citizens, it is very important for everyone to figure out the meaning of public data using proper visualization techniques. In this work, to indicate the significance of widespread public data, we consider UN voting record as public data in which many people may be interested. In general, it has high utilization value by diplomatic and educational purposes, and is available in public. If we use proper data mining and visualization algorithms, we can get an insight regarding the voting patterns of UN members. To visualize, it is necessary to measure the voting similarity values among UN members and then a social graph is created by the similarity values. Next, using a graph layout algorithm, the social graph is rendered on the screen. If we use the existing method for visualizing the social graph, it is hard to understand the meaning of the social graph because the graph is usually dense. To improve the weak point of the existing social graph visualization, we propose Friend-Matching, Friend-Rival Matching, and Bubble Heap algorithms in this paper. We also validate that our proposed algorithms can improve the quality of visualizing social graphs displayed by the existing method. Finally, our prototype system has been released in http://datalab.kunsan.ac.kr/politiz/un/. Please, see if it is useful in the aspect of public data utilization.

• Health Policy and Management
• /
• v.24 no.1
• /
• pp.100-106
• /
• 2014

Background: There is no consensus on the definition of primary care in South Korea. This study's objective was to define the concept of primary care using a Delphi method. Methods: Three expert panels were formed, consisting of 16 primary care policy researchers, 45 stakeholders, and 16 primary care physicians. Three rounds of voting, using 9-point appropriateness scales, were conducted. The first round involved rating the appropriateness of 20 previously established attributes of primary care. In the second round, panelists received a summary of the first-round results and were asked to once again vote on the 10 undetermined attributes and the provisional definition. The final round involved voting on the appropriateness of the revised definition. The Korean Language Society reviewed the revised definition. Results: Four core (first-contact care, comprehensiveness, coordination, and longitudinality) and three ancillary (personalized care, family and community context, and community base) attributes were selected. The Korean definition of primary care was accomplished with all three panel groups arriving at a 'very good' level of consensus. Conclusion: The Korean definition of primary care will provide a framework for evaluating performance of primary care in South Korea. It will also contribute to resolving confusion about the concept of primary care.

• The Korean Journal of Applied Statistics
• /
• v.15 no.1
• /
• pp.33-43
• /
• 2002

Various kinds of estimation methods have been developed for imputation of categorical missing data. They include category method, logistic regression, and association rule. In this study, we propose two fusions algorithms based on both neural network and voting scheme that combine the results of individual imputation methods. A Mont-Carlo simulation is used to compare the performance of these methods. Five factors used to simulate the missing data pattern are (1) input-output function, (2) data size, (3) noise of input-output function (4) proportion of missing data, and (5) pattern of missing data. Experimental study results indicate the following: when the data size is small and missing data proportion is large, modal category method, association rule, and neural network based fusion have better performances than the other methods. However, when the data size is small and correlation between input and missing output is strong, logistic regression and neural network barred fusion algorithm appear better than the others. When data size is large with low missing data proportion, a large noise, and strong correlation between input and missing output, neural networks based fusion algorithm turns out to be the best choice.

• Journal of Korea Society of Industrial Information Systems
• /
• v.20 no.4
• /
• pp.39-45
• /
• 2015

In this paper, variable rate CDR(Clock and Data Recovery) circuit adopting blind oversampling architecture is presented. The clock recovery circuit is implemented by using wide range voltage controlled oscillator and band selection method and the data recovery circuit is designed to digital circuit used majority voting method in order to low power and small area. The designed low power variable clock and data recovery is implemented by wide range voltage controlled oscillator and digital data recovery circuit. The designed variable rate CDR is operated from 10 bps to 2 Mbps. The total power consumption is about 4.4mW at 1MHz clock. The supply voltage is 1.2V. The designed die area is $120{\mu}m{\times}75{\mu}m$ and this circuit is fabricated in $0.13{\mu}m$ CMOS process.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.11
• /
• pp.427-432
• /
• 2019

In this paper, we propose three approaches to modeling Android malware. The first method involves human security experts for meticulously selecting feature sets. With the second approach, we choose 300 features with the highest importance among the top 99% features in terms of occurrence rate. The third approach is to combine multiple models and identify malware through weighted voting. In addition, we applied a novel method of eliminating permission information which used to be regarded as a critical factor for distinguishing malware. With our carefully generated feature sets and the weighted voting by the ensemble algorithm, we were able to reach the highest malware detection accuracy of 97.8%. We also verified that discarding the permission information lead to the improvement in terms of false positive and false negative rates.