• Korea Information Processing Society Review
• /
• v.30 no.3
• /
• pp.20-26
• /
• 2023

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.16 no.7
• /
• pp.4918-4924
• /
• 2015

In this paper, we consider the capacity improvement of systems exploiting dual-polarized antennas for two-user transmission. To this end, we analyze the upper bounds of ergodic capacities for multicast and unicast data services in the systems, and propose the condition for adjusting the complex cross-polarization discriminations (XPDs) to maximize the ergodic capacities. In addition, we present the adjustment condition of the complex XPDs that can achieve spectral efficiencies close to the maximum ergodic capacities with lower system complexity. Lastly simulation results demonstrate that the systems using the proposed conditions can obtain higher spectral efficiencies than the ones employing different adjustment conditions including the exiting adjustment condition.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.3 no.1
• /
• pp.47-52
• /
• 2010

In this paper, we introduce a new approach for real-time 3D face pose discrimination based on active IR illumination from a monocular view of the camera. Under the IR illumination, the pupils appear bright. We develop algorithms for efficient and robust detection and tracking pupils in real time. Based on the geometric distortions of pupils under different face orientations, an eigen eye feature space is built based on training data that captures the relationship between 3D face orientation and the geometric features of the pupils. The 3D face pose for an input query image is subsequently classified using the eigen eye feature space. From the experiment, we obtained the range of results of discrimination from the subjects which close to the camera are from 94,67%, minimum from 100%, maximum.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.17 no.3
• /
• pp.364-377
• /
• 2014

In this paper, we have developed a target recognition algorithm based on a template matching technique using Synthetic Aperture Radar (SAR) images. For efficient computations, Radon transform-based azimuth estimation algorithm was used with the template matching. MSTAR data set was divided into two groups according to the depression angles, which were a train set and a test set. Template data were generated by rotating and cropping chips which were from MSTAR train set using the azimuth estimation algorithm. Then the template matching process between test data and template data was performed under various conditions. Performance variation according to contrast enhancement preprocessing which is scarce in open literature was also presented. The analysis results show that the target recognition algorithm could be useful for the automatic target recognition using SAR images.

• Journal of Information Management
• /
• v.40 no.2
• /
• pp.95-115
• /
• 2009

Author resolution is to disambiguate same-name author occurrences into real individuals. For this, pair-wise author similarities are computed for author name entities, and then clustering is performed. So far, many studies have employed hierarchical clustering techniques for author disambiguation. However, various hierarchical clustering methods have not been sufficiently investigated. This study covers an empirical evaluation and analysis of hierarchical clustering applied to Korean author resolution, using multiple distance functions such as Dice coefficient, Cosine similarity, Euclidean distance, Jaccard coefficient, Pearson correlation coefficient.

• Journal of Information Management
• /
• v.42 no.1
• /
• pp.47-62
• /
• 2011

While an ideal author graph should have its nodes to represent authors, automatically-generated author graphs mostly use author names as their nodes due to the difficulty of resolving author names into individuals. However, employing author names as nodes of author graphs merges namesakes, otherwise separate nodes in the author graph, into the same node, which may distort the characteristics of the author graph. This study proposes an algorithm which resolves author ambiguities based on co-authorship and then yields an author graph consisting of not author name nodes but author nodes. Scientific collaboration relationship this algorithm depends on tends to produce the clustering results which minimize the over-clustering error at the expense of the under-clustering error. In experiments, the algorithm is applied to the real citation records where Korean namesakes occur, and the results are discussed.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.1
• /
• pp.617-623
• /
• 2024

Identifying emotions through speech poses a significant challenge due to the complex relationship between language and emotions. Our paper aims to take on this challenge by employing feature engineering to identify emotions in speech through a multimodal classification task involving both speech and text data. We evaluated two classifiers-Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM)-both integrated with a BERT-based pre-trained model. Our assessment covers various performance metrics (accuracy, F-score, precision, and recall) across different experimental setups). The findings highlight the impressive proficiency of two models in accurately discerning emotions from both text and speech data.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.48 no.3
• /
• pp.91-101
• /
• 2011

Internet is becoming increasingly popular due to the rapid development of information and communication technology. There has been a convenient social activities such as the mutual exchange of information, e-commerce, internet banking, etc. through cyberspace on a computer. However, by using the convenience of the internet, the personal IDs(identity card, driving license, passport, student ID, etc.) represented by the electronic media are exposed on the internet frequently. Therefore, this study propose a feature extraction method to analyze the characteristics of image files containing personal information and a image retrieval method to find the images using the extracted features. The proposed method selects the feature information from color, texture, and shape of the images, and the images as searched by similarity analysis between feature information. The result which it experiments from the image which it acquires from the web-based image DB and correct image retrieval rate is 89%, the computing time per frame is 0.17 seconds. The proposed method can be efficiently apply a system to search the image files containing personal information and to determine the criteria of exposure of personal information.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.16 no.1
• /
• pp.169-175
• /
• 2016

Geofence is a virtual perimeter for a real-world geographical area, which could be statically or dynamically established the specified area if necessary. Many geofencing applications incorporate 2D(two-dimensional) map such as the Google map, allowing administrators to define boundaries on top of a satellite view of a specific geographical area. But these applications do not provide 3D(three-dimensional) spatial information as well as 2D location information no matter where indoor or outdoor. Therefore we propose a mechanism to identify indoor or outdoor location for 3D geofence, and implement 3D geofence using smartphone. The proposed mechanism identifies the position information on 3D geofence regardless of indoor or outdoor, inter-floor with only GPS and WiFi. In the near future, 3D geofence as well as LBS are promising applications that become possible when IoT can become organized and connected by location.

• Journal of KIISE:Software and Applications
• /
• v.32 no.12
• /
• pp.1205-1216
• /
• 2005

Component-based Development (CBD) is gaining popularity as an effective reuse technology. Especially commercial-off-the-shelf (COTS) components are mainly for inter-organizational reuse, rather than intra-organizational reuse. One of the essential tasks in CBD is to reuse the right components that provide the functionality and interface required by component consumers. If candidate components provide a limited applicability and customizability, a component consumer cannot reuse the components in application development. To resolve this partial matching problem, we need smart connectors that fill the gap between candidate components and the specification of components required. Previous researches about smart connector describe only connector types to resolve mismatch problems. However, previous researches do not address the identification and design method to resolve the problems. In this paper, we suggest taxonomy of various mismatch problems to identify partial match problems between requirements of the application and components. We identify smart connector types and suggest a systematic process for designing smart connectors using the taxonomy. Therefore, components that have the problems can be used to develop applications.