Korean German Institute of Technology

Seoul, South Korea

Korean German Institute of Technology

Seoul, South Korea
SEARCH FILTERS
Time filter
Source Type

Hwang W.,Samsung | Wang H.,Samsung | Kim H.,Korean German Institute of Technology | Kee S.-C.,Mando Corporation | Kim J.,KAIST
IEEE Transactions on Image Processing | Year: 2011

The authors present a robust face recognition system for large-scale data sets taken under uncontrolled illumination variations. The proposed face recognition system consists of a novel illumination-insensitive preprocessing method, a hybrid Fourier-based facial feature extraction, and a score fusion scheme. First, in the preprocessing stage, a face image is transformed into an illumination-insensitive image, called an "integral normalized gradient image," by normalizing and integrating the smoothed gradients of a facial image. Then, for feature extraction of complementary classifiers, multiple face models based upon hybrid Fourier features are applied. The hybrid Fourier features are extracted from different Fourier domains in different frequency bandwidths, and then each feature is individually classified by linear discriminant analysis. In addition, multiple face models are generated by plural normalized face images that have different eye distances. Finally, to combine scores from multiple complementary classifiers, a log likelihood ratio-based score fusion scheme is applied. The proposed system using the face recognition grand challenge (FRGC) experimental protocols is evaluated; FRGC is a large available data set. Experimental results on the FRGC version 2.0 data sets have shown that the proposed method shows an average of 81.49% verification rate on 2-D face images under various environmental variations such as illumination changes, expression changes, and time elapses. © 2011 IEEE.


Bae K.,Korean German Institute of Technology | Kim H.,Chonbuk National University
International Journal of Multimedia and Ubiquitous Engineering | Year: 2013

The conventional search methods have computational complexity problem and imprecision problem in correspondence matching process. To resolve these problems, we propose how to effectively make feature space (distance map) and how to rapidly search the optimal point correspondence. The proposed distance map named Voronoi distance map is a 2-Dimensional surface that contains the distance information between each element(x-y coordinates) of image and the nearest feature point. The proposed distance map is efficiently created based on the priority-based calculation algorithm. The general distance calculation algorithm has a time complexity of O(w * h * n). (w = width, h = height, n = the number of feature points) whereas the priority-based distance calculation algorithm is a effective method with a computational cost of O(w * h * logn). Also, The partition search algorithm is a efficient method that can detect corresponding points very rapidly because this method can reduce the search range by a quarter at a time. Experimental results show that the proposed method outperforms conventional methods in reducing computation time and detecting the optimal correspondence. © 2013 SERSC.


Kim H.,Korean German Institute of Technology | Lee S.,Sungkyunkwan University
Proceedings - IEEE International Conference on Robotics and Automation | Year: 2010

This paper presents a novel line matching method based on the intersection context of coplanar line pairs especially working in poorly textured indoor scenes. To overcome the matching ambiguity of single line segments, intersecting line pairs in 2D images are utilized for line matching. Coplanarity of the intersecting line pairs and their corresponding intersection context discriminate true intersecting line pairs from false intersecting ones in 3D world. Compared to previous approaches, the proposed method can match line segments and estimate camera geometry simultaneously not knowing camera geometry in advance or not considering topological relations of all line segments. Comparison studies and experimental results prove the accuracy and speed of the proposed method in real world situations. ©2010 IEEE.


Jang S.,Yonsei University | Yang J.,Korean German Institute of Technology | Kim D.K.,Yonsei University
IEEE Communications Letters | Year: 2010

We study a joint optimal design of transceiver and relaying architecture in terms of mean square error (MSE) performance for multi-user multiple-input multiple-output (MIMO) relaying system. We show a duality relationship in the design of MSE-optimal transceiver and relaying architecture. Based on that, we devise an alternating algorithm, which converges and provides a suboptimal joint design of transceiver and relaying. By simulation, we show that the proposed joint design improves on the average MSE performance of the existing transceiver and relaying architectures. © 2010 IEEE.


Kim J.,Yonsei University | Choe Y.,Yonsei University | Kim Y.-G.,Korean German Institute of Technology
Digest of Technical Papers - IEEE International Conference on Consumer Electronics | Year: 2013

The recursive quad-tree structure of Coding Unit (CU) in HEVC can improve coding efficiency while the encoding complexity is increased significantly. In this paper, a fast CU size decision algorithm to reduce the complexity of HEVC intra coding is proposed. It introduces the early termination method based on the statistics of rate-distortion costs in CU splitting process. Experimental results show the encoding time can be reduced by 24% on average without significant loss of BD-rate. © 2013 IEEE.


Lee Y.-B.,Sungkyunkwan University | Choi Y.-J.,Korean German Institute of Technology | Kim M.-H.,Ewha Womans University
Computers in Biology and Medicine | Year: 2010

The intima-media thickness (IMT) of the carotid artery, obtained from B-mode ultrasound images, has recently been proposed as one of the most useful indices of atherosclerosis and can also be used to predict major cardiovascular events. Ultrasonic measurements of the IMT are conventionally obtained by time-consuming manual tracing of the interfaces between tissue layers. We propose a computerized method to detect the boundary of the intima-media complex using a directional Haar-like filter that can account for the slope of the boundary in an image. The directional Haar-like filter extracts a directional boundary feature as an image feature in the region of interest, which is used to compute a cost function. A cost function includes not only the directional Haar-like filtering value but also the geometric continuity that is computed for every pixel in the region of interest. The optimal boundary pixels are detected by using a dynamic programming approach that searches for the pixel that minimizes the cost function in each column of the image. We compared the performance of the proposed method with that of manual methods performed by two radiologists. The results showed that our approach produces very similar results to those based on manual tracing, and there was no statistically significant difference between the IMT measurements segmented manually and those analyzed using our method. © 2010 Elsevier Ltd.


Yang J.,Korean German Institute of Technology
Computational and Mathematical Organization Theory | Year: 2014

The key player problem (KPP) identifies a set of key nodes that have a central role in a network. In this paper, we propose a generalized KPP (GKPP) that extends existing work on KPP-Pos and KPP-Neg in such a way that it can consider network structure, node attributes, and the characteristics of edges. We also articulate a novel concept called the key player problem for exclusion (KPP-E), which selects a set of nodes to enforce the centrality of a given set of nodes of interest. To solve this problem efficiently, we propose a sequential greedy algorithm that significantly reduces computational complexity. To corroborate the conceptual meaning and effectiveness of the proposed sequential greedy algorithm, we apply GKPP and KPP-E to several real and random networks. © 2014 Springer Science+Business Media New York.


Jeon M.,Michigan Technological University | Lee J.-H.,Korean German Institute of Technology
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2013

With the rapid development of the touch screen technology, some usability issues of smartphones have been reported [1]. To tackle those user experience issues, there has been research on the use of non-speech sounds on the mobile devices [e.g., 2, 3-7]. However, most of them have focused on a single specific task of the device. Given the varying functions of the smartphone, the present study designed plausibly integrated auditory cues for diverse functions and evaluated user acceptance levels from the ecological interface design perspective. Results showed that sophisticated auditory design could change users' preference and acceptance of the interface and the extent depended on usage contexts. Overall, participants gave significantly higher scores on the functional satisfaction and the fun scales in the sonically-enhanced smartphones than in the no-sound condition. The balanced sound design may free users from auditory pollution and allow them to use their devices more pleasantly. © 2013 Springer-Verlag.


Lee J.-H.,Korean German Institute of Technology
Communications in Computer and Information Science | Year: 2011

The present study has explored the expansibility of using everyday objects for game interface to provide additional multisensory stimuli. The perceived expansibility of game interfaces was investigated by measuring user various responses such as the degree of presence, immersion, and enjoyment. Offering actual sensations assumed in the virtual environment of game could enhance user experience by effectively accessing their emotions. Participants were provided with one of four output modes: basic condition, room light stimulation, chair vibration stimulation, & both light and vibration. Consequently, tactile stimulation has resulted in higher presence and immersion than visual did. The result also suggests that a gamer could experience the expanded interfaces with additional multisensory output, bringing more enjoyment. These tendencies are especially recognized for high immersive tendency group. These findings can be applied to various fields of game interface design. © 2011 Springer-Verlag.


Yang J.,Korean German Institute of Technology
KSII Transactions on Internet and Information Systems | Year: 2014

Conventional mobile state (MS) and base station (BS) association based on average signal strength often results in imbalance of cell load which may require more powerful processor at BSs and degrades the perceived transmission rate of MSs. To deal with this problem, a Markov decision process (MDP) for load balancing in a multi-cell system with multi-carriers is formulated. To solve the problem, exploiting Sarsa algorithm of on-line learning type [12], α -controllable load balancing algorithm is proposed. It is designed to control tradeoff between the cell load deviation of BSs and the perceived transmission rates of MSs. We also propose an ɛ -differential soft greedy policy for on-line learning which is proven to be asymptotically convergent to the optimal greedy policy under some condition. Simulation results verify that the α -controllable load balancing algorithm controls the behavior of the algorithm depending on the choice of α. It is shown to be very efficient in balancing cell loads of BSs with low α. © 2014 KSII.

Loading Korean German Institute of Technology collaborators
Loading Korean German Institute of Technology collaborators