Entity

Time filter

Source Type

Daejeon, South Korea

Park S.I.,Sejong University | Lim S.-J.,SW Content Research Laboratory
ETRI Journal | Year: 2014

In this paper, we present a method for reconstructing a surface mesh animation sequence from point cloud animation data. We mainly focus on the articulated body of a subject - the motion of which can be roughly described by its internal skeletal structure. The point cloud data is assumed to be captured independently without any inter-frame correspondence information. Using a template model that resembles the given subject, our basic idea for reconstructing the mesh animation is to deform the template model to fit to the point cloud (on a frame-by-frame basis) while maintaining inter-frame coherence. We first estimate the skeletal motion from the point cloud data. After applying the skeletal motion to the template surface, we refine it to fit to the point cloud data. We demonstrate the viability of the method by applying it to reconstruct a fast dancing motion. © 2014 ETRI. Source


Yoo J.-H.,SW Content Research Laboratory | Kang B.J.,Hyundai Mobis
IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops | Year: 2014

We describe an image acquisition system based on an integrated dual-sensor with a visible and infrared spectrum, which enables a multimodal biometric system, including non-intrusive iris recognition. To implement this as capable of simultaneously acquiring facial and iris images, a beam splitter for reflecting or transmitting visible and infrared light representing an image of a target object is used along an identical ray path divided into different bands. Namely, the beam splitter divides incident light from an object into the dual-sensor. Accordingly, an image of the iris area and an image of the facial region can be simultaneously acquired from a single image acquisition system, and an iris image can be acquired from a relatively long distance. The iris biometric system is implemented by using the facial detection and iris recognition SDK. In experiments, we have successfully evaluated the performance of the proposed image acquisition system with a non-intrusive iris recognition algorithm. © 2015 IEEE. Source


Kang B.O.,SW Content Research Laboratory | Kang B.O.,Chungbuk National University | Kwon O.-W.,Chungbuk National University
IEICE Transactions on Information and Systems | Year: 2016

We propose a new method to combine multiple acoustic models in Gaussian mixture model (GMM) spaces for robust speech recognition. Even though large vocabulary continuous speech recognition (LVCSR) systems are recently widespread, they often make egregious recognition errors resulting from unavoidable mismatch of speaking styles or environments between the training and real conditions. To handle this problem, a multi-style training approach has been used conventionally to train a large acoustic model by using a large speech database with various kinds of speaking styles and environment noise. But, in this work, we combine multiple sub-models trained for different speaking styles or environment noise into a large acoustic model by maximizing the log-likelihood of the sub-model states sharing the same phonetic context and position. Then the combined acoustic model is used in a new target system, which is robust to variation in speaking style and diverse environment noise. Experimental results show that the proposed method significantly outperforms the conventional methods in two tasks: Non-native English speech recognition for second-language learning systems and noise-robust point-of-interest (POI) recognition for car navigation systems. Copyright © 2016 The Institute of Electronics, Information and Communication Engineers. Source


Lee H.,SW Content Research Laboratory | Yoo J.-H.,SW Content Research Laboratory | Park D.,Korea University
ETRI Journal | Year: 2014

Most hyper-ellipsoidal clustering (HEC) approaches use the Mahalanobis distance as a distance metric. It has been proven that HEC, under this condition, cannot be realized since the cost function of partitional clustering is a constant. We demonstrate that HEC with a modified Gaussian kernel metric can be interpreted as a problem of finding condensed ellipsoidal clusters (with respect to the volumes and densities of the clusters) and propose a practical HEC algorithm that is able to efficiently handle clusters that are ellipsoidal in shape and that are of different size and density. We then try to refine the HEC algorithm by utilizing ellipsoids defined on the kernel feature space to deal with more complex-shaped clusters. The proposed methods lead to a significant improvement in the clustering results over K-means algorithm, fuzzy Cmeans algorithm, GMM-EM algorithm, and HEC algorithm based on minimum-volume ellipsoids using Mahalanobis distance. © 2014 ETRI. Source


Youn T.-Y.,SW Content Research Laboratory | Chang K.-Y.,SW Content Research Laboratory
ETRI Journal | Year: 2014

Fairness of exchange is a significant property for secure online transactions, and a fair exchange scheme is a useful tool for ensuring the fairness of exchanges conducted over networks. In this paper, we propose an ID-based optimistic fair exchange scheme based on the RSA function, one which is designed by combining a wellknown RSA-based signature scheme and the (naive) RSA function. Note that the main contribution of this paper is to give the first provably secure ID-based fair exchange scheme based on the RSA function, whose security can be proved under fully formalized security models. Our scheme has the following additional strongpoints. The scheme is setup-free; hence, there is no registration step between a user and an arbitrator. Moreover, the proposed scheme is designed in an ID-based setting; thus, it is possible to eliminate the need for certificates and avoid some related problems. © 2014 ETRI. Source

Discover hidden collaborations