• Proceedings of the Korean Information Science Society Conference
• /
• 2004.10b
• /
• pp.697-699
• /
• 2004

Stochastic Neighbor Embedding(SNE) is a probabilistic method of mapping high-dimensional data space into a low-dimensional representation with preserving neighbor identities. Even though SNE shows several useful properties, the gradient-based naive SNE algorithm has a critical limitation that it is very slow to converge. To overcome this limitation, faster optimization methods should be considered by using trust region method we call this method fast TR SNE. Moreover, this paper presents a couple of useful optimization methods(i.e. conjugate gradient method and Newton's method) to embody fast SNE algorithm. We compared above three methods and conclude that TR-SNE is the best algorithm among them considering speed and stability. Finally, we show several visualizing experiments of TR-SNE to confirm its stability by experiments.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.6 no.10
• /
• pp.2548-2566
• /
• 2012

This paper derives two main end-to-end performance metrics, namely the spatial capacity density and the average end-to-end delay of the multi-hop wireless ad hoc networks with multi-antenna communications. Based on the closed-form expressions of these performance metrics, three hopping strategies, i.e., the closest neighbor, the furthest neighbor and the randomly selected neighbor hopping strategies have been investigated. This formulation provides insights into the relations among node density, diversity gains, number of hops and some other network design parameters which jointly determine network performances, and a method of choosing the best hopping strategy which can be formulated from a network design perspective.

• Proceedings of the Korean Information Science Society Conference
• /
• 2006.10c
• /
• pp.56-61
• /
• 2006

최근 공간 네트워크 데이터베이스를 위한 질의처리 알고리즘에 관한 연구가 많이 진행되어 왔다. 그러나 현재 좌표-기반 질의에 대한 연구는 활발히 진행중인 반면, 경로-기반 질의에 대한 연구는 매우 미흡한 실정이다. 공간 네트워크 데이터베이스에서는 이동객체가 공간 네트워크상에서만 이동하기 때문에 경로-기반 질의의 유용성이 매우 증대되므로, 경로-기반 질의에 대한 효율적인 질의처리 알고리즘 연구가 필수적이다. 따라서 본 논문에서는 경로-기반 질의의 대표적인 방법인 In-Route Nearest Neighbor 질의처리 알고리즘을 분석하여 기존 연구에서 고려하지 않은 시간 및 공간제약을 고려한 경로-기반 질의처리 알고리즘을 설계한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2005.11a
• /
• pp.735-738
• /
• 2005

기존의 주성분 분석을 이용한 물체 인식 기술은 모델 영상내의 각각의 물체의 대표 값을 만든 후에 실험 영상을 물체 공간에 투영 시켜서 나온 성분과 대표 값의 거리를 비교하여 인식하게 된다. 그러나 단순히 기존의 방법인 point to point 방식인 단순 거리 계산은 오차가 많기 때문에 본 논문에서는 개선된 Class to Class방식인 k-Nearest Neighbor을 이용하여 몇 개의 연속적인 입력영상에 대해 각 각의 모델영상들을 인식의 단위로 이용하였다. 또한, 물체 인식을 하는데 있어 본 논문에서 제안한 주성분 분석법을 물체 영상 자체를 계산하여 인식하는 게 아니라 물체 영상 공간이라는 고유 공간을 구성한 후에 단지 기여도가 큰 8개의 벡터로만 인식을 수행하기 때문에 자원 축소의 효과까지 얻을 수 있었다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2011.06c
• /
• pp.188-190
• /
• 2011

Image interpolation is a technique that pervades many an application. Interpolation is almost never the goal in itself, yet it affects both the desired results and the ways to obtain them. In this paper, we proposed a technique that is capable to find out the error when the common two methods (bilinear and nearest neighbor interpolation) are applied on an image for rotation. The proposed technique also includes the comparison results of bilinear interpolation and nearest neighbor interpolation. Among them nearest neighbor interpolation gives us a better result regarding to the enhancement and due to least error. The error is found by using Mean Square Error (MSE).

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2019.05a
• /
• pp.155-156
• /
• 2019

Traditional machine learning based emotion analysis methods such as Classification and Regression Tree (CART), Support Vector Machine (SVM), and k-nearest neighbor classification (kNN) are less accurate. In this paper, we propose an improved kNN classification method. Improved methods and data normalization achieve the goal of improving accuracy. Then, three classification algorithms and an improved algorithm were compared based on experimental data.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.2
• /
• pp.742-762
• /
• 2015

In this study, we propose a distance metric learning approach called discriminant metric learning (DML) for face verification, which addresses a binary-class problem for classifying whether or not two input images are of the same subject. The critical issue for solving this problem is determining the method to be used for measuring the distance between two images. Among various methods, the large margin nearest neighbor (LMNN) method is a state-of-the-art algorithm. However, to compensate the LMNN's entangled data distribution due to high levels of appearance variations in unconstrained environments, DML's goal is to penalize violations of the negative pair distance relationship, i.e., the images with different labels, while being integrated with LMNN to model the distance relation between positive pairs, i.e., the images with the same label. The likelihoods of the input images, estimated using DML and LMNN metrics, are then weighted and combined for further analysis. Additionally, rather than using the k-nearest neighbor (k-NN) classification mechanism, we propose a verification mechanism that measures the correlation of the class label distribution of neighbors to reduce the false negative rate of positive pairs. From the experimental results, we see that DML can modify the relation of negative pairs in the original LMNN space and compensate for LMNN's performance on faces with large variances, such as pose and expression.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.4 no.4
• /
• pp.529-546
• /
• 2010

Greedy forwarding is the key mechanism of geographic routing and is one of the protocols used most commonly in wireless sensor networks. Greedy forwarding uses 1-hop local information to forward packets to the destination and does not have to maintain the routing table, and thus it takes small overhead and has excellent scalability. However, the signal intensity reduces exponentially with the distance in realistic wireless sensor network, and greedy forwarding consumes a lot of energy, since it forwards the packets to the neighbor node closest to the destination. Previous proposed greedy forwarding protocols are the sender-based greedy forwarding that a sender selects a neighbor node to forward packets as the forwarding node and hence they cannot guarantee energy efficient forwarding in unpredictable wireless environment. In this paper, we propose the receiver-based greedy forwarding called RGF where one of the neighbor nodes that received the packet forwards it by itself. In RGF, sender selects several energy efficient nodes as candidate forwarding nodes and decides forwarding priority of them in order to prevent unnecessary transmissions. The simulation results show that RGF improves delivery rate up to maximum 66.8% and energy efficiency, 60.9% compared with existing sender-based greedy forwarding.

• Proceedings of the IEEK Conference
• /
• 2008.06a
• /
• pp.255-256
• /
• 2008

Distributed localization algorithms are required for large-scale wireless sensor network applications. In this paper, we introduce an efficient algorithm, termed weighted neighbor-node distribution localization(WNDL), which emphasizes simple refinement and low system-load for low-cost and low-rate wireless sensors. We inspect WNDL algorithm through MATLAB simulation.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2007.05a
• /
• pp.1104-1106
• /
• 2007

IPv6 환경에서 SEcure Neighbor Discovery(SEND) 프로토콜은 Neighbor Discovery 프로토콜의 위협 요소를 제거하고 단말 노드의 안전한 통신을 위한 환경을 제공함으로써 보다 안정적으로 인터넷 서비스를 제공할 수 있게 되었다. 하지만 SEND 프로토콜을 Mobile IPv6와 연동하는 과정에서 모바일 노드가 다른 네트워크로 이동하였을 경우 SEND 프로토콜 동작 과정에서의 취약점을 발견하고 기존의 방식과 차별화된 RSA기반의 위임 정보에 기반한 대리 서명 기법을 사용하여 문제를 해결하였다.