• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2005.05a
• /
• pp.599-606
• /
• 2005

Novelty detection involves identifying novel patterns. They are not usually available during training. Even if they are, the data quantity imbalance leads to a low classification accuracy when a supervised learning scheme is employed. Thus, an unsupervised learning scheme is often employed ignoring those few novel patterns. In this paper, we propose two ways to make use of the few available novel patterns. First, a scheme to determine local thresholds for the Self Organizing Map boundary is proposed. Second, a modification of the Learning Vector Quantization learning rule is proposed so that allows one to keep codebook vectors as far from novel patterns as possible. Experimental results are quite promising.

• 한국정보통신설비학회:학술대회논문집
• /
• 2009.08a
• /
• pp.387-390
• /
• 2009

Wireless Sensor Network (WSN) can easily compromised from attackers because it has the limited resource and deployed in exposed environments. When the sensitive packets are occurred such as enemy's movement or fire alarm, attackers can selectively drop them using a compromised node. It brings the isolation between the basestation and the sensor fields. To detect selective forwarding attack, Xiao, Yu and Gao proposed checkpoint-based multi-hop acknowledgement scheme (CHEMAS). The check-point nodes are used to detect the area which generating selective forwarding attacks. However, CHEMAS has static probability of selecting check-point nodes. It cannot achieve the flexibility to coordinate between the detection ability and the energy consumption. In this paper, we propose the control method for the number fo check-point nodes. Through the control method, we can achieve the flexibility which can provide the sufficient detection ability while conserving the energy consumption.

• Journal of Advanced Navigation Technology
• /
• v.16 no.1
• /
• pp.145-150
• /
• 2012

This paper explains a performance analysis of an error detection system running on nested session management of cloud computing collaboration environment using rule-based DEVS modeling and simulation techniques. In DEVS, a system has a time base, inputs, states, outputs, and functions. This paper explains the design and implementation of the FDA(Fault Detection Agent). FDA is a system that is suitable for detecting software error for multimedia remote control based on nested session management of cloud computing collaboration environment.

• Journal of Biomedical Engineering Research
• /
• v.17 no.2
• /
• pp.263-273
• /
• 1996

The time instants, at which QRS complexes are detected, are used in the electrocardioyam rhythm analysis. Hence, it is necessary that all QRS complexes are detected and that no other waves or artifacts are wrongly labeled as such. These time instants are also used in other tasks as an indication of the location of significant events in the ECG. For example, the QRS typification algorithm uses these points to define the region of interest for complex comparison and alignment. When waveform recognition is drone for each complex, these points are used to define search intervals in which the onset and the end of the QRS nmplex have to be found This paper proposes the method for the detection of QRS complexes and decision rule for the classification scheme. The efficiency of the detection is demonstrated with the aid of an internationally validated CSE(Common Standard for Quantitative Electrocardioyaph) data set 3 and 4.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2002.11b
• /
• pp.1031-1034
• /
• 2002

본 논문에서는 커널수준의 침입탐지를 위한 동적 침입탐지 규칙 변경 기법을 제안한다. 제안하는 기법은 침입탐지 규칙은 규칙타입 프로토콜 타입, 패킷 헤더와 패킷 페이로드에 대한 검사를 수행하기 위한 규칙들로 세분화하여 LVR로 표현하고 이들 LVR이 계층적으로 구성된 IDRL로 관리한다. 침입탐지는 IDRL을 이용하여 수행하며, 규칙에 대한 변경은 변경된 규칙에 대한 LVR을 구성하고 LV를 이용한 포인터 변경을 이용하여 IDRL에 반영하는 방법이다. 제안하는 기법은 IDRL을 이용한 침입탐지와 탐지규칙의 변경을 IDRL에 최소한의 비용으로 수행하고, LVR을 이용하여 침입탐지 규칙을 디스크와 메모리에 동일한 형태로 저장 및 관리하여 탐지규칙 초기화 비용과 변경 비용을 최소화할 수 있다. 이를 통하여 보다 안전한 커널 수준에서의 네트워크 보안을 위한 효율적인 동적 침입탐지 규칙 변경을 지원할 수 있다는 장점을 가진다.

• Proceedings of the KSRS Conference
• /
• 2003.11a
• /
• pp.402-404
• /
• 2003

We present two methods for the automatic selection of the threshold values in unsupervised change detection. Both methods consist of the same two procedures: 1) to determine the parameters of Gaussian mixtures from a difference image or ratio image, 2) to determine threshold values using the Bayesian rule for minimum error. In the first method, the Expectation-Maximization algorithm is applied for estimating the parameters of the Gaussian mixtures. The second method is based on the iterative thresholding that successively employs thresholding and estimation of the model parameters. The effectiveness and applicability of the methods proposed here are illustrated by an experiment on the multi-temporal KOMPAT-1 EOC images.

• International Journal of Control, Automation, and Systems
• /
• v.4 no.3
• /
• pp.372-381
• /
• 2006

The purpose of this study was to develop a process management system to manage ingot fabrication and improve ingot quality. The ingot is the first manufactured material of wafers. Trace parameters were collected on-line but measurement parameters were measured by sampling inspection. The quality parameters were applied to evaluate the quality. Therefore, preprocessing was necessary to extract useful information from the quality data. First, statistical methods were used for data generation. Then, modeling was performed, using the generated data, to improve the performance of the models. The function of the models is to predict the quality corresponding to control parameters. Secondly, rule extraction was performed to find the relation between the production quality and control conditions. The extracted rules can give important information concerning how to handle the process correctly. The dynamic polynomial neural network (DPNN) and decision tree were applied for data modeling and rule extraction, respectively, from the ingot fabrication data.

• The KIPS Transactions:PartC
• /
• v.11C no.4
• /
• pp.455-462
• /
• 2004

Network attacks are various types with development of internet and are a new types. The existing intrusion detection systems need a lot of efforts and costs in order to detect and respond to unknown or modified attacks because of detection based on signatures of known attacks. In this paper, we present a design and implementation for mining prototype system to predict unknown or modified attacks through network protocol attributes analysis. In order to analyze attributes of network protocols, we use the association rule and the frequent episode. The collected network protocols are storing schema of TCP, UDP, ICMP and integrated type. We are generating rules that can predict the types of network attacks. Our mining prototype in the intrusion detection system aspect is useful for response against new attacks as extra tool.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.21 no.2
• /
• pp.101-110
• /
• 2011

Recently, an attack to an application incapacitates the intrusion detection rule, the defense policy for a network and database and induces intrusion incidents. Thus, it is necessary to study integration security to ensure the security of an internal network and database from that attack. This article is about building an integration security system to prevent an attack to an application set with intrusion detection rules. It responds to network-based attack through detection, disperses attack with the internal integration security system through virtual clustering and load balancing, and sets up defense policy for attacking destination packets, analyzes and records attack packets, and updates rules through monitoring and analysis. Moreover, this study establishes defense policy according to attacking types to settle access traffic through virtual machine partition policy and suggests an integration security system applied to prevent attack and tests its defense. The result of this study is expected to provide practical data for integration security defense for hacking attack from outside.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.10
• /
• pp.419-426
• /
• 2022

In this paper, we propose a method to improve the accuracy of 3D human pose estimation model in various move motions. Existing human pose estimation models have some problems of jitter, inversion, swap, miss that cause miss coordinates when estimating human poses. These problems cause low accuracy of pose estimation models to detect exact coordinates of human poses. We propose a method that consists of detection and correction methods to handle with these problems. Deep learning-based outlier detection method detects outlier of human pose coordinates in move motion effectively and rule-based correction method corrects the outlier according to a simple rule. We have shown that the proposed method is effective in various motions with the experiments using 2D golf swing motion data and have shown the possibility of expansion from 2D to 3D coordinates.