• The KIPS Transactions:PartB
• /
• v.11B no.1
• /
• pp.13-20
• /
• 2004

In the real-time automatic face recognition technique, accurate face detection is essential and very important part because it has the effect to face recognition performance. In this paper, we use color information, edge information, and binary information to detect candidate regions of eyes from Input image, and then detect face candidate region using the center point of the detected eyes. We verify both eye candidate region and face candidate region using Support Vector Machines(SVM). It is possible to perform fast and reliable face detection because we can protect false detection through these verification process. From the experimental results, we confirmed the Proposed algorithm in this paper shows excellent face detection rate over 99%.

• Journal of Information Processing Systems
• /
• v.16 no.5
• /
• pp.1008-1014
• /
• 2020

Aiming at the defect detection quality of denim fabric, this paper designs an improved algorithm based on the optimized Gabor filter. Firstly, we propose an improved defect detection algorithm of jean fabric based on the maximum two-dimensional image entropy and the loss evaluation function. Secondly, 24 Gabor filter banks with 4 scales and 6 directions are created and the optimal filter is selected from the filter banks by the one-dimensional image entropy algorithm and the two-dimensional image entropy algorithm respectively. Thirdly, these two optimized Gabor filters are compared to realize the common defect detection of denim fabric, such as normal texture, miss of weft, hole and oil stain. The results show that the improved algorithm has better detection effect on common defects of denim fabrics and the average detection rate is more than 91.25%.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.14 no.1
• /
• pp.35-43
• /
• 2018

MANET composed of only wireless nodes is increasingly utilized in various fields. However, it is exposed to many security vulnerabilities because it doesn't have any infrastructure and transmits data by using multi-hop method. Therefore, MANET should be applied the intrusion detection technique that can detect efficiently malicious nodes and decrease impacts of various attacks. In this paper, we propose a distributed intrusion detection technique that can detect the various attacks while improving the efficiency of attack detection and reducing the false positive rate. The proposed technique uses the cluster structure to manage the information in the center and monitor the traffic of their neighbor nodes directly in all nodes. We use three parameters for attack detection. We also applied an efficient authentication technique using only key exchange without the help of CA in order to provide integrity when exchanging information between cluster heads. This makes it possible to free the forgery of information about trust information of the nodes and attack nodes. The superiority of the proposed technique can be confirmed through comparative experiments with existing intrusion detection techniques.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.8
• /
• pp.4021-4037
• /
• 2018

In network intrusion detection research, two characteristics are generally considered vital to building efficient intrusion detection systems (IDSs): an optimal feature selection technique and robust classification schemes. However, the emergence of sophisticated network attacks and the advent of big data concepts in intrusion detection domains require two more significant aspects to be addressed: employing an appropriate big data computing framework and utilizing a contemporary dataset to deal with ongoing advancements. As such, we present a comprehensive approach to building an efficient IDS with the aim of strengthening academic anomaly detection research in real-world operational environments. The proposed system has the following four characteristics: (i) it performs optimal feature selection using information gain and branch-and-bound algorithms; (ii) it employs machine learning techniques for classification, namely, Logistic Regression, Naïve Bayes, and Random Forest; (iii) it introduces bulk synchronous parallel processing to handle the computational requirements of large-scale networks; and (iv) it utilizes a real-time contemporary dataset generated by the Information Security Centre of Excellence at the University of Brunswick (ISCX-UNB) to validate its efficacy. Experimental analysis shows the effectiveness of the proposed framework, which is able to achieve high accuracy, low computational cost, and reduced false alarms.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.4
• /
• pp.1795-1811
• /
• 2019

In order to achieve rapid and accurate detection of vehicle objects in complex traffic conditions, we propose a novel vehicle detection method. Firstly, more contextual and small-object vehicle information can be obtained by our Joint Feature Network (JFN). Secondly, our Evolved Region Proposal Network (EPRN) generates initial anchor boxes by adding an improved version of the region proposal network in this network, and at the same time filters out a large number of false vehicle boxes by soft-Non Maximum Suppression (NMS). Then, our Mask Network (MaskN) generates an example that includes the vehicle occlusion, the generator and discriminator can learn from each other in order to further improve the vehicle object detection capability. Finally, these candidate vehicle detection boxes are optimized to obtain the final vehicle detection boxes by the Fine-Tuning Network(FTN). Through the evaluation experiment on the DETRAC benchmark dataset, we find that in terms of mAP, our method exceeds Faster-RCNN by 11.15%, YOLO by 11.88%, and EB by 1.64%. Besides, our algorithm also has achieved top2 comaring with MS-CNN, YOLO-v3, RefineNet, RetinaNet, Faster-rcnn, DSSD and YOLO-v2 of vehicle category in KITTI dataset.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2013.06a
• /
• pp.338-341
• /
• 2013

본 논문에서는 동일 주파수 대역 내 DRM(Digital Radio Mondiale) Plus 신호와 FM(Frequeny Modulation) 신호가 모두 서비스 되는 Hybrid 방송 모드에 적합한 DRM Plus 신호 검출 기법을 제안한다. OFDM(Orthogonal Frequency Division Multiplexing) 신호의 Guard-Interval 상관 기법을 그대로 적용할 경우 DRM Plus 신호 검출 성능은 우수하나, FM 신호 수신 시 False Alarm 발생 확률이 증가하는 문제점이 발생한다. 제안한 방법은 Guard-Interval 상관 방법에 Reference 블록을 포함하고 있어, 약한 DRM Plus 신호에서도 판별 확률이 높고, 아날로그 FM 신호 수신 시 False Alarm이 발생할 확률이 낮은 특징이 있다. 모의 실험을 통하여 제안한 기법의 성능을 확인한다.

• Proceedings of the IEEK Conference
• /
• 2008.06a
• /
• pp.669-670
• /
• 2008

In this paper, we present a real-time system to detect abnormal events on gas pipes, based on the signals which are observed through the audio sensors attached on them. First, features are extracted from these signals so that they are robust to noise and invariant to the distance between a sensor and a spot at which an abnormal event like an attack on the gas pipes occurs. Then, a classifier is constructed to detect abnormal events using neural networks. It is a combination of two neural network models, a Gaussian mixture model and a multi-layer perceptron, for the reduction of miss and false alarms. The former works for miss alarm prevention and the latter for false alarm prevention. The experimental result with real data from the actual gas system shows that the proposed system is effective in detecting the dangerous events in real-time with an accuracy of 92.9%.

• The Journal of the Acoustical Society of Korea
• /
• v.10 no.5
• /
• pp.69-76
• /
• 1991

본 논문에서는 통계적 방법에 의한 음소의 자동분할에 관한 알고리즘을 제안하였다. 우선 음성 신호를 AR 모델로 모델링한 후 스펙트럼이 변화하기 전과 변화한 후의 모델에 대해서 likelihood ratio 와 mutual information을 고려한 test statistics 로부터 모델 계수가 변화하는 곳을 예측해 내고 이 곳을 음소의 경계로 판단한다. 이 경우 검파되지 못하는 대부분의 음소는 짧은 자음이었으며 Signed front-to-back maximum area ratio을 이용하여 개선하였다. 또한 false alarm error을 줄이기 위해 두 segment 사이의 distortion 으로부터 smoothing을 하였다. 3명의 화자에 대한 실험 결과 non-detection error는 10%, false alarm error는 20% 정도로 나타났지만 화자간에 알고리즘의 성능 변화가 거의 없으 며 특히 분할된 경계치 분포는 전체 음소의 90% 이상이 이 30ms 이내에 위치하였다.

• Journal of Information Processing Systems
• /
• v.15 no.6
• /
• pp.1265-1276
• /
• 2019

Data deduplication is a common method to improve cloud storage efficiency and save network communication bandwidth, but it also brings a series of problems such as privacy disclosure and dictionary attacks. This paper proposes a secure deduplication scheme for cloud storage based on Bloom filter, and dynamically extends the standard Bloom filter. A public dynamic Bloom filter array (PDBFA) is constructed, which improves the efficiency of ownership proof, realizes the fast detection of duplicate data blocks and reduces the false positive rate of the system. In addition, in the process of file encryption and upload, the convergent key is encrypted twice, which can effectively prevent violent dictionary attacks. The experimental results show that the PDBFA scheme has the characteristics of low computational overhead and low false positive rate.

• Annual Conference of KIPS
• /
• 2002.04a
• /
• pp.767-770
• /
• 2002

일반적으로 방사선 의사들(radialogists)이 폐 노쥴(pulmonary nodule)을 탐지하는 데는 실제적으로 30%의 실패율을 가진다고 알려져 있다. 만약 자동화된 시스템이 체스트 영상에서 의심스런 노쥴들의 위치들을 방사선 의사에게 알려줄 수 있다면 잘못 판단되는 노쥴들의 수를 잠재적으로 줄일 수 있다. 우리는 형태학적 필터들(morphological filters)과 두가지 특징-추출(feature-extraction) 기술들을 포함하는 컴퓨터 자동 처리 시스템을 구현하였다. 본 시스템에서는 첫째로 형태학적 필터(morphological filtering) 처리를 행한다. 이 과정은 원래의 영상에 침식(erosion)과 확장 (dilation)을 연이어서 행하는 것으로 처리가 어려운 X 선 영상을 좀 더 다루기 쉬운 상태로 바꿔주는 역할을 하게 된다. 둘째는 일차적으로 노쥴로서 컴퓨터에 선택된 의심 부분에 가해주는 특징-추출 테스트로서 이 작용은 노쥴로 감지되었으나 실제로는 노쥴이 아닌 경우인 false-positive 갑지들을 줄이기 위해서 사용된다. 그리하여 본 시스템은 노쥴의 정확한 판독이 어려운 폐의 X 선 영상에 적용되어 false-positive 들을 효과적으로 줄임으로써 보다 효율적인 폐 노쥴의 탐지를 가능하게 하였다.