• Journal of Institute of Control, Robotics and Systems
• /
• v.21 no.12
• /
• pp.1109-1114
• /
• 2015

This paper presents an improved learning-based visual inspection method for auto parts inspection in severe lighting changes. Automobile sunroof frames are produced automatically by robots in most production lines. In the sunroof frame manufacturing process, there is a quality problem with some parts such as volts are missed. Instead of manual sampling inspection using some mechanical jig instruments, a learning-based machine vision system was proposed in the previous research[1]. But, in applying the actual sunroof frame production process, the inspection accuracy of the proposed vision system is much lowered because of severe illumination changes. In order to overcome this capricious environment, some selective feature vectors and cascade classifiers are used for each auto parts. And we are able to improve the inspection accuracy through the re-learning concept for the misclassified data. The effectiveness of the proposed visual inspection method is verified through sufficient experiments in a real sunroof production line.

• Proceedings of the IEEK Conference
• /
• 2002.07a
• /
• pp.188-191
• /
• 2002

To reduce the bit-rate of video sequences by removing temporal redundancy, motion estimation techniques have been developed. However, the high computational complexity of the problem makes such techniques very difficult to be applied to high-resolution applications in a real time environment. For this reason, low computational complexity motion estimation algorithms are viable solutions. If a priori knowledge about the motion of the current block is available before the motion estimation, a better starting point for the search of n optimal motion vector on be selected and also the computational complexity will be reduced. In this paper, we present an adaptive block matching algorithm based on temporal correlations of consecutive image frames that defines the search pattern and the location of initial starting point adaptively to reduce computational complexity. Experiments show that, comparing with DS(Diamond Search) algorithm, the proposed algorithm is about 0.1∼0.5(㏈) better than DS in terms of PSNR and improves as much as 50% in terms of the average number of search points per motion estimation.

• Journal of Korea Multimedia Society
• /
• v.21 no.5
• /
• pp.626-637
• /
• 2018

Human exchanges information not only through words, but also through body gesture or hand gesture. And they can be used to build effective interfaces in mobile, virtual reality, and augmented reality. The past 2D gesture recognition research had information loss caused by projecting 3D information in 2D. Since the recognition of the gesture in 3D is higher than 2D space in terms of recognition range, the complexity of gesture recognition increases. In this paper, we proposed a real-time gesture recognition deep learning model and application in 3D space using deep learning technique. First, in order to recognize the gesture in the 3D space, the data collection is performed using the unity game engine to construct and acquire data. Second, input vector normalization for learning 3D gesture recognition model is processed based on deep learning. Thirdly, the SELU(Scaled Exponential Linear Unit) function is applied to the neural network's active function for faster learning and better recognition performance. The proposed system is expected to be applicable to various fields such as rehabilitation cares, game applications, and virtual reality.

• Journal of Information Processing Systems
• /
• v.9 no.1
• /
• pp.173-188
• /
• 2013

In Facial Expression Recognition Systems (FERS), only particular regions of the face are utilized for discrimination. The areas of the eyes, eyebrows, nose, and mouth are the most important features in any FERS. Applying facial features descriptors such as the local binary pattern (LBP) on such areas results in an effective and efficient FERS. In this paper, we propose an automatic facial expression recognition system. Unlike other systems, it detects and extracts the informative and discriminant regions of the face (i.e., eyes, nose, and mouth areas) using Haar-feature based cascade classifiers and these region-based features are stored into separate image files as a preprocessing step. Then, LBP is applied to these image files for facial texture representation and a feature-vector per subject is obtained by concatenating the resulting LBP histograms of the decomposed region-based features. The one-vs.-rest SVM, which is a popular multi-classification method, is employed with the Radial Basis Function (RBF) for facial expression classification. Experimental results show that this approach yields good performance for both frontal and near-frontal facial images in terms of accuracy and time complexity. Cohn-Kanade and JAFFE, which are benchmark facial expression datasets, are used to evaluate this approach.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.6 no.11
• /
• pp.2849-2865
• /
• 2012

In order to block mobile junk messages automatically, many studies on spam filters have applied machine learning algorithms. Most previous research focused only on the accuracy rate of spam filters from the view point of the algorithm used, not on individual user's preferences. In terms of individual taste, the spam filters implemented on a mobile device have the advantage over spam filters on a network node, because it deals with only incoming messages on the users' phone and generates no additional traffic during the filtering process. However, a spam filter on a mobile phone has to consider the consumption of resources, because energy, memory and computing ability are limited. Moreover, as time passes an increasing number of feature words are likely to exhaust mobile resources. In this paper we propose a spam filter model distributed between a users' computer and smart phone. We expect the model to follow personal decision boundaries and use the uniform resources of smart phones. An authorized user's computer takes on the more complex and time consuming jobs, such as feature selection and training, while the smart phone performs only the minimum amount of work for filtering and utilizes the results of the information calculated on the desktop. Our experiments show that the accuracy of our method is more than 95% with Na$\ddot{i}$ve Bayes and Support Vector Machine, and our model that uses uniform memory does not affect other applications that run on the smart phone.

• Annual Conference on Human and Language Technology
• /
• 1998.10c
• /
• pp.73-80
• /
• 1998

본 연구는 자연어 질의의 형태 및 구문 정보를 바탕으로 불리언 질의를 생성하는데 그 목적을 둔다. 일반적으로 대부분의 상용정보검색시스템은 입력형식을 검색성능이 종은 불리언 형태로 하고 있으나, 일반 사용자는 자신이 원하는 정보를 불리언 형태로 표현하는데 익숙하지 않다. 그러므로 본 정보검색시스템은 자연어 질의를 기본 입력형태로 하여 사용자의 편의성을 높이고, 이 질의를 범주문법에 기반한 구문분석 결과에 의해 복합명사를 고려한 불리언 형태로 변환하여 검색을 수행함으로써 시스템의 검색 성능의 향상을 도모하였다. 정보검색 실험용 데이터 모음인 KTSET2.0으로 실험한 결과 본 논문에서 제안한 자연어 질의로부터 자동 생성된 불리언 질의의 검객성능이 KTSET2.0에서 제공하는 수동으로 추출한 불리언 질의보다 8% 더 우수한 성능을 보였고, 기존 자연어질의 시스템이 수용해온 방법인 형태소 분석을 거쳐 불용어를 제거한 후 Vector 모델을 적용하여 검색을 수행한 경우보다는 23% 더 나은 성능을 보였다.

• Journal of the Semiconductor & Display Technology
• /
• v.4 no.1 s.10
• /
• pp.61-64
• /
• 2005

This study presents a new algorithm to detect the facial feature in a color image entered from the mobile device with complex backgrounds and undefined distance between camera's location and the face. Since skin color model with Hough transformation spent approximately 90$\%$ of running time to extract the fitting ellipse for detection of the facial feature, we have changed the approach to the simple geometric vector operation, called a TS(Triangle-Square) transformation. As the experimental results, this gives benefit of reduced run time. We have similar ratio of face detection to other methods with fast speed enough to be used on real-time identification system in mobile environments.

• Journal of Industrial Technology
• /
• v.38 no.1
• /
• pp.41-52
• /
• 2018

Recently, object detection is a critical function for any system that uses computer vision and is widely used in various fields such as video surveillance and self-driving cars. However, the conventional methods can not detect the objects clearly because of the dynamic background change in the beach. In this paper, we propose a new technique to detect humans correctly in the dynamic videos like shores. A new background modeling method that combines spatial GMM (Gaussian Mixture Model) and temporal GMM is proposed to make more correct background image. Also, the proposed method improve the accuracy of people detection by using SVM (Support Vector Machine) to classify people from the objects and KCF (Kernelized Correlation Filter) Tracker to track people continuously in the complicated environment. The experimental result shows that our method can work well for detection and tracking of objects in videos containing dynamic factors and situations.

• ETRI Journal
• /
• v.42 no.3
• /
• pp.311-323
• /
• 2020

Encrypted traffic classification plays a vital role in cybersecurity as network traffic encryption becomes prevalent. First, we briefly introduce three traffic encryption mechanisms: IPsec, SSL/TLS, and SRTP. After evaluating the performances of support vector machine, random forest, naïve Bayes, and logistic regression for traffic classification, we propose the combined approach of entropy estimation and artificial neural networks. First, network traffic is classified as encrypted or plaintext with entropy estimation. Encrypted traffic is then further classified using neural networks. We propose using traffic packet's sizes, packet's inter-arrival time, and direction as the neural network's input. Our combined approach was evaluated with the dataset obtained from the Canadian Institute for Cybersecurity. Results show an improved precision (from 1 to 7 percentage points), and some application classification metrics improved nearly by 30 percentage points.

• Proceedings of the KSME Conference
• /
• 2002.08a
• /
• pp.717-720
• /
• 2002

The present study is to evaluate the performances of flow velocity and wall shear stress in the stenosed coronary artery using human in vivo hemodynamic Parameters and computer simulation. Initial and follow-up coronary angiographics in the patients with angulated coronary stenosis are performed. Follow-up coronary angiogram demonstrated significant difference in the percent of diameter in the stenosed coronary between two groups ($Group\;1:\;40.3{\%},\;Group\;2:\;25.5{\%}$). Flow-velocity wave obtained from in vivo intracoronary Doppler ultrasound data is used for the boundary condition for the computer simulation. Spatial and temporal variations of flow velocity vector and recirculation area are drawn throughout the selected segment of coronary models. The WSS of pre- and post-intracoronary stenting are calculated from three-dimensional computer simulation. Then negative shear stresses area on 3D simulation we noted on the inner wall of the post-stenotic area before stenting. The negative WSS is disappeared after stenting. High spatial and temporal WSS before stenting fell into within physiologic WSS after stenting. This finding was prominent in Model 2. The present study suggest that hemodynamic forces exerted by pulsatile coronary circulation termed WSS might affect on the evolution of atherosclerosis within the angulated vascular curvature. The local recirculation area which has low or negative WSS, might lead to progression of atherosclerosis.