가우시안 배경혼합모델을 이용한 Tracking기반 사고검지 알고리즘의 적용 및 평가

오주택,민준영 한국도로학회 2012 한국도로학회논문집 Vol.14 No.3

자동사고검지 알고리즘의 대부분은 사고가 발생했을 때 사고로 검지하지 못하고, 혼잡으로 검지하는 경우가 많다는 문제점을 가지고 있다. 또한 교통정보센터 운영자들은 교통사고검지시스템을 운영하면서 대부분 CCTV 육안감시 또는 운전자들의신고에 의존하여 사고처리를 하고 있는 실정이다. 그 이유는 현재 운영되고 있는 교통사고검지시스템에서는 실제 사고가 아닌데도 불구하고, 사고라는 오검지 경고가 많이 발생되어 시스템 전체의 신뢰도가 떨어진다는 문제점이 있기 때문이다. 다시 말해 교통사고검지시스템의 알고리즘은 검지율(Detection probability)이 높아야 함과 동시에, 오검지율(False alarm probability)은 낮아야 하고, 정확한 사고지점과 시간을 검지해 낼 수 있어야 한다. 이에 본 연구는 검지율을 높이고 동시에,오검지율을 낮추는 방법으로 기 개발된 가우시안 혼합모델(Gaussian Mixture Model)과 개별차량 Tracking을 이용하여 개발한 사고검지 알고리즘을 교통정보센터 관리시스템(Center Management System)에 적용하고, 실제 교통상황에서 사고검지율과 오검지의 빈도를 측정하여 그 효과를 검증 및 평가하고자 한다. Most of Automatic Accident Detection Algorithm has a problem of detecting an accident as traffic congestion. Actually, center、s managers deal with accidents depend on watching CCTV or accident report by drivers even though they run the Automatic Accident Detection system. It is because of the system、s detecting errors such as detecting non-accidents as accidents, and it makes decreasing in the system、s overall reliability. It means that Automatic Accident Detection Algorithm should not only have high detection probability but also have low false alarm probability, and it has to detect accurate accident spot. The study tries to verify and evaluate the effectiveness of using Gaussian Mixture Model and individual vehicle tracking to adapt Accident Detection Algorithm to Center Management System by measuring accident detection probability and false alarm probability、s frequency in the real accident.

A Network Intrusion Detection Model Based on K-means Algorithm and Information Entropy

Gao Meng,Li Dan,Wang Ni-hong,Liu Li-chen 보안공학연구지원센터 2014 International Journal of Security and Its Applicat Vol.8 No.6

Many factors could influence the clustering performance of K-means algorithm, selection of initial cluster centers was an important one, traditional method had a certain degree of randomness in dealing with this problem, for this purpose, information entropy was introduced into the process of cluster centers selection, and a fusion algorithm combining with information entropy and K-means algorithm was proposed, in which, information entropy value was used to measure the similarity degree among records, the least similar record would be regarded as a cluster center. In addition, a network intrusion detection model was built, it could make cluster centers change dynamically along with the network changes, and the model could real-time update the cluster centers according to actual needs. Experiment results show that the improved algorithm proposed is better than the traditional K-means algorithm in detection ratio and false alarm ratio, and the network intrusion detection model is proved to be feasible.

Wheat Rows Detection Based on Machine Vision

Changdong Ma,Hongtu Zhao,Xiaojie Wang 보안공학연구지원센터 2016 International Journal of Signal Processing, Image Vol.9 No.9

This study describes a new method for wheat rows detection at the middle growth stage based on Machine Vision. The algorithm includes three steps: (i) vegetation segmentation, (ii) centers points extraction and (iii) wheat rows detection. In the first step, color images were transformed into gray-level images and Otsu’s method was used to implement binarization. Based on the fact that the corresponding center points on two adjacent horizontal scanning lines can’t have a large deviation, in the second step, we firstly extracted the initial center points on the first scanning line based on a sliding window, and then gave a small shift based on positions of the initial center points which have been extracted on the previous scanning line to extract the center points for the next scanning line. Finally, the Randomized Hough transform (RHT) method was employed to locate the wheat rows. Test results indicate that the proposed method can effectively detect the wheat rows at the middle growth stage.

A New Finger-Knuckle-Print ROI Extraction Method Based on Two-Stage Center Point Detection

보안공학연구지원센터(IJSIP) 보안공학연구지원센터 2015 International Journal of Signal Processing, Image Vol.8 No.2

Finger-knuckle-print (FKP) pattern has been utilized in biometric recognition systems. This paper proposes a new FKP region of interest (ROI) extraction method based on two-stage center point detection. In our method, a center point preliminary detection is constructed to capture the center point initially. Then, an efficient center point positioned algorithm is presented to locate the center point more precisely in real time. Finally, we select the Hong Kong Polytechnic University (PolyU) database to verify the efficiency of the proposed method. The experimental results show that the proposed method can extract ROI not only accurately but also in real time.

Eye pupil detection system using an ensemble of regression forest and fast radial symmetry transform with a near infrared camera

Jeong, Mira,Nam, Jae-Yeal,Ko, Byoung Chul Elsevier 2017 Infrared physics & technology Vol.85 No.-

<P><B>Abstract</B></P> <P>In this paper, we focus on pupil center detection in various video sequences that include head poses and changes in illumination. To detect the pupil center, we first find four eye landmarks in each eye by using cascade local regression based on a regression forest. Based on the rough location of the pupil, a fast radial symmetric transform is applied using the previously found pupil location to rearrange the fine pupil center. As the final step, the pupil displacement is estimated between the previous frame and the current frame to maintain the level of accuracy against a false locating result occurring in a particular frame. We generated a new face dataset, called Keimyung University pupil detection (KMUPD), with infrared camera. The proposed method was successfully applied to the KMUPD dataset, and the results indicate that its pupil center detection capability is better than that of other methods and with a shorter processing time.</P> <P><B>Highlights</B></P> <P> <UL> <LI> Using a near IR camera to detect eye pupil region. </LI> <LI> Using an ensemble of regression forest and fast radial symmetry transform. </LI> <LI> The pupil displacement is estimated to maintain the level of accuracy. </LI> <LI> Pupil detection accuracy is higher than those of related algorithms. </LI> </UL> </P>

Center point prediction using Gaussian elliptic and size component regression using small solution space for object detection

( Yuantian Xia ),( Shuhan Lu ),( Longhe Wang ),( Lin Li ) 한국인터넷정보학회 2023 KSII Transactions on Internet and Information Syst Vol.17 No.8

The anchor-free object detector CenterNet regards the object as a center point and predicts it based on the Gaussian circle region. For each object's center point, CenterNet directly regresses the width and height of the objects and finally gets the boundary range of the objects. However, the critical range of the object's center point can not be accurately limited by using the Gaussian circle region to constrain the prediction region, resulting in many low-quality centers' predicted values. In addition, because of the large difference between the width and height of different objects, directly regressing the width and height will make the model difficult to converge and lose the intrinsic relationship between them, thereby reducing the stability and consistency of accuracy. For these problems, we proposed a center point prediction method based on the Gaussian elliptic region and a size component regression method based on the small solution space. First, we constructed a Gaussian ellipse region that can accurately predict the object's center point. Second, we recode the width and height of the objects, which significantly reduces the regression solution space and improves the convergence speed of the model. Finally, we jointly decode the predicted components, enhancing the internal relationship between the size components and improving the accuracy consistency. Experiments show that when using CenterNet as the improved baseline and Hourglass-104 as the backbone, on the MS COCO dataset, our improved model achieved 44.7%, which is 2.6% higher than the baseline.

Design and Analysis of Improved Iris-Based Gaze Estimation Model

Sharma, Anjana,Abrol, Pawanesh Korean Institute of Information Scientists and Eng 2018 Journal of Computing Science and Engineering Vol.12 No.2

The detection accuracy of gaze direction mainly depends on the performance of features extracted from eye images. Limitations on the estimation of gaze direction include harmful infrared (IR) light, expensive devices, static thresholding, inappropriate and complex segmentation techniques, corneal reflections, etc. In this study, an efficient appearance cum feature-based detection model, namely, iris center-based gaze estimation (ICGE), has been proposed. The model is an extension of the earlier proposed glint-based gaze direction estimation (GDE) model and overcomes the above limitations. The ICGE model has been analyzed for GDE based on iris center coordinates using a local adaptive thresholding technique. An indigenous database using more than two hundred images of different subjects on a five quadrant map screen generates almost 90% accurate results for iris and gaze quadrant detection. The distinguishing features of the low cost, non-intrusive proposed model include a lack of IR and affordable ubiquitous H/W designing, large subject-camera distance and screen dimensions, no glint dependency, and many more. The proposed model also shows significantly better results in the lower periphery corners of the quadrant map than traditional models. In addition, aside from the comparison with the GDE model, the proposed model has also been compared with other existing techniques.

봉쇄-탐지-대응 기반 보안관제 대시보드 설계

한충희,Han, Choong-Hee 한국융합보안학회 2021 융합보안 논문지 Vol.21 No.3

효율적인 보안관제센터 운영을 위해서는 보안관제 대시보드의 표준화가 반드시 필요하다. 보안관제 대시보드는 24시간 365일 내내 함께 생활해야 하는 보안관제근무자들에게 많이 활용되도록 구성해야 한다. 또한 보안관제센터의 업무활동을 종합적으로 표출할 수 있어야 한다. 추가적으로 보안관제센터의 업무활동들을 쉽게 설명할 수 있어야 할 것이다. 이에 본 논문에서 사례기관에 실제 적용한 봉쇄·탐지·대응 기반의 보안관제 대시보드 디자인을 설명하고자 한다. 이를 통해 불필요한 귀빈 맞춤형 대시보드 구성작업에 대한 노력과 시간을 줄이고 보안관제센터의 효율적인 운영에 이바지하고자 한다. Standardization of the security operation dashboard is essential for efficient operation of security operation center. The security operation dashboard should be configured so that it is widely used by security operation workers who have to live together 24 hours a day, 365 days a year. In addition, it must be able to comprehensively express the business activities of the security operation center. In addition, it should be possible to easily explain the business activities of the security operation center. Therefore, in this paper, we would like to explain the design of a security control dashboard based on blockade, detection, and response that is actually applied to case organizations in the power sector. Through this, it is intended to reduce the effort and time required for configuring a custom dashboard for VIPs, and contribute to the efficient operation of the security operation center.

Design and Analysis of Improved Iris-Based Gaze Estimation Model

Anjana Sharma,Pawanesh Abrol 한국정보과학회 2018 Journal of Computing Science and Engineering Vol.12 No.2

The detection accuracy of gaze direction mainly depends on the performance of features extracted from eye images. Limitations on the estimation of gaze direction include harmful infrared (IR) light, expensive devices, static thresholding, inappropriate and complex segmentation techniques, corneal reflections, etc. In this study, an efficient appearance cum feature-based detection model, namely, iris center-based gaze estimation (ICGE), has been proposed. The model is an extension of the earlier proposed glint-based gaze direction estimation (GDE) model and overcomes the above limitations. The ICGE model has been analyzed for GDE based on iris center coordinates using a local adaptive thresholding technique. An indigenous database using more than two hundred images of different subjects on a five quadrant map screen generates almost 90% accurate results for iris and gaze quadrant detection. The distinguishing features of the low cost, non-intrusive proposed model include a lack of IR and affordable ubiquitous H/W designing, large subject-camera distance and screen dimensions, no glint dependency, and many more. The proposed model also shows significantly better results in the lower periphery corners of the quadrant map than traditional models. In addition, aside from the comparison with the GDE model, the proposed model has also been compared with other existing techniques.

Salient Object Detection via Multiple Random Walks

( Jiyou Zhai ),( Jingbo Zhou ),( Yongfeng Ren ),( Zhijian Wang ) 한국인터넷정보학회 2016 KSII Transactions on Internet and Information Syst Vol.10 No.4

In this paper, we propose a novel saliency detection framework via multiple random walks (MRW) which simulate multiple agents on a graph simultaneously. In the MRW system, two agents, which represent the seeds of background and foreground, traverse the graph according to a transition matrix, and interact with each other to achieve a state of equilibrium. The proposed algorithm is divided into three steps. First, an initial segmentation is performed to partition an input image into homogeneous regions (i.e., superpixels) for saliency computation. Based on the regions of image, we construct a graph that the nodes correspond to the superpixels in the image, and the edges between neighboring nodes represent the similarities of the corresponding superpixels. Second, to generate the seeds of background, we first filter out one of the four boundaries that most unlikely belong to the background. The superpixels on each of the three remaining sides of the image will be labeled as the seeds of background. To generate the seeds of foreground, we utilize the center prior that foreground objects tend to appear near the image center. In last step, the seeds of foreground and background are treated as two different agents in multiple random walkers to complete the process of salient object detection. Experimental results on three benchmark databases demonstrate the proposed method performs well when it against the state-of-the-art methods in terms of accuracy and robustness.