• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.8 no.2
• /
• pp.36-44
• /
• 2009

In this paper, we suggest a system of detecting a vehicle with lane violation, which can detect the vehicle with lane violation, by using the feature point tracking. The whole algorithm in the suggested system of detecting a vehicle with lane violation is composed of three stages such as feature extraction, register and tracking in feature for the tracking-targeted vehicle, and detecting a vehicle with lane violation. In the stage of feature extraction, the feature is extracted from the inputted image by sing the feature-extraction algorithm available for the real-time processing. The extracted features are again selected the racking-targeted feature. The registered feature is tracked by using NCC(normalized cross correlation). Finally, whether or not lane violation is finally detected by using information on the tracked features. As a result of experimenting the suggested system by using the acquired image in the section with a ban on intervention, the excellent performance was shown with 99.09% for positive recognition ratio and 0.9% for error ratio. The fast processing speed could be obtained in 34.48 frames per second available for real-time processing.

• Journal of Korea Multimedia Society
• /
• v.10 no.8
• /
• pp.981-996
• /
• 2007

In this paper, we describe a method that can recognize gestures by obtaining motion features information with principal factor analysis from sequential gesture images. In the algorithm, firstly, a two dimensional silhouette region including human gesture is segmented and then geometric features are extracted from it. Here, global features information which is selected as some meaningful key feature effectively expressing gestures with principal factor analysis is used. Obtained motion history information representing time variation of gestures from extracted feature construct one gesture subspace. Finally, projected model feature value into the gesture space is transformed as specific state symbols by grouping algorithm to be use as input symbols of HMM and input gesture is recognized as one of the model gesture with high probability. Proposed method has achieved higher recognition rate than others using only shape information of human body as in an appearance-based method or extracting features intuitively from complicated gestures, because this algorithm constructs gesture models with feature factors that have high contribution rate using principal factor analysis.

• Journal of the Korea Convergence Society
• /
• v.10 no.3
• /
• pp.39-47
• /
• 2019

In this paper, we proposed extracting modified Histogram of Oriented Gradients (HOG) features using background removal when tracking pedestrians in real time. HOG feature extraction has a problem of slow processing speed due to large computation amount. Background removal has been studied to improve computation reductions and tracking rate. Area removal was carried out using S and V channels in HSV color space to reduce feature extraction in unnecessary areas. The average S and V channels of the video were removed and the input video was totally dark, so that the object tracking may fail. Histogram equalization was performed to prevent this case. HOG features extracted from the removed region are reduced, and processing speed and tracking rates were improved by extracting clear HOG features. In this experiment, we experimented with videos with a large number of pedestrians or one pedestrian, complicated videos with backgrounds, and videos with severe tremors. Compared with the existing HOG-SVM method, the proposed method improved the processing speed by 41.84% and the error rate was reduced by 52.29%.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2007.06a
• /
• pp.149-152
• /
• 2007

비전 연구에 있어서 객체 추적은 무엇보다도 중요시 되어 왔다. 특히 비디오 감시 시스템에서의 객체 추적은 매우 중요하다. 본 논문에서는 영상 내에서 움직이는 객체를 추출하고 객체내의 다중 후보블록의 통계적 특징을 이용한 추적 시스템을 구성하였다. 객체를 추적하기 위해서는 먼저 움직이는 객체 추출이 선행되어야 한다. 객체 추출은 영상 내에서 배경 프레임과 매 프레임에서의 현재 프레임간의 차 연산에 의한 가중치를 이용하여 객체의 움직임을 판단하고 추출하였다. 움직이는 객체는 본 논문에서 제안한 다중 후보 블록 알고리즘을 수행하여 추적에 필요한 통계 값을 획득한다. 통계 값으로는 방향성에 필요한 블록의 중심 좌표 값과 객체추적에 필요한 객체간의 매칭 정도를 사용하였다. 본 논문에서 제안한 추적 시스템은 민감한 빛의 변화에도 강건하였으며, 특정 블록에 대해서만 연산 수행을 수행하므로 컴퓨터의 연산을 줄여 실시간 추적도 가능하다.

• Journal of Internet Computing and Services
• /
• v.5 no.1
• /
• pp.59-66
• /
• 2004

Symmetry region searching can extract face region without a prior information in an image by using symmetric. However, this method requires a plenty of the computation time because the mask size to process symmetry region searching must be larger than the size of object such as eye, nose and mouth in face. in this paper, it proposed symmetric by using symmetry region searching and Wavelet Transform to reduce computation time of symmetry region searching, and It was applied to this method in an original image. To extract exact face region, we also experimented face region searching by using domain division in extraction region.

• Proceedings of the Korean Information Science Society Conference
• /
• 2000.04b
• /
• pp.119-121
• /
• 2000

문자 정보 및 X-Ray, MRI, CT등과 같은 의료영상 정보를 취급하는 의료정보 시스템에서의 유사객체 검색을 지원하는 협력 의료정보 시스템의 설계에 대하여 논한다. 이를 위해 객체간 의미적 관련성을 기반으로 한 유사도 자동 추출 방식 및 지식베이스 구성 방식을 제안하고 이를 활용한 유사객체 검색에 대하여 논한다. 특히 의료영상을 객체 값으로 갖는 경우 객체간 유사도는 영상처리의 특징추출 방식에 의해 추출된 영상내에 출현하는 공간 객체의 위치, 면적, 둘레, 공간 객체간의 위상 관계 등의 공간 속성을 이용한다. 여기서 공간적 위치에 근거한 유사도는 공간 위치를 대표하는 Hilbert값의 분포와 빈도를 토대로 계산한다.

• Journal of the Korea Society of Computer and Information
• /
• v.13 no.5
• /
• pp.125-133
• /
• 2008

As society develops, the importance of safety for individuals and facilities in public places is getting higher. Not only the areas such as the existing parking lot, bank and factory which require security or crime prevention but also individual houses as well as general institutions have the trend to increase investment in guard and security. This study suggests face feature extract and the method to simply divide face region and head region that are import for face recognition by using color transform. First of all, it is to divide face region by using color transform of Y image of YIQ image and head image after dividing head region with K image among CMYK image about input image. Then, it is to extract features of face by using labeling after Log calculation to head image. The clearly divided head and face region can easily classify the shape of head and face and simply find features. When the algorism of the suggested method is utilized, it is expected that security related facilities that require importance can use it effectively to guard or recognize people.

• Journal of the Korea Society of Computer and Information
• /
• v.18 no.12
• /
• pp.21-26
• /
• 2013

In this paper, We are suggested methods of feature extraction in molecular images. The result of image transform DWT examination by suggested method, we are obtained as follows. 1-level and 2-levels of decomposition results showed the composition of the low frequency region. But, 3-level decomposition results did not appear in the data component is almost. Observed not with the naked eye is not, but the 3-level output data values of the results were decomposed. We are printed the horizontal and vertical directions of low-frequency region of the data, the high frequency region of the horizontal and vertical data, and diagonal high frequency region of the horizontal and vertical directions data. If the output data using molecular imaging and CT, PET, MR imaging will be compared with the data.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.25 no.4
• /
• pp.279-287
• /
• 2007

The increased availability of portable, low-cost, high resolution video equipments have resulted in a rapid growth of the applications for video sequences. These video devices can be mounted in handhold unit, mobile unit and airborne platforms like maned or unmaned helicopter, plane, airship, etc. This paper describes the feasibility fur generating image map from the experimental results we designed to track the interested points extracted by KLT operator in the neighboring frames and implement image matching for each frames taken from UAV (Unmaned Aerial Vehicle). In the image registration for neighbourhood frames of aerial video, the results demonstrate the successful rate of matching slightly decreases as the drift between frames increases, and also that the stable photographing is more important matching condition than the pixel shift.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2008.05a
• /
• pp.337-342
• /
• 2008

최근 들어 기존의 녹색 바탕 차량 번호판에서, 흰색 바탕의 신 차량 번호판으로 교체되고 있다. 하지만 아직 기존 차량 번호판이 신 차량 번호판으로 전면 교체되지 않아 두 번호판 모두 사용되고 있다. 따라서 주차관리 시스템, 속도위반, 신호 위반 등 무인 카메라를 이용한 시스템에서, 기존 차량 번호판과 신 차량 번호판의 특징에 맞는 인식 시스템이 요구된다. 본 논문에서는 이러한 문제를 해결하기 위해 기존 차량 번호판과 신 차량 번호판을 통합한, 지능형 차량 번호판 인식 시스템을 제안한다. 무인 카메라에서 획득된 차량 영상에서 번호판의 색상 정보를 이용하여 기존 차량 번호판과 신 차량 번호판을 구분한다. 기존 차량 번호판인 경우에는 HSI 컬러 공간을 이용하여 이진화를 적용하며, 신 차량 번호판인 경우에는 블록 이진화를 적용한다. 이진화된 영상을 대상으로 차량의 형태학적 특징을 이용하여 잡음을 제거한 후, 차량 번호판 영역을 추출한다. 추출된 차량 번호판 영역에 대해 Labeling 알고리즘을 적용하여 개별 문자를 추출한다. 추출된 개별 문자는 FCM 알고리즘을 적용하여 인식한다. 제안된 차량 번호판 추출 및 인식 방법의 성능을 평가하기 위해 160장의 기존 차량 영상과 100장의 신 차량 영상을 대상으로 실험한 결과, 제안된 차량 번호판 추출 및 인식 방법이 실험을 통해서 효율적인 것을 확인하였다.