• Journal of KIISE:Software and Applications
• /
• v.32 no.2
• /
• pp.128-139
• /
• 2005

Moving object boundary is very important for moving object segmentation. But the moving object boundary shows broken boundary We invent a novel space-oriented boundary linking algorithm to link the broken boundary The boundary linking algorithm forms a quadrant around the terminating pixel in the broken boundary and searches forward other terminating pixel to link within a radius. The boundary linking algorithm guarantees shortest distance linking. We also register the background from image sequence. We construct two object masks, one from the result of boundary linking and the other from the registered background, and use these two complementary object masks together for moving object segmentation. We also suppress the moving cast shadow using Roberts gradient operator. The major advantages of the proposed algorithms are more accurate moving object segmentation and the segmentation of the object which has holes in its region using these two object masks. We experiment the algorithms using the standard MPEG-4 test sequences and real video sequence. The proposed algorithms are very efficient and can process QCIF image more than 48 fps and CIF image more than 19 fps using a 2.0GHz Pentium-4 computer.

• Journal of Broadcast Engineering
• /
• v.24 no.6
• /
• pp.992-1001
• /
• 2019

In many applications, images acquired from multiple cameras are stitched to form an image with a wide viewing angle. We propose a method of estimating a seam line using motion information to stitch multiple images without distortion of the moving object. Existing seam estimation techniques usually utilize an energy function based on image gradient information and parallax. In this paper, we propose a seam estimation technique that prevents distortion of moving object by adding temporal motion information, which is calculated from the gradient information of each frame. We also propose a measure to quantify the distortion level of stitched images and to verify the performance differences between the existing and proposed methods.

• Proceedings of the KAIS Fall Conference
• /
• 2009.12a
• /
• pp.605-607
• /
• 2009

본 논문은 감시 시스템의 객체 추적 시, 정확한 객체 추출을 위해 블록 기반으로 객체를 추적하는 방법을 제안한다. 객체 움직임 추적은 주어진 환경에 따라 변수가 많고, 변수를 대처하는 알고리즘을 많이 추가 할 경우 실시간 추적에 어려움이 발생한다. 특히 배경이 조명이나 바람 등의 환경적 요인에 의해 변화되는 문제는 객체를 추적하는데 가장 큰 문제점이다. 특히 사람이나 멧돼지의 경우 움직임에 의한 객체 구성 요소의 흔들림에 의해 고정 블록의 연산에 의해 움직임 객체를 추적할 때 정확한 객체의 윤곽선을 검출하기 힘들다. 따라서 연속되는 프레임에서 전체 화면의 차영상을 이용하여 움직임 관심 영역을 설정하고, 관심 영역에 해당하는 블록들을 분석하였다. 이를 기반으로 움직임 객체의 최외곽 사각형의 객체 영역을 추출하여 기존 고정 블록 방법에 의한 객체 추적보다 좀 더 정확하게 객체를 추출하고 추적할 수 있다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.41 no.9
• /
• pp.1120-1131
• /
• 2016

In this paper, we propose a smoker recognition algorithm, detecting smokers in a video sequence in order to prevent fire accidents. We use description-based method in hierarchical approaches to recognize smoker's activity, the algorithm consists of background subtraction, object detection, event search, event judgement. Background subtraction generates slow-motion and fast-motion foreground image from input image using Gaussian mixture model with two different learning-rate. Then, it extracts object locations in the slow-motion image using chain-rule based contour detection. For each object, face is detected by using Haar-like feature and smoke is detected by reflecting frequency and direction of smoke in fast-motion foreground. Hand movements are detected by motion estimation. The algorithm examines the features in a certain interval and infers that whether the object is a smoker. It robustly can detect a smoker among different objects while achieving real-time performance.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.860-863
• /
• 2021

고정 카메라 환경에서 움직이는 객체만을 탐지하는 것은 비디오 모니터링의 중요한 응용 분야이다. 본 논문에서는 비디오의 특성인 움직임 정보가 포함된 영상에서 GMM을 이용하여 움직인 돼지와 움직이지 않은 돼지의 위치를 대략적으로 구분하고, 추가적인 영상 처리 기법과 딥러닝 기반 객체 탐지기를 적용한 박스 단위 객체 탐지 결과를 활용하여 움직인 돼지의 외곽선을 보정한다. 돈사에서 촬영된 비디오 데이터로 실험한 결과, 제안 방법은 효과적으로 움직인 돼지를 탐지할 수 있음을 확인하였다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.5
• /
• pp.631-637
• /
• 2010

In this paper, We propose a method of motion interpretation of pedestrian for active traffic signal control. We detect pedestrian object in a movie of crosswalk area by using the code book method and acquire contour information. To do this stage fast, we use parallel processing based on CUDA (Compute Unified Device Architecture). And we remove shadow which causes shape distortion of objects. Shadow removed object is judged by using the hilbert scan distance whether to human or noise. If the objects are judged as a human, we analyze pedestrian objects' motion, face area feature, waiting time to decide that they have intetion to across a crosswalk for pdestrians. Traffic signal can be controlled after judgement.

• Journal of Broadcast Engineering
• /
• v.2 no.2
• /
• pp.114-124
• /
• 1997

In object based coding, binary shape ccx:ling plays an important role by ccx:ling the outer shape of object. Here we propose a new shape ccx:ling tool, which enccx:les the outline of shape from a baseline. Different from 2-D (Vertex) shape ccx:ling algorithms. the proposed method encodeds the data that are extracted in a I-D fashion. The enccx:led data consist of the starting position, distance lists, and turning point lists. In the lossless ccx:ling mode, every contour pixel is input for ccx:ling, whereas variable sampling has been employed to enccx:le fewer contour pixels while preserving reasonable distortion. For interframe ccx:ling, a fast motion compensation was achieved by use of distance and turning point lists. Subjective viewing tests proved that the proposed method outperforms the current shape ccx:ling standard, CAE, in MPEG-4. In objective results for compression efficiency, the proposed method was significantly better in intraframe coding than CAE, whereas CAE was better in interframe ccx:ling.