http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
기하학적 모델 및 색상 모델에 기반한 이동물체 추적에 관한 연구
김우현(Woo-hyun Kim),기창두(Chang-doo Kee) 한국자동차공학회 2003 한국자동차공학회 지부 학술대회 논문집 Vol.2003 No.5
As there have been many interests about a real-time detection in a face recognition field, the research of face detection is being progressed lively through skin-color model for quick processing. However, if the color of background and skin-color is almost alike, unexpected recognition will happen. Also, face recognition which is a important field of pattern-recognition includes complex difficulty like size, shape, and position of target object, background, and lighting, but the faces give us wide information like inherent face-color, head movement, face shape, eyes, a nose, a lip.<br/> Using this much information, we present face-tracking method in this paper by combining skin-color chaser and ellipse chaser in given conditions.<br/> We use ellipse chaser as the first algorithm, and skin-color chaser which has human's skin-color. By experiment using two chasers, this algorithm shows robustness in face-tracking<br/> <br/>
Hough Transform 을 이용한 차선인식과 응용
장윤(Yun Jang),기창두(Chang Doo Kee) 한국자동차공학회 2002 한국자동차공학회 Symposium Vol.2002 No.11
Automated driving control needs information of visuality. At first, lane detection is neede in order to study the Advanced Safety Vehicle(ASV) or ITS project. In this study, Hough transform algorithm is used to detect the lanes of a road. This algorithm is effective to detect a line, because main principle is simple and method of statistics is used. In this method, we don't have to know a preliminary knowledge of the location of object to find a boundary line. But this transform method demands much amount of space of memory. To reduce this memory space, scanning region is cut down into small segments. Finally, to verify the usefulness of this algorithm, experiments are executed for autonomous mobile robot's drive.