http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
휴대폰과 센서를 통한 Structure-from-Motion 배율에 대한 연구
유수곤(Soogon Yoo),감동욱(Donguk Kam),김대욱(DaeWook Kim),이건우(Gunwoo Lee) (사)한국CDE학회 2015 한국 CAD/CAM 학회 학술발표회 논문집 Vol.2015 No.하계
3D reconstruction and scanning of images have recently become researched in computer graphics and manufacturing industry. Especially, Structure-from-Motion, called SfM, is one of the most popular methods for finding 3D structure from the captured images, even by using mobile phones. However, SfM does not exactly calculate the scale of images compared with the real scenes. In order to check an exact scale of the scenes, IMU sensors in mobile phones can be used. Also, extended Kalman filter is applied for camera location estimation from sensor data, and from comparison of camera locations between filtered data and SfM, a scale factor can be calculated and applied in estimating scene depth or mapping scenes. We propose the method that a scale factor between captured images and real scenes is calculated by using sensors in mobile phones, as well as camera locations are estimated in real world case.
김대욱(Dae-wook Kim),감동욱(Dong-uk Kam),유수곤(Soogon Yoo),이건우(Kunwoo Lee) (사)한국CDE학회 2016 한국 CAD/CAM 학회 학술발표회 논문집 Vol.2016 No.동계
Smart application for a novel recognition system of commercial tire was suggested. Consumers who buy their car tire through online usually make wrong order because they hardly have knowledge about interpreting tire information, which is carved on left side of tire. The suggested system only needs smart phone and provides the tire information (maker & model) from the picture taken by consumer’s phone. At first, wheel rim of tire is detected using texture-based segmentation algorithm and rectified image is acquired. After the large text area is cropped, bounding boxes which have large possibility of including text were extracted. Through HOG feature machine learning and convolutional neural network deep learning, tire information can be recognized. With suggested system, consumers can easily understand the tire information only by taking the picture.