作者: In Su Kim , Dai Jin Kim , Jae Sik Yun
DOI:
关键词: Pattern recognition 、 Mathematics 、 Covariance 、 Artificial intelligence 、 Point (geometry) 、 Feature (computer vision) 、 Object (computer science) 、 Computer vision 、 Orientation (computer vision) 、 Image (mathematics)
摘要: Disclosed are object recognizing method and apparatus. The may comprise extracting a feature point of an to be recognized in input image; generating plurality images by orientation information region adjacent the normalizing based information; deriving covariance descriptor on images; image comparing with reference stored advance. Thus, can fast robustly changes size image, rotations illumination, whereby recognition performance enhanced.