公开(公告)号：US08903124B2

公开(公告)日：2014-12-02

申请号：US12968040

申请日：2010-12-14

申请人： Woon Tack Woo ,  Won Woo Lee ,  Young Min Park

发明人： Woon Tack Woo ,  Won Woo Lee ,  Young Min Park

IPC分类号： G06K9/00 ,  G06K9/62 ,  G06T7/20

CPC分类号： G06K9/6211 ,  G06T7/248 ,  G06T2207/20081

摘要： The present invention relates to an object learning method that minimizes time required for learning an object, an object tracking method using the object learning method, and an object learning and tracking system. The object learning method includes: receiving an image to be learned through a camera to generate a front image by a terminal; generating m view points used for object learning and generating first images obtained when viewing the object from the m view points using the front image; generating second images by performing radial blur on the first images; separating an area used for learning from the second images to obtain reference patches; and storing pixel values of the reference patches.

摘要翻译： 本发明涉及一种使学习对象所需的时间最小化的对象学习方法，使用对象学习方法的对象跟踪方法以及对象学习跟踪系统。 对象学习方法包括：通过摄像机接收要学习的图像以通过终端生成正面图像; 生成用于对象学习的m个视点，并且使用前图像从m个视点观察对象时产生第一图像; 通过对所述第一图像执行径向模糊来生成第二图像; 将用于学习的区域与第二图像分离以获得参考斑块; 并存储参考片的像素值。

公开(公告)号：US20110262003A1

公开(公告)日：2011-10-27

申请号：US12968040

申请日：2010-12-14

申请人： Woon Tack WOO ,  Won Woo Lee ,  Young Min Park

发明人： Woon Tack WOO ,  Won Woo Lee ,  Young Min Park

IPC分类号： G06K9/62 ,  G06K9/00

CPC分类号： G06K9/6211 ,  G06T7/248 ,  G06T2207/20081

摘要： The present invention relates to an object learning method that minimizes time required for learning an object, an object tracking method using the object learning method, and an object learning and tracking system. The object learning method includes: receiving an image to be learned through a camera to generate a front image by a terminal; generating m view points used for object learning and generating first images obtained when viewing the object from the m view points using the front image; generating second images by performing radial blur on the first images; separating an area used for learning from the second images to obtain reference patches; and storing pixel values of the reference patches.

摘要翻译： 本发明涉及一种使学习对象所需的时间最小化的对象学习方法，使用对象学习方法的对象跟踪方法以及对象学习跟踪系统。 对象学习方法包括：通过摄像机接收要学习的图像以通过终端生成正面图像; 生成用于对象学习的m个视点，并且使用前图像从m个视点观察对象时产生第一图像; 通过对所述第一图像执行径向模糊来生成第二图像; 将用于学习的区域与第二图像分离以获得参考斑块; 并存储参考片的像素值。

公开(公告)号：US07607077B2

公开(公告)日：2009-10-20

申请号：US11413101

申请日：2006-04-28

申请人： Won Woo Lee

发明人： Won Woo Lee

IPC分类号： G06F17/00

CPC分类号： G11B27/10

摘要： The present invention relates to a mobile communication terminal and operating method thereof in which data sync error occurring when multimedia data is played can be compensated for. The mobile communication terminal includes a speaker through which played multimedia data is output, a display unit on which text data are output, an multimedia controller that analyzes and plays multimedia data, compensates for error of an output time, which is generated as the multimedia data are played, and outputs compensated text data, and a controller that manages data when multimedia data is played, and controls the driving of the speaker and the display unit. Temporal error between output time information and an audio signal that is actually output, and proportional temporal error occurring in proportion to an output time are calculated. Temporal information is compensated for based on the calculated temporal error and proportional temporal error.

摘要翻译： 移动通信终端及其操作方法技术领域本发明涉及移动通信终端及其操作方法，其中可以补偿播放多媒体数据时发生的数据同步错误。 移动通信终端包括通过其播放的多媒体数据被输出的扬声器，输出文本数据的显示单元，分析和播放多媒体数据的多媒体控制器，补偿作为多媒体数据产生的输出时间的错误 并且输出补偿文本数据;以及控制器，其在播放多媒体数据时管理数据，并且控制扬声器和显示单元的驱动。 计算输出时间信息与实际输出的音频信号之间的时间误差，以及与输出时间成正比的时间误差。 基于计算出的时间误差和比例时间误差补偿时间信息。