文献类型：专利

英文题名：Method for diving video moving target for use in local descriptors, involves measuring similarity between pixels by OSF graph cutting model, and obtaining updated seed point of foreground and background object by utilizing cutting model

作者：CHEN Z;WANG J;LI C;GAO T;DAI C;SHENG B

机构：[1]UNIV EAST CHINA SCI & TECHNOLOGY

申请号：CN106997599-A

申请日：2017-04-17

公开日：2017-08-01

语种：英文

收录：DERWENT

摘要：NOVELTY - The method involves processing an original video into multiple image sequence. A value of OSF frame image is calculated by optimization of changeable model. A target position in two frames of the sequence is predicted. A foreground seed points and background seed points is selected by OSF descriptor tracking foreground target. The foreground and background target is updated based on seed point position. A similarity between pixels is measured by OSF graph cutting model. An updated seed point of the foreground object and the background object is obtained by utilizing the cutting model. USE - Method for diving video moving target for use in local descriptors. DESCRIPTION OF DRAWING(S) - The drawing shows a flowchart of a method for diving video moving target. '(Drawing includes non-English language text)'