权益分类	功能权益	普通用户	{{item.name}}会员
{{category.name}}	{{benefitItem.name}}

Dynamical System Approach to Optical Flow Estimation

光流估计的动力系统方法

基本信息

批准号：
23656273
负责人：
SEBE Noboru
金额：
$ 2.33万
依托单位：
Kyushu Institute of Technology
依托单位国家：
日本
项目类别：
Grant-in-Aid for Challenging Exploratory Research
财政年份：
2011
资助国家：
日本
起止时间：
2011 至 2013
项目状态：
已结题

项目摘要

Optical flow is a vector which represents the motion of objects in images. The ordinary methods estimate optical flow instantaneously from two images. On the other hand, the proposed estimation method considers the intensity preservation law as an equation of time evolution of dynamical system and estimate from time sequence of images. The proposed method provides not only a new strategy for optical flow estimation but also some advantages. The first advantage of the method is that it simultaneously estimates the confidence of estimation, which enable us to evaluate the estimation results without the knowledge of true values. The second advantage of the method is that it detects discontinuity of optical flow, which is a higher level information of object motion.

光流是一种表示图像中物体运动的矢量。通常的方法是从两幅图像中瞬时估计光流。另一方面，所提出的估计方法将强度保持定律视为动力系统的时间演化方程，并根据图像的时间序列进行估计。该方法不仅为光流场估计提供了一种新的策略，而且具有一定的优势。该方法的第一个优点是同时估计估计的置信度，这使得我们能够评估估计结果，而不需要知道真值。该方法的第二个优点是它检测光流的不连续性，这是物体运动的更高层次的信息。