Using Energy Flow Information for Video Segmentation

ZHANG Zhong-wei; LIU Gui-zhong; LI Hong-liang; LI Yong-li

您当前的位置：

首页 >

文章列表页 >

Using Energy Flow Information for Video Segmentation

更新时间：2025-07-16

- Using Energy Flow Information for Video Segmentation
- Acta Electronica Sinica Vol. 33, Issue 1, Pages: 177-180(2005)
- 作者机构：
  
  西安交通大学电子与信息学院信息与通讯工程系,陕西,西安,710049
- 作者简介：
- 基金信息：
- DOI：
  CLC： TN9411.73
- Published：2005
- 稿件说明：
移动端阅览
ZHANG Zhong-wei, LIU Gui-zhong, LI Hong-liang, et al. Using Energy Flow Information for Video Segmentation[J]. Acta Electronica Sinica, 2005, 33(1): 177-180.
DOI：

ZHANG Zhong-wei, LIU Gui-zhong, LI Hong-liang, et al. Using Energy Flow Information for Video Segmentation[J]. Acta Electronica Sinica, 2005, 33(1): 177-180. DOI：

摘要

通过把视频序列中图像的像素看作运动的气体分子

本文提出了一个图像温度场的概念

并利用描述气体分子做无规则运动的麦克斯韦-玻尔兹曼分布律

来描述图像序列中像素的运动规律.与传统光流信息相比

由图像温度场估算出的能流信息更加精确地描述了图像间的运动场

把该能流信息进一步应用到视频镜头的自动分割中

仿真实验结果表明

基于能流分析的视频分割方法更加准确地检测出了镜头的切变和渐变.

Abstract

By looking up pixels of image sequence as those moving gas molecules

a novel concept

image temperature

has been proposed to describe 2-D image motion during video analysis.The idea is motivated by the well-known Maxwell-Boltzman Distribution Law in the gas dynamic theories.The Energy Flow Field (EFF) estimated by the image temperature presents more accurate motion vectors than the traditional Optical Flow Field (OFF).When the information of EFF is applied for video segmentation

the experimental results show that EFF provides better performance.

关键词

Keywords

references

Views

935

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Accurate Motion Estimation of High Speed Moving Target about Stepped-Frequency Chirp Signal

Optimization on Motion Estimation Algorithm Based on H.264

Multi-Frame Super-Resolution Reconstruction Based on Anisotropic Markov Random Field Modeling

A New Motion Estimation and Intra-frame Prediction Method for H.264 Video Coding

Related Author

王辉

郑南宁

徐维朴

FONT

WANG Hui

ZHENG Nan-ning

XU Wei-pu

LI Shen

Related Institution

The Institute of Artificial Intelligence and Robotics,Xi'an Jiaotong University

School of Electronic Science and Engineering,National University of Defense Technology,,Hunnan

Harbin Institute of Technology Shenzhen Graduate School

School of Computer Science and Technology,Nanjing University of Science and Technology

Department of Communication and Control Engineering Hunan Institute of Humanities Science and Technology Loudi Hunan China

⁰