Genetic sequential Monte Carlo methods for 2D motion tracking

遺傳連續蒙特卡羅算法下的二維运动跟踪

Student thesis: Master's Thesis

View graph of relations

Author(s)

  • Zhu YE

Related Research Unit(s)

Detail(s)

Awarding Institution
Supervisors/Advisors
  • Zhi-Qiang LIU (Supervisor)
Award date15 Feb 2007

Abstract

Image sequence based human tracking is a fundamental problem in computer vision research and has been widely studied. The main goal of visual tracking is to imitate the motion sensibility of physical visual system, empower the machine with the ability of perceiving the object motion and their relations in the scene and provide an important way for image sequence understanding. Visual tracking technique has many applications, such as video surveillance and analysis,video based motion analysis and synthesis, motion-based human identification. After more than 40 years’ development, visual tracking technique has made great progress especially in the past ten years. However, practical experience has shown that visual tracking technologies are currently far from mature. A great number of challenges need to be solved before one can implement a robust visual tracking system for commercial applications. Within this context, this work based on the framework of sequential Monte Carlo methods and genetic algorithm try to get insights on some key issues in visual tracking with application scenarios on human motion tracking. Some important technologies and solutions are proposed for robust and practical tracking systems, especially concentrate on how to model and estimate the object complex dynamics and enhance tracker’s robust and efficiency. The contribution of this work is the development of a new optimal solution for the human motion tracking, called Genetic Sequential Monte Carlo (GSMC) method which is an evolutionary online learning algorithm. The experimental result of the new method for tracking hand is an improvement over the standard SMC method with the advantages of recovering from tracking failure and reducing the computational load.

    Research areas

  • Motion perception (Vision), Computer vision, Mathematical models, Monte Carlo method, Image processing