Announcement for Downloading full text filePlease respect the Copyright Act.
All digital full text dissertation and theses from this website are authorized the copyright owners. These copyrighted full-text dissertation and theses can be only used for academic, research and non-commercial purposes. Users of this website can search, read, and print for personal usage. In respect of the Copyright Act of the Republic of China, please do not reproduce, distribute, change, or edit the content of these dissertations and theses without any permission. Please do not create any work based upon a pre-existing work by reproduction, Adaptation, Distribution or other means.
URN etd-1119104-113452 Statistics This thesis had been viewed 3713 times. Download 1845 times. Author Lung-Sheng Jang Author's Email Address No Public. Department Electrical Engineering Year 2003 Semester 2 Degree Master Type of Document Master's Thesis Language English Page Count 51 Title Detection and Tracking of Human in Dynamic Scenes Keyword motion estimation motion detection human tracking human detection human detection human tracking motion detection motion estimation Abstract ABSTRACT
Detection and tracking human in image sequences is a fundamental and crucial research topic. It has vast applications in many visual systems, e.g. video surveillance, traffic monitoring, human detection, and video editing, etc. In this thesis, we aim at explore and implement a detection and tracking of human system in dynamic scenes.
The system consists of four major blocks: (i) Dominant motion estimation: We adopt the hierarchical dominant motion estimation method in the form of a Gaussian pyramid to estimate the background motion. (ii) Motion detection: An efficient and effective method based on background subtraction is adopted to detect moving objects. (iii) Image blending: Once the dominant motion of the consecutive frames is obtained, we then use a triangular weighting function to blend overlapping regions into a panoramic background image. (iv) Tracking of moving objects: Finally, we utilize a set of three geometric features extracted from the detected moving objects to track moving object in complex environment.
We demonstrate our tracking system with real-life image sequences. Experimental results show that our algorithm is fair robust and effective.
Advisor Committee Jia-Ching Cheng - advisor
Jiun-Shian Jou - co-chair
Yung-An Gau - co-chair
Files Date of Defense 2004-07-30 Date of Submission 2004-11-19