Date of Award
5-1-2011
Degree Name
Master of Science
Department
Computer Science
First Advisor
Cheng, Qiang
Abstract
Image analysis and classification have become a very active research topic in recent decades. The practical applications where classification of images is of importance range from technology to geographical, civil, military, and biological sciences. This article proposes an efficient image classification algorithm based on a recently developed regression method. Groves of regression trees are used with image feature sampling to learn patterns from images that are later used to define a probabilistic framework for classification purposes. This new method can generalize well with small number of training subwindows. By comparing several publicly available datasets, this paper shows how the new method outperforms several state of the art techniques.
Access
This thesis is only available for download to the SIUC community. Others should
contact the interlibrary
loan department of your local library.