Automated feature extraction for object recognition
- Ilya Levner
- Vadim Bulitko
- Lihong Li
- Greg Lee
- Russ Greiner, Dept of Computing Science; PI of AICML
Automated image interpretation is an importanttask in numerous applications ranging from securitysystems to natural resource inventorization basedon remote-sensing. Recently, a second generationof adaptive machine-learned image interpretationsystems have shown expert-level performance inseveral challenging domains. While demonstratingan unprecedented improvement over hand-engineeredand first generation machine-learned systems in termsof cross-domain portability, design-cycle time, androbustness, such systems are still severely limited.This paper reviews the anatomy of the state-of-theartMulti resolution Adaptive Object Recognitionframework (MR ADORE) and presents extensionsthat aim at removing the last vestiges of humanintervention still present in the original design ofADORE. More specifically, feature selection is stilla task performed by human domain experts therebyprohibiting automatic creation of image interpretationsystems. This paper focuses on autonomous featureextraction methods aimed at removing the need forhuman expertise in the feature selection process.
Citation
I. Levner, V. Bulitko, L. Li, G. Lee, R. Greiner. "Automated feature extraction for object recognition". Image and Vision Computing New Zealand (IVCNZ), New Zealand, pp 309-313, November 2003.| Keywords: | MrAdore, machine learning |
| Category: | In Conference |
BibTeX
@incollection{Levner+al:IVCNZ03,
author = {Ilya Levner and Vadim Bulitko and Lihong Li and Greg Lee and Russ
Greiner},
title = {Automated feature extraction for object recognition},
Pages = {309-313},
booktitle = {Image and Vision Computing New Zealand (IVCNZ)},
year = 2003,
}Last Updated: May 29, 2007Submitted by Staurt H. Johnson
