NASA Logo

NTRS

NTRS - NASA Technical Reports Server

Back to Results
Knowledge-based low-level image analysis for computer vision systemsTwo algorithms for entry-level image analysis and preliminary segmentation are proposed which are flexible enough to incorporate local properties of the image. The first algorithm involves pyramid-based multiresolution processing and a strategy to define and use interlevel and intralevel link strengths. The second algorithm, which is designed for selected window processing, extracts regions adaptively using local histograms. The preliminary segmentation and a set of features are employed as the input to an efficient rule-based low-level analysis system, resulting in suboptimal meaningful segmentation.
Document ID
19890046307
Document Type
Conference Paper
Authors
Dhawan, Atam P. (Houston Univ. TX, United States)
Baxi, Himanshu (Houston Univ. TX, United States)
Ranganath, M. V. (Houston, University TX, United States)
Date Acquired
August 14, 2013
Publication Date
January 1, 1988
Subject Category
COMPUTER PROGRAMMING AND SOFTWARE
Meeting Information
Applications of Artificial Intelligence VI(Orlando, FL)
Distribution Limits
Public
Copyright
Other