Share Email Print
cover

Proceedings Paper

Scene interpretation using a nonmonotonic reasoning approach
Author(s): Miao-Li M. Pai; Robin L. Ying
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Automated image interpretation systems can use non-monotonic reasoning to resolve many of the inherent ambiguities and uncertainties present in real world sensor data. In order to recognize objects of interest in the sensor imagery, an efficient method is needed to match features extracted from image data, such as lines and regions, to corresponding features in known object models. This matching process is based on the accuracy of the extracted features. It is typically the case, however, that these features are themselves uncertain. This uncertainty is present because the image processing algorithms used for feature extraction do not consider global context, are noise sensitive, and use scene dependent parameters. A non-monotonic system identifies global inconsistencies and provides the means to recover from them. It does this by retracting invalid deductions based on the underlying causes of inconsistencies during the model matching process. A nonmonotonic system can also complete a partial match by predicting the existence of additional features in the scene. With this prediction capability, non-monotonic reasoning provides the focus of attention mechanism to confine spatial domains for object detection. Therefore, greatly reduces the processing time as well as increases the accuracy. Such a scene interpretation system can analyze various sensor imagery. This makes it useful in applications such as terrain mapping for robot path finding, parts inspection for product quality control, and object recognition for medical imageiy.

Paper Details

Date Published: 22 September 1993
PDF: 5 pages
Proc. SPIE 2101, Measurement Technology and Intelligent Instruments, (22 September 1993); doi: 10.1117/12.156379
Show Author Affiliations
Miao-Li M. Pai, AT&T Bell Labs. (United States)
Robin L. Ying, AT&T Bell Labs. (United States)


Published in SPIE Proceedings Vol. 2101:
Measurement Technology and Intelligent Instruments

© SPIE. Terms of Use
Back to Top