Content area
Full text
Int J Comput Vis (2011) 93: 141161 DOI 10.1007/s11263-010-0343-9
Predicate Logic Based Image Grammars for Complex Pattern Recognition
Vinay Shet Maneesh Singh Claus Bahlmann
Visvanathan Ramesh Jan Neumann Larry Davis
Received: 15 October 2009 / Accepted: 13 April 2010 / Published online: 28 September 2010 Springer Science+Business Media, LLC 2010
Abstract Predicate logic based reasoning approaches provide a means of formally specifying domain knowledge and manipulating symbolic information to explicitly reason about different concepts of interest. Extension of traditional binary predicate logics with the bilattice formalism permits the handling of uncertainty in reasoning, thereby facilitating their application to computer vision problems. In this paper, we propose using rst order predicate logics, extended with a bilattice based uncertainty handling formalism, as a means of formally encoding pattern grammars, to parse a set of image features, and detect the presence of different
Application and experimental validation of the reasoning framework on aerial images has been funded by US Government contract# NBCHC080029. Aerial images provided by DigiGlobe.
J. Neumann contributed to the work presented in this paper while he was afliated with Siemens Corporate Research.
V. Shet ( ) M. Singh C. Bahlmann V. Ramesh
Siemens Corporate Research, a division of Siemens Corporation, 755 College Road East, Princeton, NJ 08540, USAe-mail: mailto:[email protected]
Web End [email protected]
M. Singhe-mail: mailto:[email protected]
Web End [email protected]
C. Bahlmanne-mail: [email protected]
V. Rameshe-mail: [email protected]
J. NeumannStreamsage/Comcast, 1110 Vermont Ave NW, Washington, DC 20005, USAe-mail: mailto:[email protected]
Web End [email protected]
L. DavisUniversity of Maryland, Department of Computer Science, College Park, MD 20742, USAe-mail: [email protected]
patterns of interest. Detections from low level feature detectors are treated as logical facts and, in conjunction with logical rules, used to drive the reasoning. Positive and negative information from different sources, as well as uncertainties from detections, are integrated within the bilattice framework. We show that this approach can also generate proofs or justications (in the form of parse trees) for each hypothesis it proposes thus permitting direct analysis of the nal solution in linguistic form. Automated logical rule weight learning is an important aspect of the application of such systems in the computer vision domain. We propose a rule weight optimization method which casts the instantiated inference tree as a knowledge-based neural network, interprets rule uncertainties as link weights in the...