Directed Acyclic Graph Matcher

Research Interests
Bio and Picture

Curriculum Vitae

Buenos Aires Links

Contact Information

Home (or click on the eye)


	DAG Matcher
	Introduction The directed acyclic graph (DAG) matcher is an extension to the indexing and matching framework for hierarchical structures proposed by Shokoufandeh et al. The indexing performance is improved by a vote accumulation algorithm that efficiently solves a multiple one-to-one assignment of votes. In turn, the matching algorithm is extended by ensuring the satisfaction of all the hierarchical constraints encoded in the graphs. The version number "X.Y.Z" of the DAG Matcher X changes only when there is a large reorganization of the code. Y changes when the update implies that you must rebuild the databases created with older versions. Z means that there have been minor changes and that it is safe to use the databases created with any X.Y version. Download DAG Matcher version 5.1 (Tue Jan 24 17:58:09 EST 2006) Source code and the required library (password required, 10MB) Note: access to the source code is granted only to research groups working in collaboration with us. The code compiles well with g++ 3.2, Linux 8.0 and LEDA 4.4, and with g++ 3.0.1, Linux 7.0, and LEDA 4.3.1. Datasets for shape matching experiments Dataset 1: 8 objects, 15 views per object, clean and noisy versions (ppm files, 2MB) 120 images per dataset, 13 datasets. Dataset 2: 21 objects, 128 views per object, clean version only (ppm files, 5MB) 2688 images of 21 objects with a 128 views per object. DAG Matcher dependencies SR-Tree Newmat LEDA (see the LEDA manual) Doxygen