A Graph Lattice Approach to Maintaining and Learning Dense Collections of Subgraphs as Image Features
1. The graph lattice contains a threshold number of Accepted graph lattice nodes at a given level.
2. The graph lattice contains a threshold number of Accepted graph lattice nodes in total.
3. The list of Candidate graph lattice nodes is exhausted.
4. Quality measures for Candidate nodes falls below a threshold.
The author is with the Intelligent Systems Laboratory, Palo Alto Research Center, Palo Alto, CA 94304.
Manuscript received 27 Feb. 2012; revised 4 Sept. 2012; accepted 1 Dec. 2012; published online 19 Dec. 2012.
Recommended for acceptance by M. Brown.
For information on obtaining reprints of this article, please send e-mail to: email@example.com, and reference IEEECS Log Number TPAMI-2012-02-0148.
Digital Object Identifier no. 10.1109/TPAMI.2012.267.
1. Known in the trade as doctype classification.
2. Single core, 2.4 GHz PC, Java.
Eric Saund received the BS degree in Engineering and Applied Science from the California Institute of Technology and the PhD degree in Cognitive Science from the Massachusetts Institute of Technology. He is with the Intelligent Systems Laboratory, Palo Alto Research Center. His research in computational vision specializes in perceptual organization, with applications in document image analysis, perceptually supported image editing, and video analysis. He was honored to serve as an associate editor of the IEEE Transactions on Pattern Analysis and Machine Intelligence from 2006 to 2011. He is a member of the IEEE and the IEEE Computer Society.