Talks and Poster Presentations (without Proceedings-Entry):
A Chwatal, G. Raidl:
"Applying branch-and-cut for compressing fingerprint templates";
Talk: EURO XXII: 22nd European Conference on Operational Research,
Prag, Tschechische Republik;
We consider the problem of encoding fingerprint minutiae data in a highly compact way. An application is the embedding as watermarks in images of ID-cards. Our new approach is based on connecting a subset of the given data points by an arborescence whose arcs can be efficiently encoded by indices into a small codebook. For achieving highest possible compression, the connected points, the arborescence, and the codebook are simultaneously optimized. We transform this problem into an extended variant of the minimum label spanning tree problem and solve an ILP formulation by branch-and-cut.
Created from the Publication Database of the Vienna University of Technology.