Optimal reconstruction of a surface using a reference library
To reconstruct (approximate) an arbitrary surface using subsurfaces (patches) from a library of surfaces in an optimal way is an interesting algorithmic problem and has many applications in image processing. This paper presents an efficient algorithm for an optimal reconstruction of a query surface using patches from a reference library of surfaces, under the constraint that the smallest patch size is above some specified value. In this algorithm, a surface is given as an integer function f(x, y) over a finite 2-D grid. The algorithm partitions a query surface into patches in such a way that each patch is represented by a similar patch from a library surface, and the total difference between the query surface and the representing (composite) surface is minimized, where the boundary of a patch is not pre-determined but solely determined by the optimization process. By using a minimum spanning tree-based data structure, this optimization problem can be solved efficiently. An application of this technique in computational forensics is outlined.
- Research Organization:
- Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
- Sponsoring Organization:
- USDOE Office of Energy Research, Washington, DC (United States)
- DOE Contract Number:
- AC05-96OR22464
- OSTI ID:
- 522739
- Report Number(s):
- CONF-9706151-1; ON: DE97008473; TRN: 97:005032
- Resource Relation:
- Conference: International conference on imaging science, systems, and technology, Las Vegas, NV (United States), 30 Jun - 2 Jul 1997; Other Information: PBD: 1997
- Country of Publication:
- United States
- Language:
- English
Similar Records
A Parallel Computing Framework for Finding the Supported Solutions to a Biobjective Network Optimization Problem
AGGREGATE: dAta-driven modelinG preservinG contRollable dEr for outaGe mAnagemenT and rEsiliency (Final Report)