An Efficient Exact Algorithm for the 'Least Squares Image Registration Problem
Abstract
Image registration involves estimating how one set of n-dimensional points is rotated, scaled, and translated into a second set of n-dimensional points. In practice, n is usually 2 or 3. We give an exact algorithm to solve the least-squares formulation of the two-dimensional registration problem. The algorithm, which is based on parametric linear programming, can be viewed as a refinement of the O(k cubed) approximation method proposed by Zikan and Silberberg. Keywords: Parametric linear programming, Frobenius norm.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1989
- Accession Number
- ADA208725
Entities
People
- Karel Zikan
Organizations
- Stanford University