The Hough Transform without the Accumulators
The least-squares method (LSM) efficiently solves the model-fitting problem, if we assume a model equation. For the fitting to a collection of models, the classification of data is required as pre-processing. The Hough transform, achieves both the classification of sample points and the model fitting concurrently. However, as far as adopting the voting process is concerned, the maintenance of the accumulator during the computation cannot be neglected. We propose a Hough transform without the accumulator expressing the classification of data for the model fitting problems as the permutation of matrices which are defined by data.
KeywordsSample Point Orthogonal Matrix Projection Matrix Hough Transform Moment Matrix
- 1.Ballard, D. and Brown, Ch. M., Computer Vision, Prentice-Hall; New Jersey, 1982.Google Scholar
- 4.Becker, J.-M., Grousson, S., and Guieu, D., Space of circles: its application in image processing, Vision Geometry IX, Proceedings of SPIE, 4117, 243–250, 2000.Google Scholar
- 6.Mattavelli, M., Noel, V., and Ammaldi, E., Fast line detection algorithms based on combinatorial optimization, LNCS, 2051, 410–419, 2001.Google Scholar