
Invariant Pattern Recognition
The general invariant pattern recognition problem is to construct a system which takes as input an element f of V (set of possible patterns) and computes a value s(f), with the intention that s(f)=c(f) for all f
V • c(f) is the desired classification off, invariant under the action of all transformations (translations, rotations, scalings, etc.).
Digital Hartley Transform and Experimental Results
| Table 1. Two Hartley feature vectors of a defect. | ||
| Rotation (20°) | Rot. (20 °) + Scaling (50%) | |
| H[2] | 0.254127 | 0.260218 |
| H[3] | 0.111315 | 0.114237 |
| H[4] | 0.061237 | 0.062647 |
| H[5] | 0.042525 | 0.044109 |
| H[6] | 0.028964 | 0.031119 |
| H[7] | 0.021795 | 0.023398 |
| H[8] | 10.016215 | 0.016770 |
Conclusion
An invariant pattern recognition method, based on the Hartley transform, applied toradiographic images is presented. The results show that this method is capable to describe anyweld flaw into a small feature vector, allowing its recognition by the inspection system.
Abstract Source:
Book of Abstracts, 7th European Conference on Non-Destructive Testing, 26-29 May 1998, ISBN: 87-986898-0-00
Full-Text Source:
Proceedings of the 7th European Conference on Non-Destructive Testing, 26-29 May 1998, ISBN:
Publication Contact:
7th ECNDT, Park Alle 345, DK-2605 Broendby, Denmark, Fax: +45 46 26 70 11, Email: 101373.3414@compuserve.com
| NDTnet |