A novel algorithm for a rotation invariant template matching

Abstract

A novel algorithm for a rotation invariant template matching is proposed when the fluctuating scope of the rotation angle is limited within the region of [−20°,20°]. The matching candidates are selected using a computationally low cost improved correlation algorithm. “AND” operation is adopted to reduce the computational cost. Therefore the algorithm improves the matching speed consumedly. The simulation results verify the efficiency of the proposed method. Moreover, when the size of reference image is fixed, the advantage of this time-saving algorithm is more obvious as the increase of the size of the real time image. The matching speed of the proposed method is over 20 times faster than the speed of the two-level pyramid decomposing accelerating method.

References

[1] BakerE. S., DegroatR. D.. IEEE Trans on Signal Proceesing, 1998, 46: 3112

[2] BrownL. G.. ACM Computing Surveys, 1992, 24: 325

[3] Cahnvon Seelen U M, and Bajcsy R. University of Pennsylvania: GRASP Laboratory Technical Report, 1996.

[4] CavesR.G., HarleyP.J., QueganS.. IEEE Transactions on geoscience and remote sensing, 1992, 30: 680

[5] WongR. Y., HallE. L.. IEEE Trans on Computer, 1978, 27: 359

[6] RosenfeldA., KakA.. Digital Image Processing, 1982, 2nd EditionOrlando, Academic Press

[7] ProkopR.J., ReevesA.P.. CVGIP Graph. Models Image Process, 1992, 54: 438

[8] LeeS.-W., KimW.-Y.. Proc KITE, 1996, 19: 475

[9] Di StefanoL.. Maschine Vision and Applications, 2003, 13: 213

[10] KimS.-G., KimW.-Y.. Proceedings KSPC, 1996, 96: 1335

[11] ChoiM.-S., KimW.-Y.. Pattern Recognition, 2002, 35: 119