Gradient descent approach for minimizing dissimilarity measure in log-polarimagery

Abstract

Log-polar mapping has been proposed as a very appropriate space-variant imaging model in active vision applications. There is no doubt about the importance of translation estimation in active visual tracking. In this paper an approach is presented, and its performances are evaluated. The approach uses a gradient descent for minimizing a dissimilarity measure. The experimental results reveal that this method is efficient for approaching active image translations.

References

[1] SchwartzE. L.. Biological Cybernetics, 1977, 25: 181-181

[2] OshiroN., MaruN., NishikawaA., MiyazakiF.. IEEE/RSJ International Conference on Intelligent Robots and Systems, Osaka Japan, 1996, 2: 791-791

[3] BernardinoA., Santos-VictorJ.. Robotics and Autonomous Systems, 1998, 25: 137-137

[4] C. F. Weiman and R. D. Juday,SPIE, 1989, 843.

[5] H. Tunley and D. Young,Dynamic Fixation of a Moving Surface using Log Polar Sampling, British Machine Vision Conference, vol. 2, BMVA Press, 1994, 579.

[6] F. Panerai, C. Capurro, and G. Sandini,Space variant vision for an active camera mount, Technical Report TR 1/95,LIRA, DIST, University of Genova, Italy, February 1995.

[7] TraverV. J., PlaF.. Spanish Symposium on Pattern Recognition and Image Analysis (SNRFAI), Castello'n, Spain, 2001, 2: 281-281

[8] I. Ahrns and H. Neumann,Proceedings of Artificial Intelligence, 1998, 89.

[9] OshiroN., MaruN., NishikawaA., MiyazakiF.. IEEE/RSJ International Conference on Intelligent Robots and Systems, Osaka Japan, 1996, 2: 791-791