Optimisation of shape kernel and threshold in image-processing motion analysers
- 65 Downloads
The aim of the work is to optimise the image processing of a motion analyser. This is to improve accuracy, which is crucial for neurophysiological and rehabilitation applications. A new motion analyser, ELITE-S2, for installation on the International Space Station is described, with the focus on image processing. Important improvements are expected in the hardware of ELITE-S2 compared with ELITE and previous versions (ELITE-S and Kinelite). The core algorithm for marker recognition was based on the current ELITE version, using the cross-correlation technique. This technique was based on the matching of the expected marker shape, the so-called kernel, with image features. Optimisation of the kernel parameters was achieved using a genetic algorithm, taking into account noise rejection and accuracy. Optimisation was achieved by performing tests on six highly precise grids (with marker diameters ranging from 1.5 to 4 mm), representing all allowed marker image sizes, and on a noise image. The results of comparing the optimised kernels and the current ELITE version showed a great improvement in marker recognition accuracy, while noise rejection characteristics were preserved. An average increase in marker co-ordinate accuracy of +22% was achieved, corresponding to a mean accuracy of 0.11 pixel in comparison with 0.14 pixel, measured over all grids. An improvement of +37%, corresponding to an improvement from 0.22 pixel to 0.14 pixel, was observed over the grid with the biggest markers.
KeywordsMotion analysis Genetic algorithms Correlation International Space Station Rehabilitation instrumentation Constrained optimisation over finite domain
Unable to display preview. Download preview PDF.
- Beasley, D., Bull, D. R., andMartin, R. R. (1993): ‘An overview of genetic algorithms’,Univ. Comput. Google Scholar
- Ferrigno, G., andPedotti, A. (1985): ‘ELITE: a digital dedicated hardware system for movement analysis via real time TV signal processing’,IEEE Trans. Biomed. Eng.,32, pp. 943–950Google Scholar
- Ferrigno, G., Borghese, N. A., andPedotti, A. (1990): ‘Pattern recognition in 3D automatic human motion analysis’,ISPRS J. Photogrammetry Remote Sensing,45, pp. 227–246Google Scholar
- Homaifar, A., Lai, S. H., andQi, H. (1994): ‘Constrained optimisation by genetic algorithms’,Simulations,62, pp. 242–254Google Scholar
- Joines, J. A., Culbreth, C. T. andKing, R. E. (1993): ‘Manufacturing cell design: an integer programming model employing genetic algorithms’. Technical report NCSU-IE Technical Report 93-16, North Carolina State University, (revised 1996)Google Scholar
- Joines, J. A., andHouck, C. (1994): ‘On the use of non-stationary penalty functions to solve non-linear constrained optimization problems with gas’. Proc. 1st IEEE Conf. Evol. Comput., New York, pp. 579–584Google Scholar
- Michalewicz, Z., andNazhiyath, G. (1995): ‘Genocop III. A coevolutionary algorithm for numerical optimisation problems with non-linear constraints’. Proc. 2nd IEEE Conf. Evol. Comput., New York, pp. 647–651Google Scholar
- Michalewicz, Z. (1996): ‘Genetic algorithms + Data structures = Evolution programs’, (Springer, Berlin, Germany), 3rd ednGoogle Scholar
- Pedrocchi, A., Baroni, G., Ferrigno, G., andPedotti, A. (1998): ‘ELITE-S2: potential application and technological requirements for multifactorial movement analysis in microgravity’. Second Eur. Symp. Utilization of the International Space Station, Estec, pp. 549–553Google Scholar
- Riuz-Andino, A., Araujo, L., Saenz, F., andRuz, J. (2000): ‘A hybrid evolutionary approach for solving constrained optimisation problems over finite domains’,IEEE Trans. Evol. Comput.,4, pp. 353–372Google Scholar
- Schnecke, V., andVornberger, O. (1997): ‘Hybrid genetic algorithms for constrained placement problems’,IEEE Trans. Evol. Comput.,1, pp. 266–277Google Scholar