Automated human identification from their walking behavior is a challenge attracting much interest among machine vision researchers. However, the systems which are able to detect pedestrian attributes based on their walking behavior remain to be developed. Here, a soft computing approach to determine walking behavior based on motion imagery is studied as the basis for developing pedestrian safety information systems. Gender and age are classified based on motion pattern derived in experiments. At the front end, image and video processing was performed to separate foreground from background images. The widths of silhouette were analyzed using two-dimensional (2D) Fourier transformation to extract human motion features. Feature sub-sets were then selected to find salient, effective classification features. Finally, Choquet integral agent networks (CHIAN) with a competitive learning algorithm were employed to classify gender and age into its classes. The experimental results demonstrated capability of the proposed system to classify gender and age in highly accurately.

1. [1] Pedestrian and Bicycle Information Center,
   http://www.walkinginfo.org.
2. [2] T. Hummel, “Dutch Pedestrian Safety Research Review,” tech. report FHWA-RD-99-092, SWOV Institute for Road Safety Research, Dec 1999.
3. [3] I. Banerjee, S.E. Shladover, J. A. Misener, C. Y. Chan, and D. R. Ragland, “Impact of Pedestrian Presence on Movement of Left-Turning Vehicles: Method, Preliminary Research & Possible Use in Intersection Decision Support,” University of California, Berkeley, 2004.
4. [4] A. K. Bechtel, J. Geyer, and D. R. Ragland, “A Review of ITS-based Pedestrian Injury Countermeasures,” University of California, Berkeley, 2003.
5. [5] S. Kamijo, Y. Matsushita, K. Ikeuchi, and M. Sakauchi, “Traffic Monitoring and Accident Detection at Intersections,” IEEE Transactions on Intelligent Transportation Systems, Vol.1(2), pp. 108-118, June 2000.
6. [6] C. Stauffer , W. E. L. Grimson, “Learning patterns of activity using real-time tracking,” IEEE Transaction on Pattern Analysis & Machine Intelligence, Vol.22(8), pp. 747-757, 2000.
7. [7] A. Prati, I. Mikic, C. Grana, and M. M. Trivedi, “Shadow Detection Algorithms for Traffic Flow Analysis: a Comparative Study,” Submitted to IEEE Int. Conf. On Intelligent Transportation Systems, Oakland, California, Aug 2001.
8. [8] A. Kale, N. Cuntoor, A. N. Rajagopalan, B. Yegnanarayana, and R. Chellappa, “Gait analysis for human identification,” Proc. of 3rd Int. Conf. on Audio and Video Based Person Authentication, June 2003.
9. [9] Wavelet Tutorial Part II by R. Polikar,
   http://users.rowan.edu/ 〜 polikar/WAVELETS/WTpart2.html.
10. [10] P. Pavel and J. Novovicova, “Novel Methods for feature subset selection with respect to problem domain. In Feature Extraction, Construction and Selection : A Data Mining Perspective,” Kluwer Academic Publisher, 2^nd edition, 2001.
11. [11] S. Theodoridis and K. Koutroumbas, “Pattern Recognition,” Academic Press., mathrm^rd Edition, 2005.
12. [12] K. Nakamura, “A Scheme for information fusion by Choquet Integral Agent Networks,” Eighth IFSA Congress, pp. 954-958, 1999.
13. [13] G. C. Cawley, and N. L. C. Talbot, “Efficient leave-one-out cross-validation of kernel fisher discriminant classifiers,” Pattern Recognition, Vol.36(11), pp. 2585-2592, November 2003.