97 
have been found in an example image. Regardless of possible movement patterns a rule-based combination of 
found limb regions already enables the recognition system to yield an alarm expressing that a person might be 
mapped in the analysed image area. This is an important result of a sound scene description which is attainable 
with the techniques explained in this contribution. 
4 Conclusion 
Time-efficient procedures for contour following and edge based segmentation have been presented as the first 
steps towards forming an object hypothesis. The algorithms are tailored to the particular instance of recognizing 
humans because a closed-form solution for this optimization problem is not in sight. Only with these restrictions 
fast analysis with cycle times around 150 ms has been made possible. The results show that a fully automated in- 
itialization of movement models is feasible for this class of complex objects, too. 
5 References 
Akita, Koichiro: Analysis of body motion image sequence. IEEE Conf. Pattern Recognition, pp.320-325, 1982. 
Attwood, C.L; Sullivan, G.D.; Baker, K.D.: Model-based recognition of human posture using single synthetic 
images. Proc. Fifth Alvey Vision Conf. AVC 89, Sheffield, pp.25-30, Sept. 1989. 
Bartolini, F.; Cappellini, V.; Mecocci, A.: Counting people getting in and out of a bus by a real-time 
image-sequence processing. Image and Vision Computing 12(1) 36-41, Jan/Feb 1994. 
Borghese, N.A.; et. al: ELITE - A goal oriented vision system for moving objects detection. Robotica, Vol.9, 
pp.275-282, 1991. 
Chen, Zen; Lee, Hsi-Jian: Knowledge-guided visual perception of 3-D human gait from a single image Sequence. 
IEEE Trans. Systems, Man, and Cybernetics 22(2) 336-42, March/April 1992. : 
Dickmanns, Ernst D.; Mysliwetz, Birger: Recursive 3D road and relative ego-state recognition. IEEE PAMI 14, 
Special Issue on Interpretation of 3D Scenes Part II, pp.199-213, Feb. 92. 
Dolan, John; Riseman, Edward: Computing curvilinear structure by token-based grouping. CVPR, 264-270, 1992. 
Edelsbrunner, Herbert: Algorithms in combinatorial geometry. Springer, Berlin, 1987. 
Fletcher, M.J.; Warwick, K.; Mitchell, R.J.: The application of a hybrid tracking algorithm to motion analysis. Proc. 
IEEE CVPR 91, pp.84-89, 1991. 
Hanson, Allen R.; Riseman, Edward M. (eds.): VISIONS — a computer system for interpreting scenes. In Hanson 
& Riseman (eds.): Computer vision systems. Academic Press, New York, pp.303-334, 1978. 
Hinton, G.E.: Relaxation and its role in vision. Ph.D. dissertation, Univ. of Edinburgh, December 1979. 
Hogg, David: Model-based vision - a program to see a walking person. Image and Vision Computing 1(1) 5-20, Feb. 
1983. 
Kinzel, W.; Dickmanns, E.D.: Moving humans recognition using spatio-temporal models. XVII ISPRS Congress, 
Commission V, 1992. 
Leung, M.K.; Yang, Yee-Hong: A region based approach for human body motion analysis. Pattern Recognition 
20(3) 321-39, 1987. 
Okawa, Yoshikuni; Hanatani, Shingo: Recognition of human body motions by robots. Proc. IEEE/RSJ Int. Conf. 
Intelligent Robots and Systems, Raleigh, NC, pp. 2139-46, July 1992. 
Oscarsson, E.: TV-camera detecting pedestrians for traffic light control. Technological and methodological 
advances measurement (ACTA IMEKO), vol.3, pp.275-82, 1982. : 
Perales, Francisco J.; Villanueva, Juan J.; Luo, Y.: An automatic two-camera human motion perception system 
based on biomechanic model matching. IEEE Proc. Int. Conf. SMC "Decision Aiding for Complex Systems’, 
vol.2 of 3, pp.853-8, Oct 1991. 
Quinlan, J.R.: Induction of decision trees. Machine Learning 1(1) 81-106, 1986. 
Rohr, K; Nagel Hans-Hellmut: Modellgestützte Bestimmung des Bewegungszustandes von FuBgingern in 
Realweltbildfolgen. In R.E. Grofkopf (ed.): Mustererkennung '90, Proc. 12. DAGM-Symposium, pp.52-58, 
1990. 
O'Rourke, J.; Badler, N.L: Model-based image analysis of human motion using constraint propagation. IEEE 
Trans. PAMI 2(6) 522-536, Nov. 1980. 
Segen, Jakub: From features to symbols; learning relational models of shape. In Jean C. Simon (ed.): From pixels 
to features, Elsevier, North-Holland, pp.237-48, 1989. 
IAPRS, Vol. 30, Part 5W1, ISPRS Intercommission Workshop "From Pixels to Sequences”, Zurich, March 22-24 1995