Faber, Petko
by the model seat occupation detection. In fig. 5 k) + 1) the projection of the 3-D point set forming the approximation
of a sucessfully approximated human’s head is sketched. Assuming, that the approximation is correct, we can use the
determined parameter of the ellipsoid to control the airbag deployment. The parameter set of the sketched ellipsoid are:
Fm = 204.56 mm, y,, = 302.25 mm, zn = —579.66 mm, and A =— 116.92 mm, B — 94.24 mm, C z 76.78 mm
corresponding the the head length 2- A, head height 2- P and head breadth 2-C' (cf. the values in section 3.1).
The reliability of the results of the module seat occupation detection is very high. If a seat is occupied by a human (or
any significant object) in reality it is always classified as occupied. The reliability of the results of the module passenger
detection and localization shows some weak points up. Thus e.g. the "plausible" possibility of the approximation of a
shoulder or the head restraint exists as human's head. Likewise the reliability of a reliable detection causes difficulties. In
these cases the justified assumption exists that a sequence processing brings the desired stability.
5 CONCLUSION
The presented results show that the objective, the detection and localization of passengers inside vehicles, can be reached
using the described computational structure. If a seat is occupied by a human it is possible to detect and localize the
human’s head as the most distinguished part of the body.
The achievable results then can give the necessary information wether and how an airbag has to be deployed in the case
of an accident to protect the passengers inside a vehicle against heavy violations.
REFERENCES
The National Highway Traffic Safety Administration, 1999. Airbags. http://www.nhtsa.dot .gov.
The Insurance Institute for Highway Safety, 1996. Airbags. http://www.highwaysafety. org.
Aggarwal, J. K. and Cai, Q., 1999. Human motion analysis: A review. CVIU 73(3), pp. 428-440.
Birchfeld, S., 1998. Elliptical head tracking using intensity gradients and color histograms. In: CVPR.
Bookstein, F. L., 1979. Fitting conic sections to scattered data. Computer Graphics and Image Processing 9, pp. 56-71.
Faber, P., 2000. Seat occupation detection inside vehicles. In: Proc. Southwest Symp. on Image Analysis and Interpreta-
tion.
Fieguth, P. and Terzopoulos, D., 1997. Color-based tracking of heads and other mobile objects at video frame rates. In:
CVPR.
Fischler, M. A. and Bolles, R. C., 1981. Random sample consensus: A paradigm for model fitting with applications to
image analysis and automated cartography. Comm. of the ACM 24(6), pp. 381-395.
Fitzgibbon, A. W., Pilu, M. and B.Fisher, R., 1996. Direct least squares fitting of ellipses. Technical report, Univerity of
Edinburgh.
Fórstner, W. and Gülch, E., 1987. A Fast Operator for Detection and Precise Location of Distrinc Points, Corners and
Centres of Circular Features. In: Intercommision Conference of ISPRS on Fast Processing of Photogrammetric Data.
Gander, W., 1981. Least squares with a quadratic constraint. Numerische Mathematik 36, pp. 291—307.
Gavrila, D. M., 1999, The visual analysis of human movement: A survey. CVIU 73(1), pp. 82-98.
Haralick, R. M. and Shapiro, L. G., 1993. Computer and Robot Vision. Vol. 2, Addison-Wesley.
Krumm, J. and Kirk, G., 1998. Video occupant detection for airbag deployment. In: 4th IEEE Workshop on Applications
of Computer Vision.
Schubert, A. and Dickmanns, E. D., 1998. Real-time gaze observation for tracking human control of attention. In: Face
Recognition: From theory to Applications, Springer, pp. 617—626.
Sobottka, K. and Pitas, L, 1996. Segmentation and tracking of faces in color images. In: 2nd Int. Conf. on Automatic
Face and Gesture Recognition.
Yow, K. C. and Cipolla, R., 1996. Feature-based human face detection. Technical report, University of Cambridge.
236 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B5. Amsterdam 2000.
— OC + + © ee
