7A-5-8 
7 Conclusions 
(a) (b) (c) 
Figure 14: (a) Original imager(b) through a text-area 
filterr(c) text region located 
(a) (b) (c) 
Figure 15: (a) Original imager(b) through a text-area 
filterr(c) text region located 
window while a traffic sign tracked in a sequence of 
images is becoming bigger. 
The other method is to adaptively determine the 
size of the local window. Vertical edges of a text region 
can be considered as an uniform distribution. Statis 
tical variables can be used to find a suitable window 
size. For instancerassume the window size w varies 
from 15x1 to 50x1 ran accumulator cell A(w) is at 
tached to each window size w. Scan the whole imager 
if the distribution of the vertical edges in the local 
window of the size w is uniformly distributedT vote 
the cell A(w). Large accumulated values in the cells 
mean there exist potential text lines with a certain 
font size corresponding to the window size w in the 
image. 
(a) (b) (c) (d) 
Figure 16: (a) Original imager(b) through a text-area 
filterr(c) text region locatedr(d) characters located 
In this paperFwe describe automatic traffic sign de 
tection. The results here cover the first module: sign 
detection. Future work will integrate the remaining 
three modules: sign recognitionTsign localization and 
sign prediction shown in Figure 2. 
Acknowledgments : Thanks go to Dr. Alan Selby 
for his assistance in proofreading this paper. 
References 
[1] R. Li. “Mobile Mapping - An Emerging Technology 
for Spatial Data Acquisition”. Journal of Photogram- 
metric Engineering and Remote Sensing, Vol.63, 
No.9, pp.1085-1092. 1997. 
[2] G. Piccioli, E. D. Michelli, and M. Campani. “A ro 
bust method for road sign detection and recognition”. 
In Proc. European Conference on Computer Vision 
1994, pages 495-500, 1994. 
[3] N. Kehtaxnavaz, N. C. Griswold, and D. S. Kang. 
“Stop-sign recognition based on color-shape process 
ing”. Machine Vision and Application, Vol.6, Pages 
206-208, 1993. 
[4] L. Priese, J. Klieber, R. Lakmann, V. Rehrmann, and 
R. Schian. “New results on traffic sign recognition”. 
In Symposium on intelligent vehicles. IEEE, 1994. 
[5] R. Janssen, W. Ritter, F. Stein, and S. Ott. “Hybrid 
approach for traffic sign recognition”. In Symposium 
on intelligent vehicles. IEEE, July 1993. 
[6] W. Sjarbek abd A. Koschan. Color Image segmenta 
tion. Technical report, Institute for Technical Infor 
matics, Technical University of Berlin, October 1994. 
[7] S. Z. Li. Markov Random Field Modeling in Com 
puter Vision. Springer-Verlag, 1995 
[8] R. Gonzalez and P. Wintz. Digital Image Process 
ing. 2nd edition, Addison-Wesley Publishing Com 
pany, 1987. 
[9] J. Besag. “Spatial interaction and the statistical anal 
ysis of lattice systems”. Journal of the Royal Statis 
tical Society, Series B, 36:192-236, 1974. 
[10] P. Cohen, W. B. Tong and J. Y. Herve. “Automated 
traffic sign recognition”, submitted to 32nd Inter 
national Symposium on Automotive Technology and 
Automation (ISATA), June 14-18, 1999, Vienna, Aus 
tria.