The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B3b. Beijing 2008 
Figure 8: Histograms over the features of each weak classi 
fier h t : /9, /10, /12, /13, /14, /15, fi6, fis, /113, /121 and 
/129 are the only features that are involved in classifying 
roof regions at scale a = 1. 
Table 6: Error rates on automatically segmented regions. 
scale 
window 
window pane 
all 
21.0% 
51.4% 
1 
18.2% 
13.1% 
2 
19.9% 
19.0% 
4 
86.4% 
31.3% 
8 
38.6% 
7.5% 
027113 which is funded by The European Union. 
The authors would also like to thank Marko Pilger for im 
plementing the functions for feature extraction. 
References 
Bangham, J. A., Moravec, K., Harvey, R. and Fisher, M., 
1999. Scale-space trees and applications as filters for 
stereo vision and image retrieval. In: BMVC, pp. 1 IS 
MS. 
Bishop, C. M., 2006. Pattern Recognition and Machine 
Learning. Springer. 
Breiman, L., 2001. Random forests. Machine Learning 45, 
pp. 5-32. 
Douglas, D. H. and Peucker, T. K., 1973. Algorithms for 
the reduction of the number of points required to repre 
sent a digitized line or its caricature. Canadian Cartog 
rapher 10(2), pp. 112-122. 
Guyon, I. and Elisseeff, A., 2003. An introduction to vari 
able and feature selection. Journal of Machine Learning 
Research 3, pp. 1157-1182. 
Hartz, J. and Neumann, B., 2007. Learning a knowledge 
base of ontological concepts for high-level scene inter 
pretation. In: ICMLA. 
Ho, T. K., 1998. The random subspace method for con 
structing decision forests. PAMI 20(8), pp. 832-844. 
Kumar, S. and Hebert, M., 2003. Man-made structure de 
tection in natural images using a causal multiscale ran 
dom field. In: CVPR. 
Lifschitz, I., 2005. Image interpretation using bottom-up 
top-down cycle on fragment trees. Master’s thesis, Weiz- 
mann Institute of Science, Rehovot, Israel. 
Liu, H. and Motoda, H., 1998. Feature Selection for 
Knowledge Discovery and Data Mining. Kluwer Aca 
demic. 
Mayer, H., 1999. Automatic object extraction from aerial 
imagery - a survey focusing on buildings. CVIU 74, 
pp. 138-149. 
McGlone, J. C. (ed.), 2004. Manual of Photogrammetry. 
5th edition edn, ASPRS. 
Nevada, R., Lin, C. and Huertas, A., 1997. A system for 
building detection from aerial images. In: USC Com 
puter Vision. 
Rätsch, G., Onoda, T. and Müller, K.-R., 2001. Soft mar 
gins for adaboost. Machine Learning 43(3), pp. 287- 
320. 
Reznik, S. and Mayer, H., 2007. Implicit shape models, 
model selection, and plane sweeping for 3d facade inter 
pretation. In: PIA07, IAPRS 36 (3AV49A), pp. 173-178. 
Schapire, R. E., 1990. The strength of weak leamability. 
Machine Learning 5(2), pp. 197-227. 
Schapire, R. E. and Singer, Y., 1999. Improved boosting 
algorithms using confidence-rated predictions. Machine 
Learning 37(3), pp. 297-336. 
Sivic, J., Russell, B. C., Efros, A. A., Zisserman, A. and 
Freeman, W. T., 2005. Discovering objects and their lo 
cation in images. In: Proc. 10th ICCV, Beijing, China. 
Tuytelaars, T., Turina, A. and Van Gol, L., 2003. Noncom- 
binatorial detection of regular repetitions under perspec 
tive skew. PAMI 25(4), pp. 418-432. 
Verbeek, J. and Triggs, B., 2007. Region classification with 
markov field aspect models. In: VCVPR. 
Wang, X., Totaro, S., Taillandier, F., Hanson, A. R. and 
Teller, S., 2002. Recovering facade texture and mi 
crostructure from real-world images. In: Proc. 2nd IW- 
TAS, pp. 145-149.