| Authors |
Mauthner Thomas, Kluckner Stefan , Roth Peter M., Bischof Horst |
| Appeared in |
Proceedings Annual Symposium German Association for Pattern Recognition |
| Date |
2010 |
| Abstract |
Recently descriptors based on Histograms of Oriented Gra-
dients (HOG) and Local Binary Patterns (LBP) have shown excellent
results in object detection considering the precision as well as the recall.
However, since these descriptors are based on high dimensional repre-
sentations such approaches su er from enormous memory and runtime
requirements. The goal of this paper is to overcome these problems by in-
troducing hierarchies of orthogonal Non-negative Matrix Factorizations
(NMF). In fact, in this way a lower dimensional feature representation
can be obtained without loosing the discriminative power of the orig-
inal features. Moreover, the hierarchical structure allows to represent
parts of patches on di erent scales allowing for a more robust classi ca-
tion. We show the e ectiveness of our approach for two publicly available
datasets and compare it to existing state-of-the-art methods. In addition,
we demonstrate it in context of aerial imagery, where high dimensional
images have to be processed requiring efficient methods. |
| Link |
PDF
|
