TY  - JOUR
AU  - Noha Elfiky
AU  - Jordi Gonzalez
AU  - Xavier Roca
PY  - 2012//
TI  - Compact and Adaptive Spatial Pyramids for Scene Recognition
T2  - IMAVIS
JO  - Image and Vision Computing
SP  - 492–500
VL  - 30
IS  - 8
N2  - Most successful approaches on scenerecognition tend to efficiently combine global image features with spatial local appearance and shape cues. On the other hand, less attention has been devoted for studying spatial texture features within scenes. Our method is based on the insight that scenes can be seen as a composition of micro-texture patterns. This paper analyzes the role of texture along with its spatial layout for scenerecognition. However, one main drawback of the resulting spatial representation is its huge dimensionality. Hence, we propose a technique that addresses this problem by presenting a compactSpatialPyramid (SP) representation. The basis of our compact representation, namely, CompactAdaptiveSpatialPyramid (CASP) consists of a two-stages compression strategy. This strategy is based on the Agglomerative Information Bottleneck (AIB) theory for (i) compressing the least informative SP features, and, (ii) automatically learning the most appropriate shape for each category. Our method exceeds the state-of-the-art results on several challenging scenerecognition data sets.
L1  - http://refbase.cvc.uab.es/files/EGR2012.pdf
UR  - http://dx.doi.org/10.1016/j.imavis.2012.04.002
N1  - ISE
ID  - Noha Elfiky2012
ER  -