Shadow edge detection using geometric and photometric features

The detection of shadow and shading edges is a first step towards reducing the imaging effects that are caused by interactions of the light source with surfaces that are in the scene. As most of the algorithms for shadow edge detection use photometric information, geometric information have been ignored so far. In this paper, the aim is to include geometric features for more robust shadow edge detection. First, thousands of patches are annotated as either containing a shadow edge or not. Then, geometric features of these patches are analyzed and it is shown that the combination of photometric and geometric features improves the classification of shadow edges with respect to using either one of these features with 14%. These results demonstrate the added value of geometric features, in addition to photometric features, for the detection of shadow edges.

[1]  Cheng Lu,et al.  On the removal of shadows from images , 2006, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[2]  G D Finlayson,et al.  Color constancy at a pixel. , 2001, Journal of the Optical Society of America. A, Optics, image science, and vision.

[3]  Arnold W. M. Smeulders,et al.  Color Invariance , 2001, IEEE Trans. Pattern Anal. Mach. Intell..

[4]  Antonio Torralba,et al.  Modeling the Shape of the Scene: A Holistic Representation of the Spatial Envelope , 2001, International Journal of Computer Vision.

[5]  Matti Pietikäinen,et al.  Multiresolution Gray-Scale and Rotation Invariant Texture Classification with Local Binary Patterns , 2002, IEEE Trans. Pattern Anal. Mach. Intell..

[6]  Robert M. Haralick,et al.  Textural Features for Image Classification , 1973, IEEE Trans. Syst. Man Cybern..

[7]  Jiri Matas,et al.  On Combining Classifiers , 1998, IEEE Trans. Pattern Anal. Mach. Intell..

[8]  Joost van de Weijer,et al.  Edge Classification for Color Constancy , 2008, CGIV/MCS.

[9]  Erik Reinhard,et al.  Evaluation of color spaces for edge classification in outdoor scenes , 2005, IEEE International Conference on Image Processing 2005.

[10]  Cordelia Schmid,et al.  Beyond Bags of Features: Spatial Pyramid Matching for Recognizing Natural Scene Categories , 2006, 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06).

[11]  Joost van de Weijer,et al.  Edge and corner detection by photometric quasi-invariants , 2005, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[12]  Theo Gevers,et al.  Classifying color edges in video into shadow-geometry, highlight, or material transitions , 2003, IEEE Trans. Multim..

[13]  Matthijs C. Dorst.  Distinctive Image Features from Scale-Invariant Keypoints , 2011 .