13. M. Marszalek, C. Schmid, J. Zhang, S. Lazebnik: Bag-of-features image representation: state-of-the-art and beyond

We present our bag-of-features based image classification framework and discuss the selection of different building blocks [1]. We then examine the influence of background clutter on the bag-of-features representation and present a spatial extension to this representation that reduces its sensitivity to background clutter [2]. Finally, we will show some of our recent work which extends and generalizes the framework to perform simultaneous object localization and segmentation.

[1] J. Zhang, M. Marszalek, S. Lazebnik, C. Schmid. Local Features and Kernels for Classification of Texture and Object Categories: A Comprehensive Study.
International Journal of Computer Vision. To appear.
[2] M. Marszalek, C. Schmid. Spatial Weighting for Bag-of-Features.
IEEE Conference on Computer Vision & Pattern Recognition, 2006.