What is Superpixels?
The term ‘superpixel’ was introduced by Ren and Malik [1]. A superpixel is an image patch which is better aligned with intensity edges than a rectangular patch. Superpixels are perceptually consistent units which carry more information than pixels and adhere well to image boundaries.
Desirable Properties of Superpixels
Perceptual meaningfulness - Superpixel algorithms aggregate pixels together to form atomic regions that have a certain meaningful perception. Superpixels are the natural representation of an image and carry more perceptual and semantic meaning than pixels.
Computational efficiency and simple to use - Superpixel segmentation showed to be a useful preprocessing step in many computer vision applications.
Algorithms for generating superpixels can be broadly categorized as either graph-based or gradient-ascent methods.
The normalized cuts algorithm [2], efficient graph-based image segmentation algorithm [3] and superpixel lattices [4] are graph-based superpixel algorithms.
The watershed approach [5], The mean-shift [6], quick-shift [7] and simple linear Iterative clustering (SLIC) [8] are gradient-ascent based superpixel algorithms.
SLIC is shown to yield state-of-the-art adherence to image boundaries on the Berkeley benchmark dataset and outperforms existing methods. Furthermore, it is faster and more memory efficient than previous methods. The algorithm has a complexity of O(N). In addition to these quantifiable benefits, SLIC is easy to use, offers flexibility in the compactness and number of the superpixels it generates and is straightforward to extend to higher dimensions.
References
[1] - X. Ren and J. Malik. Learning a classification model for segmentation. IEEE ICCV, pp. 10–17, 2003.
[2] - J. Shi and J. Malik. "Normalized cuts and Image Segmentation", In IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), pp. 888 – 905, 2000.
[3] - P.F. Felzenszwalb, and P.D. Huttenlocher. "Efficient Graph- based Image Segmentation", In International Journal of Computer Vision (IJCV), pp. 167–181, 2004.
[4] - A.P. Moore, S. Prince, J. Warrell, U. Mohammed and G. Jones. "Superpixel Lattices", In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1–8, 2008.
[5] - L. Vincent and P. Soille. "Watersheds in Digital Spaces: An efficient algorithm based on immersion simulations", IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), pp. 583–598, 1991.
[6] - D. Comaniciu and P. Meer. "Mean shift: A robust approach toward feature space analysis", In IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), pp. 603–619, 2002.
[7]- A. Vedaldi and S. Soatto. "Quick shift and Kernel Methods for Mode Seeking", In proceedings of the European Conference on Computer Vision, Springer Berlin Heidelberg, pp.705–718,2008.
[8]- R. Achanta, A. Shaji, K. Smith, A. Lucchi, P. Fua and S.Susstrunk. "SLIC Superpixels Compared to State-of-the-art Superpixel Methods", In IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), pp. 2274– 2282, 2012.
State-of-the-art superpixels algorithms
The normalized cuts algorithm [2], efficient graph-based image segmentation algorithm [3] and superpixel lattices [4] are graph-based superpixel algorithms.
The watershed approach [5], The mean-shift [6], quick-shift [7] and simple linear Iterative clustering (SLIC) [8] are gradient-ascent based superpixel algorithms.
SLIC is shown to yield state-of-the-art adherence to image boundaries on the Berkeley benchmark dataset and outperforms existing methods. Furthermore, it is faster and more memory efficient than previous methods. The algorithm has a complexity of O(N). In addition to these quantifiable benefits, SLIC is easy to use, offers flexibility in the compactness and number of the superpixels it generates and is straightforward to extend to higher dimensions.
References
[1] - X. Ren and J. Malik. Learning a classification model for segmentation. IEEE ICCV, pp. 10–17, 2003.
[2] - J. Shi and J. Malik. "Normalized cuts and Image Segmentation", In IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), pp. 888 – 905, 2000.
[3] - P.F. Felzenszwalb, and P.D. Huttenlocher. "Efficient Graph- based Image Segmentation", In International Journal of Computer Vision (IJCV), pp. 167–181, 2004.
[4] - A.P. Moore, S. Prince, J. Warrell, U. Mohammed and G. Jones. "Superpixel Lattices", In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1–8, 2008.
[5] - L. Vincent and P. Soille. "Watersheds in Digital Spaces: An efficient algorithm based on immersion simulations", IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), pp. 583–598, 1991.
[6] - D. Comaniciu and P. Meer. "Mean shift: A robust approach toward feature space analysis", In IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), pp. 603–619, 2002.
[7]- A. Vedaldi and S. Soatto. "Quick shift and Kernel Methods for Mode Seeking", In proceedings of the European Conference on Computer Vision, Springer Berlin Heidelberg, pp.705–718,2008.
[8]- R. Achanta, A. Shaji, K. Smith, A. Lucchi, P. Fua and S.Susstrunk. "SLIC Superpixels Compared to State-of-the-art Superpixel Methods", In IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), pp. 2274– 2282, 2012.
No comments:
Post a Comment