Spectral clustering ensemble for image segmentation

ABSTRACT

To make full use of information included in a dataset, a multiway spectral clustering algorithm with joint model is applied to image segmentation. To overcome the sensitivity of the joint model-based multiway spectral clustering to kernel parameter and to produce the robust and stable segmentation results, spectral clustering ensemble algorithm is proposed in this paper, which can make full use of the built-in randomness of spectral clustering and the inaccuracy of Nystrom approximation to produce diversity. Experiments on UCI dataset, textural and SAR images show that, after cluster ensemble, the new algorithm is not only more robust but also better quality. Therefore, the new algorithm is effective.

REFERENCES

• 1: Jianbo Shi , Jitendra Malik, Normalized Cuts and Image Segmentation, IEEE Transactions on Pattern Analysis and Machine Intelligence, v.22 n.8, p.888-905, August 2000 [doi>10.1109/34.868688]
• 2: Maurizio Filippone , Francesco Camastra , Francesco Masulli , Stefano Rovetta, A survey of kernel and spectral methods for clustering, Pattern Recognition, v.41 n.1, p.176-190, January, 2008 [doi>10.1016/j.patcog.2007.05.018]
• 3: Verma, D. and Meila, M. 2005. A comparison of spectral clustering algorithms. Technical report, Department of CSE University of Washington Seattle, WA 98195--2350.
• 4: Brand, M. and Huang, K. 2003. A unifying theorem for spectral embedding and clustering. In Proceedings of the 9th International Conference on Artificial Intelligence and Statistics, Key West, Florida.
• 5: Meila, M. and Shi, J. B. 2001. A random walks view of spectral segmentation. In Proceedings of Eighth International Conference on Artificial Intelligence and Statistics, Key West, FL, January 4--7.
• 6: Charless Fowlkes , Serge Belongie , Fan Chung , Jitendra Malik, Spectral Grouping Using the Nyström Method, IEEE Transactions on Pattern Analysis and Machine Intelligence, v.26 n.2, p.214-225, January 2004 [doi>10.1109/TPAMI.2004.1262185]
• 7: Luxburg, U. V., Bousquet, O., and Belkin, M. 2005. Limits of spectral clustering. Advances in Neural Information Processing Systems. Vol.17, pp.857--864.
• 8: Data Clustering Using Evidence Accumulation, Proceedings of the 16 th International Conference on Pattern Recognition (ICPR'02) Volume 4, p.40276, August 11-15, 2002
• 9: Ana L. N. Fred, Finding Consistent Clusters in Data Partitions, Proceedings of the Second International Workshop on Multiple Classifier Systems, p.309-318, July 02-04, 2001
• 10: Alexander Strehl , Joydeep Ghosh, Cluster ensembles --- a knowledge reuse framework for combining multiple partitions, The Journal of Machine Learning Research, 3, p.583-617, 3/1/2003 [doi>10.1162/153244303321897735]
• 11: Thomas M. Cover , Joy A. Thomas, Elements of information theory, Wiley-Interscience, New York, NY, 1991
• 12: Haralick, R. M., Shanmugam, K., and Dinstein, I. 1973. Texture features for image classification. IEEE Transactions on Systems, Man and Cybernet. pp.610--621.
• 13: Hu, Z. L., Guo, D. Z., and Sheng, Y. H. 2001. Extracting textural information of satellite SAR image based on wavelet decomposition. Journal of Remote Sensing. Vol.5, No.6, pp.423--427.
• 14: Chongjun Wang , Wu jun Li , Lin Ding , Juan Tian , Shifu Chen, Image Segmentation Using Spectral Clustering, Proceedings of the 17th IEEE International Conference on Tools with Artificial Intelligence, p.677-678, November 14-16, 2005

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.4 IMAGE PROCESSING AND COMPUTER VISION
      I.4.6 Segmentation
          Subjects: Region growing, partitioning

General Terms:
Algorithms, Experimentation

Keywords:
cluster ensemble, image segmentation., nystrom approximation, spectral clustering