Complexity reduction for "large image" processing

N.R. Pal1, J.C. Bezdek2
1Electronics and Communication Sciences Unit, Indian Statistical Institute, Calcutta, India
2Department of Computer Science, University of West Florida, Pensacola, FL, USA

Tóm tắt

We present a method for sampling feature vectors in large (e.g., 2000 /spl times/ 5000 /spl times/ 16 bit) images that finds subsets of pixel locations which represent c "regions" in the image. Samples are accepted by the chi-square (/spl chi//sup 2/) or divergence hypothesis test. A framework that captures the idea of efficient extension of image processing algorithms from the samples to the rest of the population is given. Computationally expensive (in time and/or space) image operators (e.g., neural networks (NNs) or clustering models) are trained on the sample, and then extended noniteratively to the rest of the population. We illustrate the general method using fuzzy c-means (FCM) clustering to segment Indian satellite images. On average, the new method can achieve about 99% accuracy (relative to running the literal algorithm) using roughly 24% of the image for training. This amounts to an average savings of 76% in CPU time. We also compare our method to its closest relative in the group of schemes used to accelerate FCM: our method averages a speedup of about 4.2, whereas the multistage random sampling approach achieves an average acceleration of 1.63.

Từ khóa

#Image processing #Image sampling #Clustering algorithms #Acceleration #Pixel #Testing #Computer networks #Neural networks #Image segmentation #Satellites

Tài liệu tham khảo

10.1007/978-3-642-88087-2 1993, Interactive Data Language (IDL) Ver 3 1 Users Guide de la paz, 1986, approximate fuzzy <formula><tex>$c$</tex></formula>-means (afcm) cluster analysis of medical magnetic resonance image (mri) data&mdash;a system for medical research and education, IEEE Transactions on Geoscience and Remote Sensing, ge 25, 815 morrison, 1994, an introduction to segmentation of magnetic resonance images, Aust Comput J, 26, 90 10.1109/TPAMI.1986.4767778 10.1109/TGRS.1986.289598 10.1016/0031-3203(94)90118-X 10.1109/NAFIPS.2001.944766 uma shankar, 1994, ffcm: an effective approach for large data sets, Proc 3rd Int Conf Fuzzy Logic Neural Nets and Soft Computing IIZUKA, 331 cheng, 1995, fast clustering with application to fuzzy rule generation, Proceedings of the IEEE International Conference on Fuzzy Systems, 2289 10.1145/223784.223812 10.1016/S0306-4379(01)00008-4 bezdek, 2002, Some Notes on Alternating Optimization in Advances in Soft Computing AFSS, 289 domingos, 2001, a general method for scaling up machine learning algorithms and its application to clustering, Proc 18th Int Conf Machine Learning, 106 10.1080/01621459.1963.10500830 haralick, 1992, Computer and Robot Vision, 1 gonzalez, 1992, Digital Image Processing kendall, 1979, The Advanced Theory of Statistics Vol 2 Inference and Relationship gibbons, 1992, Nonparametric Statistical Inference cochran, 1977, Sampling Techniques kullback, 1959, Information Theory and Statistics 10.1007/978-1-4757-0450-1 10.1111/j.1365-2818.1990.tb02958.x chen, 2000, intelligent methodology for sensing, modeling and control of pulsed gtaw: part ii&mdash;butt joint welding, Weld Res Supplement, 164 10.1109/72.159057 zhang, 1995, birch: an efficient data clustering method for very large databases, Proc ACM SIGMOD, 103 fayyad, 1995, from massive data sets to science catalogs: applications and challenges, Proc Workshop Massive Data Sets 10.1002/jmri.1880050520 10.1109/2.781633 bradley, 1998, scaling clustering algorithms to large databases, Proc 4th Int Conf Knowledge Discovery and Data Mining, 9 chen, 2000, intelligent methodology for sensing, modeling, and control of pulsed gtaw: part i&mdash;bead-on-plate welding, Weld Res Supplement, 151 10.1109/ICDE.1999.754966 bowyer, 1996, the digital database for screening mammography, Proc 3rd Int Workshop Digital Mammography 58 fisher, 1954, Statistical Methods for Research Workers 10.1109/91.660808 rosenfeld, 1982, Digital Picture Processing 10.1007/b106267 hathaway, 0, convergence of alternating optimization, Comput Optim Appl 10.1109/FUZZY.1994.343855 jain, 1996, image segmentation by clustering, Advances in Image Understanding, 65 10.1109/91.917113