Correction of Misclassifications Using a Proximity-Based Estimation Method
- PDF / 1,287,294 Bytes
- 14 Pages / 600 x 792 pts Page_size
- 94 Downloads / 234 Views
Correction of Misclassifications Using a Proximity-Based Estimation Method Antti Niemisto¨ Institute of Signal Processing, Tampere University of Technology, P.O. Box 553, 33101 Tampere, Finland Email: [email protected] Department of Pathology, The University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030, USA
Ilya Shmulevich Department of Pathology, The University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030, USA Email: [email protected]
Vladimir V. Lukin Department 504, National Aerospace University, 17 Chkalova Street, 61070 Kharkov, Ukraine Email: [email protected]
Alexander N. Dolia Department 504, National Aerospace University, 17 Chkalova Street, 61070 Kharkov, Ukraine School of Electronics and Computer Science, University of Southampton, Southampton SO17 1BJ, England, UK Email: [email protected]
Olli Yli-Harja Institute of Signal Processing, Tampere University of Technology, P.O. Box 553, 33101 Tampere, Finland Email: [email protected] Received 14 October 2003; Revised 17 December 2003; Recommended for Publication by John Sorensen An estimation method for correcting misclassifications in signal and image processing is presented. The method is based on the use of context-based (temporal or spatial) information in a sliding-window fashion. The classes can be purely nominal, that is, an ordering of the classes is not required. The method employs nonlinear operations based on class proximities defined by a proximity matrix. Two case studies are presented. In the first, the proposed method is applied to one-dimensional signals for processing data that are obtained by a musical key-finding algorithm. In the second, the estimation method is applied to two-dimensional signals for correction of misclassifications in images. In the first case study, the proximity matrix employed by the estimation method follows directly from music perception studies, whereas in the second case study, the optimal proximity matrix is obtained with genetic algorithms as the learning rule in a training-based optimization framework. Simulation results are presented in both case studies and the degree of improvement in classification accuracy that is obtained by the proposed method is assessed statistically using Kappa analysis. Keywords and phrases: misclassification correction, image recognition, training-based optimization, genetic algorithms, musical key finding, remote sensing.
1.
INTRODUCTION
Automatic classification of data is a standard problem in signal and image processing. In this context, the overall objective of classification is to categorize all data samples into different classes as accurately as possible. The selection of
classes depends naturally on the particular application. Powerful supervised classification methods based on neural networks, genetic algorithms, Bayesian methods, and Markov random fields have been developed (see, e.g., [1, 2, 3]). However, even the most advanced methods of automatic classification are typically unable to pro
Data Loading...
