TY - GEN
T1 - Using uncertainty information to combine soft classifications
AU - Gonçalves, Luisa M. S.
AU - Fonte, Cidália C.
AU - Caetano, Mário
PY - 2010/7/29
Y1 - 2010/7/29
N2 - The classification of remote sensing images performed with different classifiers usually produces different results. The aim of this paper is to investigate whether the outputs of different soft classifications may be combined to increase the classification accuracy, using the uncertainty information to choose the best class to assign to each pixel. If there is disagreement between the outputs obtained with the several classifiers, the proposed method selects the class to assign to the pixel choosing the one that presents less uncertainty. The proposed approach was applied to an IKONOS image, which was classified using two supervised soft classifiers, the Multi-layer Perceptron neural network classifier and a fuzzy classifier based on the underlying logic of the Minimum-Distance-to-Means. The overall accuracy of the classification obtained with the combination of both classifications with the proposed methodology was higher than the overall accuracy of the original classifications, which shows that the methodology is promising and may be used to increase classification accuracy.
AB - The classification of remote sensing images performed with different classifiers usually produces different results. The aim of this paper is to investigate whether the outputs of different soft classifications may be combined to increase the classification accuracy, using the uncertainty information to choose the best class to assign to each pixel. If there is disagreement between the outputs obtained with the several classifiers, the proposed method selects the class to assign to the pixel choosing the one that presents less uncertainty. The proposed approach was applied to an IKONOS image, which was classified using two supervised soft classifiers, the Multi-layer Perceptron neural network classifier and a fuzzy classifier based on the underlying logic of the Minimum-Distance-to-Means. The overall accuracy of the classification obtained with the combination of both classifications with the proposed methodology was higher than the overall accuracy of the original classifications, which shows that the methodology is promising and may be used to increase classification accuracy.
KW - combining soft classifications
KW - Soft classifiers
KW - uncertainty information
UR - http://www.scopus.com/inward/record.url?scp=77954877354&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-14049-5_47
DO - 10.1007/978-3-642-14049-5_47
M3 - Conference contribution
AN - SCOPUS:77954877354
SN - 3642140483
SN - 9783642140488
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 455
EP - 463
BT - Computational Intelligence for Knowledge-Based Systems Design - 13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010, Proceedings
T2 - 13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010
Y2 - 28 June 2010 through 2 July 2010
ER -