A technique for feature selection in multiclass problems

One of the main phases in the development of a system for the classiŽ cation of remote sensing images is the deŽ nition of an e€ ective set of features to be given as input to the classiŽ er. In particular, it is often useful to reduce the number of features available,while saving the possibility to discriminate among the di€ erent land-cover classes to be recognized. This paper addresses this topic with reference to applications that involve more than two land-cover classes (multiclass problems). Several criteria proposed in the remote sensing literature are considered and compared with one another and with the criterion presented by the authors. Such a criterion, unlike those usually adopted for multiclass problems, is related to an upper bound to the error probability of the Bayes classiŽ er. As the objective of feature selection is generally to identify a reduced set of features that minimize the errors of the classiŽ er, the aforementioned property is very important because it allows one to select features by taking into account their e€ ects on classiŽ cation errors. Experiments on two remote sensing datasets are described and discussed. These experiments conŽ rm the e€ ectiveness of the proposed criterion, which performs slightly better than all the others considered in the paper. In addition, the results obtained provide useful information about the behaviour of di€ erent classical criteria when applied in multiclass cases.