Browsing by Author "Starks, S.A."
Now showing 1 - 2 of 2
Results Per Page
Sort Options
Item An Algorithm for Extraction of More than One Optimal Linear Feature from Several Gaussian Pattern Classes(1976-04-20) de Figueiredo, Rui J.P.; Pau, K.C.; Sagar, A. D.; Starks, S.A.; Van Rooy, D.L.Two algorithms have been developed at Rice University for optimal linear feature extraction based on the minimization risk (probability) of misclassification under the assumption that the class conditional probability density functions are Gaussian. One of these algorithms, which applieds to the case in which the dimensionality of the feature space (reduced space) is unity, has been described elsewhere [Rice University ICSA Technical Reports Nos. 275-025-022 and 275-025-025 (EE Technical Reports Nos. 7520 and 7603)]. In the present report, we describe the second algorithm which is used when the dimension of the feature space is greater than one. Numerical results obtained from the application of the present algorithm to remotely sensed data from the Purdue C1 flight line are mentioned. Brief comparisons are made of these results with those obtained using a feature selection technique based on maximizing the Bhattacharyya distance. For the example considered, a significant improvement in classification is obtained by the present technique.Item An Algorithm for Optimal Single Linear Feature Extraction from Several Gaussian Pattern Classes(1975-11-20) Starks, S.A.; de Figueiredo, Rui J.P.; Van Rooy, D.L.A computational algorithm is presented for the extraction of an optimal single linear feature from several Gaussian pattern classes. The algorithm minimizes the increase in the probability of misclassification in the transformed (feature) space. The general approach used in this procedure was developed in a recent paper by R.J.P. de Figueiredo. Numerical results on the application of this procedure to the remotely sensed data from the Purdue C1 flight line as well as LANDSAT data are presented. It was found that classification using the optimal single linear feature yielded a value for the probability of misclassification on the order of 30% less than that obtained by using the best single untransformed feature. Also, the optimal single linear feature gave performance results comparable to those obtained by using the two features which maximized the average divergence.