R.O. Duda, P.E. Hart, and D.G. Stork, Pattern Classification (2d ed.), New York: Wiley & Sons, 2001.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....Trees. 5 Multivariate Decision Trees Multivariate Decision Trees (MDT s) 4] create piecewise linear approximations of regions in feature space by recursively dividing feature space with hyperplanes (figure 6) MDT s recursively subdivide the feature space by linear threshold units (LTU s) [33, 12]. The LTU s are binary tests, represented by linear combinations of feature values and associated weights. Each division attempts to separate, in a set of known instances (the training set) target instances from non targets. If two subsets are linearly separable, a single LTU will separate them ....

R.O. Duda and P.E. Hart, Pattern Classification and Scene Analysis, New York: Wiley & Sons, 1973.

....are amenable to optimization using a diversity of methods, particularly GAs. Classical Vector Quantization (VQ) algorithms try to compute an optimal group of labeled vectors, called dictionary, that represent a given sample and can be used to recognize previously unseen members of that sample [2]. A dictionary is composed of l reference vectors or codevectors, being l called number of levels of the dictionary. An lvq network is actually a dictionary, but each vector in the dictionary is conventionally called weight vectors. Each label corresponds to a class. In the case of lvq , a ....

....lvq1 after new versions were developed. lvq usually performs efficiently if the initial weight vectors are close to their final values. For this reason some heuristic procedures are used for weight initialization, like Kohonen s som [1] or any other vector quantization procedure, like k means [2]. Monte [3] has used a genetic algorithm to set the initial weights of a lvq neural net. In lvq , the number of weight vectors remains fixed during training, and it is usually set to be equal to the number of classes, or to a fixed number of weight vectors per class. Several sets of initial weight ....

Duda, R. O., Hart, P. E., Pattern classification and scene analysis, New York: Wiley & Sons, 1973.

....in feature space by recursively dividing feature space with hyper planes (figure 5) Multivariate decision trees have been developed over the past few years [4] but have not been widely applied in computer vision. MDT s recursively subdivide the feature space by linear threshold units (LTU s) [27, 9]. The LTU s are binary tests, represented by linear combinations of feature values and associated weights. Each division attempts to separate, in a set of known instances (the training set) target instances from non targets. If the two resulting subsets are linearly separable, a single LTU will ....

R.O. Duda and P.E. Hart, Pattern Classification and Scene Analysis, New York: Wiley & Sons, 1973.

....The techniques presented are demonstrated on a challenging task: detecting camouflaged vehicles in outdoor scenes. 1. Introduction Classifying objects based on their color (and in some cases, texture) is one of the oldest problems in computer vision and pattern recognition (see, for example, [Duda and Hart 1973]) In indoor settings, where lighting and other imaging conditions can be controlled, maximum likelihood classifiers recognize objects by modeling variations in color as This work was supported by ARPA through Rome Labs under contract F30602 94 C 0042. Gaussian noise in feature space, and ....

....the region; if not, find the hyperplane(s) that maximally separates instances of the two labels, divide feature space into two regions using this hyperplane, and recurse on each region. The hyperplanes used to divide feature space are represented as linear threshold units (LTUs) Nilsson 1965, Duda and Hart 1973]. Several methods exist for learning the weights in a linear threshold unit. Brodley and Utgoff [Brodley and Utgoff] discuss four such methods: the Recursive Least Squares (RLS) algorithm [Young 1984] the Pocket al..gorithm [Gallant 1986] Thermal Training [Frean 1990] and CART s coefficient ....

Duda, R. O. and Hart, P.E., Pattern Classification and Scene Analysis, New York: Wiley & Sons, 1973.

No context found.

R.O. Duda, P.E. Hart, and D.G. Stork, Pattern Classification (2d ed.), New York: Wiley & Sons, 2001.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC