Handwriting recognition has attracted enormous scientific interest because of its potential for improved man/machine interfaces. We have designed an on-line handwritten digit recognition system after the examination of different techniques based on statistical and neural pattern recognition approaches. We collected a digit database from 44 people. We use two representations. The dynamic representation is based on constant length feature vectors of equally distanced points on the pen trajectory. The static representation converts the dynamic information to an image similar to images used in off-line recognition tasks.Then, we tested the well known statistical classification method k-nearest neighbor (k-NN) and neural multi-layer perceptron (MLP) and recurrent networks using both representations. Classifiers trained with dynamic and static representations make misclassifications for different samples. We combine them first by forming a feature vector composed of dynamic and static repr...

In many pattern recognition tasks, an approach based on combining classifiers has shown a significant potential gain in comparison to the performance of an individual best classifier. This improvement turned out to be subject to a sufficient level of diversity exhibited among classifiers, which in general can be assumed as a selective property of classifier subsets. Given a large number of classifiers, an intelligent classifier selection process becomes a crucial issue of multiple classifier system design. In this paper, we have investigated three evolutionary optimization methods for the classifier selection task. Based on our previous studies of various diversity measures and their correlation with majority voting error we have adopted majority voting performance computed for the validation set directly as a fitness function guiding the search. To prevent from training data overfitting we extracted a population of best unique classifier combinations, and used them for second stage majority voting. In this work we intend to show empirically, that using efficient evolutionary-based selection leads to the results comparable to absolutely best, found exhaustively. Moreover, as we showed for selected datasets, introducing a second stage combining by majority voting has the potential for both, further improvement of the recognition rate and increase of the reliability of combined outputs.

Decision templates (DT) are a technique for classier fusion for continuous-valued individual classier outputs. The individual outputs considered here sum up to the same value (e.g., statistical classiers, yielding some estimates of the posterior probabilities for the classes). First, the DT fusion algorithm is explained. Second, we show that two similarity measures (S1 and S2 ) and two inclusion indices (I1 and I2) between fuzzy sets (see Dubois and Prade, 1980) produce the same DT classier. The equivalence is proven by showing that for every object submitted for classication, all 4 measures induce the same ordering on the set of class labels (through DT fusion), thereby assigning the object to the same class. Keywords: Pattern recognition, multiple classier fusion, aggregation, decision templates, measures of similarity and inclusion 1 Introduction The objective of combination of a set of classiers is to achieve a higher accuracy than the accuracy of the best individual in th...

Abstract. This paper describes a method for fusing a collection of classifiers where the fusion can compensate for some positive correlation among the classifiers. Specifically, it does not require the assumption of evidential independence of the classifiers to be fused (such as Dempster Shafer’s fusion rule). The proposed method is associative, which allows fusing three or more classifiers irrespective of the order. The fusion is accomplished using a generalized intersection operator (T-norm) that better represents the possible correlation between the classifiers. In addition, a confidence measure is produced that takes advantage of the consensus and conflict between classifiers. 1

We study classifier fusion by the fuzzy template (FT) technique. Given an object to be classified, each classifier from the pool yields a vector with degrees of "support " for the classes, thereby forming a decision profile. A fuzzy template is associated with each class as the averaged decision profile over the training samples from this class. A new object is then labeled with the class whose fuzzy template is closest to the objects' decision profile. We give a brief overview of the field to place the FT approach in a proper group of classifier combination techniques. Experiments with two data sets (satimage and phoneme) from the ELENA database demonstrate the superior performance of FT over: a version of majority voting; aggregation by fuzzy connectives (minimum, maximum, and product); and (unweighted) average. 1 Introduction We consider a pattern classification problem where x 2 ! p is a feature vector to be labeled into one or more of c classes. Every mapping D : ! p ! f1; ...

Abstract—This paper addresses the problem of automatically extracting perceptive information from acoustic signals, in a supervised classification context. Global labels, i.e., atomic information describing a music title in its entirety, such as its genre, mood, main instruments, or type of vocals, are entered by humans. Classifiers are trained to map audio features to these labels. However, the performances of these classifiers on individual labels are rarely satisfactory. In the case we have to predict several labels simultaneously, we introduce a correction scheme to improve these performances. In this scheme—an instance of the classifier fusion paradigm—an extra layer of classifiers is built to exploit redundancies between labels and correct some of the errors coming from the individual acoustic classifiers. We describe a series of experiments aiming at validating this approach on a large-scale database of music and metadata (about 30 000 titles and 600 labels per title). The experiments show that the approach brings statistically significant improvements. Index Terms—Feature extraction, learning systems, music, pattern classification. I.

Abstract not found

Abstract. One advantage of evolutionary multiobjective optimization (EMO) algorithms over classical approaches is that many non-dominated solutions can be simultaneously obtained by their single run. In this paper, we propose an idea of using EMO algorithms for constructing an ensemble of fuzzy rule-based classifiers with high diversity. The classification of new patterns is performed based on the vote of multiple classifiers generated by a single run of EMO algorithms. Even when the classification performance of individual classifiers is not high, their ensemble often works well. The point is to generate multiple classifiers with high diversity. We demonstrate the ability of EMO algorithms to generate various non-dominated fuzzy rule-based classifiers with high diversity by their single run. Through computational experiments on some wellknown benchmark data sets, it is shown that the vote of generated fuzzy rulebased classifiers leads to high classification performance on test patterns. 1

We introduce a multiple classifier system that incorporates a global optimization technique based on a Genetic Algorithm for dynamically selecting the set of experts to use in the majority vote approach. The proposed technique is applicable when the experts in the pool provide both the class assigned to the input sample and a measure of the reliability of the this classification. For each sample, the experts selected for participating in the majority vote are those whose reliability is larger than a given threshold. There are as many thresholds as the number of experts by the number of classes. The values of the thresholds aimed at selecting the best set of experts for each input sample are determined by a canonical Genetic Algorithm. The reliability measures provided by the experts of the pool are also used to implement the tie-break mechanism needed within the majority vote scheme. The system has been tested on a handwritten digit recognition problem, and its performance compared with those exhibited by other multiexpert systems exploiting different combining rules.

Abstract not found