We develop a common theoretical framework for combining classifiers which use distinct pattern representations and show that many existing schemes can be considered as special cases of compound classification where all the pattern representations are used jointly to make a decision. An experimental comparison of various classifier combination schemes demonstrates that the combination rule developed under the most restrictive assumptions—the sum rule—outperforms other classifier combinations schemes. A sensitivity analysis of the various schemes to estimation errors is carried out to show that this finding can be justified theoretically.

In classification tasks it may be wise to combine observations from di!erent sources. Not only it decreases the training time but it can also increase the robustness and the performance of the classi"cation. Combining is often done by just (weighted) averaging of the outputs of the di!erent classi"ers. Using equal weights for all classi"ers then results in the mean combination rule. This works very well in practice, but the combination strategy lacks a fundamental basis as it cannot readily be derived from the joint probabilities. This contrasts with the product combination rule which can be obtained from the joint probability under the assumption of independency. In this paper we will show di!erences and similarities between this mean combination rule and the product combination rule in theory and in practice. # 2000 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.

. In the problem of one-class classification target objects should

Abstract—Classifier fusion may generate more accurate classification than each of the basic classifiers. Fusion is often based on fixed combination rules like the product, average etc. This paper presents decision templates as classifier fusion method for the recognition of the handwritten English and Farsi numerals (1-9). The process involves extracting a feature vector on well-known image databases. The extracted feature vector is fed to multiple classifier fusion. A set of experiments were conducted to compare decision templates (DTs) with some combination rules. Results from decision templates conclude 97.99 % and 97.28 % for Farsi and English handwritten digits. Keywords—Decision templates, multi-layer perceptron, characteristics Loci, principle component analysis (PCA). I.

Abstract—Combining classifiers is a useful method for solving complex problems in machine learning. The ECOC (Error Correcting Output Codes) method has been widely used for designing combining classifiers with an emphasis on the diversity of classifiers. In this paper, in contrast to the standard ECOC approach in which individual classifiers are chosen homogeneously, classifiers are selected according to the complexity of the corresponding binary problem. We use SATIMAGE database (containing 6 classes) for our experiments. The recognition error rate in our proposed method is %10.37 which indicates a considerable improvement in comparison with the conventional ECOC and stack generalization methods. Keywords—Error correcting output code, combining classifiers, neural networks. I.

We present a multiscale unsupervised segmenter for automatic detection of potentially cancerous regions of interest containing fibroglandular tissue in digital screening mammography. The mammogram tissue textures are locally represented by four causal multispectral random field models recursively evaluated for each pixel and several scales. The segmentation part of the algorithm is based on the underlying Gaussian mixture model and starts with an over segmented initial estimation which is adaptively modified until the optimal number of homogeneous mammogram segments is reached. The performance of the presented method is verified on the Digital Database for Screening Mammography (DDSM) from the University of South Florida as well as extensively tested on the Prague Texture Segmentation Benchmark and compares favourably with several alternative unsupervised texture segmentation methods. 1.

Segmentation is the fundamental process which partitions a data space into meaningful salient regions. Image segmentation essentially affects the overall performance of any automated image analysis system thus its quality is of the utmost importance. Image regions, homogeneous with respect to some usually textural or colour measure, which result