Medical image modality classification using discrete Bayesian networks.

We propose and evaluate a pipeline for the use of visual descriptors extracted from medical images as input in discrete Bayesian Network Classifiers.We compare the results obtained thanks to our pipeline with other proposals in the scenario of the ImageCLEFmed 2013 competition.When coping with classification problems including large number of classes, hierarchical approaches are supplementary for increasing the baseline accuracy.The proposed discretization and feature subset selection techniques allow for a proper integration of any combination of visual descriptors. Moreover, the resulting number of variables does not necessarily increase when integrating new descriptors.In contrast to other participant's proposals, we present a generalist classification system (ranking 3rd out of 8) that has not been optimized to the competition problem.The use of probabilistic classifiers allows us for a deep result analysis, which let us identify the weak points in the discrimination capabilities. In this paper we propose a complete pipeline for medical image modality classification focused on the application of discrete Bayesian network classifiers. Modality refers to the categorization of biomedical images from the literature according to a previously defined set of image types, such as X-ray, graph or gene sequence. We describe an extensive pipeline starting with feature extraction from images, data combination, pre-processing and a range of different classification techniques and models. We study the expressive power of several image descriptors along with supervised discretization and feature selection to show the performance of discrete Bayesian networks compared to the usual deterministic classifiers used in image classification. We perform an exhaustive experimentation by using the ImageCLEFmed 2013 collection. This problem presents a high number of classes so we propose several hierarchical approaches. In a first set of experiments we evaluate a wide range of parameters for our pipeline along with several classification models. Finally, we perform a comparison by setting up the competition environment between our selected approaches and the best ones of the original competition. Results show that the Bayesian Network classifiers obtain very competitive results. Furthermore, the proposed approach is stable and it can be applied to other problems that present inherent hierarchical structures of classes.