Classification of endoscopie images using support vector machines

This paper presents an application of support vector machines (SVMs) to mu I ti class problem in endoscopie image classification. Many studies have reported that SVMs have met with success in the texture classification problem. As an endoscopie image poses rich information expressed by texture features, we therefore investigate the potential of SVMs in this task. Strategy for multiclass problem based on an ensemble of binary classifiers is also implemented since the traditional SVMs algorithm deals with single label classification problems. The proposed scheme demonstrated an excellent classification result for multiclass problem in endoscopie image classification. We also show how a distortion correction helps further improve the results.