Machine Learning Methods in Automatic Image Annotation

Abstract

Machine learning methods are successfully applied in many branches of Computer Science. One of these branches is image analysis, and being more specific – Automatic Image Annotation. Automatic Image Annotation was found an important research domain several years ago. It grew from such research domains as image recognition and cross-lingual machine translation. Increase of computational, data storage and data transfer abilities of todays’ computers has been one of key factors, making Automatic Image Annotation possible. Automatic Image Annotation methods, which have appeared during last several years, make a large use of many machine learning approaches. Clustering and classification methods are most frequently applied to annotate images. The chapter consists of three main parts. In the first, some general information concerning annotation methods is presented. In the second part, two original annotation methods are described. The last part presents experimental studies of the proposed methods.