Learning to Classify X-Ray Images Using Relational Learning

: Image understanding often requires extensive background knowledge. The problem addressed in this paper is such knowledge can be acquired. We discuss how relational machine learning methods can be used to automatically build rules for classifying types of blood vessels. We introduce a new learning system that can make use of background knowledge coded as arbitrarily complex Prolog programs to construct concept descriptions, particularly those needed to classify features in an image.

• Type: Conference
• Year: 1998
• Where: ECML
• Authors: Claude Sammut, Tatjana Zrimec
• Keywords: Background Knowledge, ECML 1998, Extensive Background Knowledge, Machine Learning, Machine Learning Methods