Text-based Medical Image Retrieval using Query Modification Methods

Web proceedings papers

Authors

Ivan Kitanovski , Katarina Trojacanec , Ivica Dimitrovski and Suzana Loshkovska

Abstract

This paper presents a strategy for text-based retrieval of medical images by applying query modification. The goal of the paper is to investigate whether query modification techniques can improve the retrieval performance. The query modification includes expansion of the queries by medical synonyms or traditional query expansion techniques based on the retrieved document collection. Images for the experiment are extracted from medical articles. The text from the medical articles is used to create a field-based textual representation of the images. We used Terrier IR search engine due to its flexibility and scalability. The results showed that query modification methods can increase the performance of the medical image retrieval.

Keywords

/