Medical Image Categorization and Retrieval System in Radiology Using Bag of Visual Words Framework

Abstract

In This work we present an efficient image categorization and retrieval system applied to Image Clef 2009 medical image retrieval task. In this task we have presented methodology is based on local patch representation of the image content and a bag-of-features approach for defining image categories, with a kernel based SVM classifier. Two main tasks are addressed: First organ identification task; second the detection and identification of pathologies, i.e. shifting from the organ level to pathology level analysis. We used a large generic archive of 12,000 radiographs (IRMA) to tune the system parameters. We demonstrate automated organ detection on the IRMA collection as well as the generalization to a new data collection. We submitted one run, using support-vector-machines trained on the visual word histograms in multiple scales. We proposed system was helped to find discriminating orientation and body regions in X-ray images also organ-level discrimination we show an application to pathology level categorization of chest X-ray data. Results indicate detection of pathology at a sensitivity of 88.4% and a specificity of 81%. This is first step towards similarity-based medical image categorization that has a major clinical importance in computer-assisted diagnostics. It can identify suspicious pathological X-rays and alert the referring clinicians to potential emergencies. Overall it is hoped that the development of such systems will contribute to the improvement of safety and quality of medical services