Automatic medical image classification for content based image retrieval systems.
Epaphrodite Uwimana, Miguel E Ruiz · PubMed · 2008
This paper discusses results and methods used to automatically classify medical images for Content Based Image Retrieval (CBIR) systems. Using a supervised learning approach, we automatically classified over 3,000 medical images according to the four facets of IRMA classification code (that's image modality, body orientation, biological system, and anatomical part). Our best results were obtained in the image modality facet classification with an overall error-rate of 1%.