Case Based Reasoning in the Detection of Retinal Abnormalities Using Decision Trees

The most common abnormalities in retina images occur due to age related macular degeneration and diabetic retinopathy. In this paper a decision support system is proposed to classify these abnormalities. The process involves the combination of contextual information with images obtained from a database. Decisions trees are proposed for this purpose as they can combine contextual information with images. Images are pre- processed and segmented to obtain the regions of interest for the individual manifestations in each of these diseases. Matching of candidate segmented images with prototype segmented images is performed along with the matching of the associated contextual information.