Image Mining for Establishing Medical Diagnosis

In this paper, a method based on association rules that support medical image diagnosis is developed. The proposed method has important characteristics that make it different from other computer assisted diagnosis methods: the process is completely automatic, having the possibility to define a great number of diagnoses; the method could be applied to any medical domain, because the visual features, the semantic indicators remain unchangeable, and the semantic rules are generated by learning from labelled images-examples; the selection of the visual characteristics set is based on their retrieval accuracy; the spatial information of the regions is considered, offering important medical information of the relationships between sick regions. Although we present the results achieved in endoscopic images analysis, our method can be used to analyze other types of medical images. The prototype system was applied to real datasets and the results show high accuracy.