Thesis (M.Sc.) - Cairo University - Faculty of Computers and Information - Department of Information Systems

Medical images became a huge problem due to the fast growing size of the medical image repositories, thousands of medical images are produced daily. Medical image repositories need to be well organized using an efficient and fast tool to allow researches or medical experts to extract useful information in the right time and as fast as possible. Organizing large medical image repositories can help in many fields as in medical fields that can be useful in diagnosis and knowing the history of a patient and in the researching area as it can be mined easily and be a necessary step before many application as content based image retrieval and medical image classification application. The objective of this thesis is to implement a new efficient clustering technique for medical images. This technique contains three main methods, the first is to extract features using gray-level co-occurrence matrix and apply PCA for dimensionality reduction, and then k-means clustering is applied. The second method where the 2D wavelet transforms is applied as a feature extraction and feature selection is used to select most efficient attributes, then k-means clustering is applied. The final and proposed technique is to combine the two models and apply k-means clustering

Issued also as CD