Rough set theory is an approach and a methodology to handle ambiguity or uncertainty. We will write this methodology into a programmed system, where the user specifies the classes in the data by selecting particular districts in 2d slices.
Rough set theory used to compute lower and upper approximation of the classes. The boundary region between the lower and the upper approximations represents the ambiguity or doubt of the classification. We will present an approach to automatically compute classification rules from the rough set classification using k-means approach. The rule generation removes redundancies, which allows us to enhance the original feature space attributes with a number of further feature and object space attributes. The rules can be transferred from one 2D slice to the entire 3D data set to produce a 3D segmentation result. The result can be refined by the user interactively adding more samples (from the same or other 2D slices) to the respective classes. Our system allows for visualization of both the classification result and the uncertainty of the individual class representations. The methods can be applied to single, as well as, multichannel imaging data. we will apply it to medical imaging data with RGB (red green blue) color channels.