Classification of Bio-medical Images Using Neuro Fuzzy Approach

The prime requirement for medical imaging systems is to be able to display images relating to a particular disease, there is increasing interest in the use of Image Retrieval techniques to aid diagnosis by identifying similar past cases. One area where computers have scored great success in biomedicine has been medical imaging. In biomedicine, searching digital bio-medical images is a challenging problem. Bio-medical images, such as pathology slides, usually have higher resolution than general-purpose pictures. Retrieval of biomedical images will find wide usage in the next decade. In this paper a neuro fuzzy technique for classification of biomedical images on the basis of combined feature vector, which combines color and texture feature into a single feature vector, is presented. The system uses concept based on pixel descriptors, which combines the human perception of color and texture into a single vector, for region extraction. The region extracted using the feature vectors represented in the form of pixel descriptor are fed as input to a neural network, which is trained for classification of images The method takes care of “imprecision”. The user always provides partial information while posing queries. In the proposed method efforts have been made to model the imprecision using fuzzy interpretation. The technique has been implemented on the database of biomedical images. Some of the experimental results are reported in the paper. Tested on a database of more than 100 pathological / biomedical images the technique is found to be quite satisfactory. Some of the results have been reported in the paper. This technique can assist the medical community in diagnosing the disease.