Data mining, Classification and Clustering with Morphological features of Microbes

The idiosyncrasies of the medical profession makes medical image mining a challenge. Image mining deals with the extraction of implicit knowledge, image data relationship, or other patterns not explicitly stored in the images. It is very difficult to determine the exact number of microorganisms under the microscope in the presence of a human expert in conventional methods. An automated tool for fast recognition of microbes is needed to examine the medical data before it expires. Digital image processing is an integral part of microscopy. Automated color image segmentation for bacterial image is proposed to classify the bacteria into two broad categories of gram images. Edge detection algorithm with 8 neighbor-connectivity contour is used. Bacterial morphological, geometric features extracted from microscopy images are used for classification and clustering. The potential and distinguished features are extracted from each bacterial cell. Experimental results with self organizing map shows that the bacterial cluster patterns obtained are better than the statistical approach.