Pattern Classification of Human Epithelial Images

This project shows an important role to diagnosis autoimmune disorder which is by a comparative analysis on the most appropriate clustering technique for the segmentation and also to develop algorithm for positivity classification. In this project, there are four stages will be used to analyze pattern classification in human epithelial (HEp-2) images. First of all, image enhancement will take part in order to boost efficiency of algorithm by implementing some of the adjustment and filtering technique to increase the visibility of image. After that, the second stage will be the image segmentation by using most appropriate clustering technique. There will be a comparative analysis on clustering techniques for segmentation which are adaptive fuzzy c-mean and adaptive fuzzy moving k-mean. Then, for features extraction, by calculating the mean of each of the properties such as area, perimeter, major axis length, and minor axis length for each images. After that, will implementing a grouping based on properties dataset that has been calculated. Last but not least, from the mean of properties, it will classify into the pattern after ranging the value of mean properties of each of the pattern itself that has been done in classification stage.