Abstract: This study deals with the fast and efficient algorithm for segmenting the cervical image using unsupervised methods. Here, the image segmentation method is based on basic region growing method. In this study, both fast k-means with optional weighting and careful initialization and Fuzzy c-means Clustering algorithm are used to deal with segmentation of the cervical image. Both algorithm segments the image in accordance with the colour for each cluster and its neighborhood. In this method, first the cervical image is smoothed, enhanced and converted into L*u*v color space. The L*u*v color space image is segmented using fast k-means algorithms with optional weighting and careful seeding and fuzzy c-means algorithm. Finally, the performance analysis of the three segmentation algorithms is carried out. Experimental results show that fast k-means segmentation with careful seeding methods are fast as compare to Fast K-means with weight and fuzzy c-means method, three algorithm gives better segmented images with finer details and accurate location but FCM takes more time.
Anantha Sivaprakasam Sivaprakasam and Naganathan Ealai Rengasari, 2015. Segmentation of Cervical Image Using Unsupervised Clustering Algorithms with L*u*v Color Transformation. Asian Journal of Information Technology, 14: 147-153.