Image clustering comparison of two color segmentation techniques

• Main Author: Subramaniam, Kavitha Pichaiyan
• Format: Thesis
• Language: English; English
• Published: 2010
• Subjects: T Technology (General); TA Engineering (General). Civil engineering (General)
• Online Access: http://eprints.utem.edu.my/id/eprint/12808/1/Image_clustering_comparison_of_two_color_segmentation_techniques24_pages.pdf; http://eprints.utem.edu.my/id/eprint/12808/3/Image%20clustering%20comparison%20of%20two%20color%20segmentation%20techniques.pdf; http://eprints.utem.edu.my/id/eprint/12808/; https://plh.utem.edu.my/cgi-bin/koha/opac-detail.pl?biblionumber=63009
• Institution: Universiti Teknikal Malaysia Melaka

The clustering research is regarding the area of data mining and implementation of the clustering algorithms. The image clustering is major part of data mining where study about how to binds the similar data together in a cluster and show the meaningful data. There are many algorithm for analysing clustering each having its own method to do clustering. This clustering technique increasingly common and has yield many insights into segmentation factors, would effect image functioning and performance. The enormous researches going on extract image with background subtraction. We focus on the outlier detection and background subtraction on image. This project proposed a two color segmentation techniques such as K-means and Fuzzy C-means clustering algorithm that are accurately segment the desired images, which have the same color as the pre-selected pixels with background subtraction. In the software development testing we examine image based clustering, as we can used clustering by distance base, by pixel (red, green, blue) value etc., The problem is solved by region based method which is based on connect component and background detection techniques. The appropriate Java codes are developed for solve this task. The developed patterns are applied in the field of real-time analysis. Finally, the algorithm found, which would solve the image segmentation problem.