M.Sc. Tezi Görüntüleme

Student: SOHEİLA VAFAEİNAHR
Supervisor: YRD. DOÇ. DR. ORHAN KESEMEN
Department: İstatistik ve Bilgisayar Bilimleri (İstatistik)
Institution: Graduate School of Natural and Applied Sciences
University: Karadeniz Technical University Turkey
Title of the Thesis: HUE-BASED SEGMENTATION OF COLOR IMAGES USING WEIGHTED DIRECTIONAL CLUSTERING ALGORITHMS
Level: M.Sc.
Acceptance Date: 10/2/2015
Number of Pages: 56
Registration Number: i2888
Summary:

      In this study, to analyze the color images, one of the pre-processing steps are focused on the image segmentation process. For this procedure, a clustering process to be performed in RGB space, a small image is seen that even reached the incredible computing time. Another approach has been a histogram based approach. However, in histogram approach can be applied in single-channel or gray-scale images. Thus, in a color image, color-based segmentation is considered an approach to extract. However, since the hue angle based showed a change, it was not possible to use the number line based operations. To solve the problem, directional-based clustering algorithms have been applied. Directional based algorithms, measuring performance on the image segmentation process are compared.

Key Words: Image Segmentation, Hue-Based Segmentation, Directional Data, Clustering Algorithms.