Master'sOpen Access

Spektral kümeleme kullanan Renkli Görüntü bölütleme

2017
0 views
0 downloads
Advisor: Doç. Dr. Oğuz Bayat

Abstract (EN)

the spectral clustering has newly arise and has become one of the most common clustering algorithms, and the learning algorithm is considered uncensored. It is easy to apply and can be efficiently solved using standard linear algebra software. Often outweigh the normal clustering algorithms, for example, the K-mean algorithm. segmentation is a digital image split that is entered into several regions and re-image representation to useful elements and more clearly for analysis. The process of color-based segmentation is greatly influenced by the color of the space. the L*a*b color space is the best representative of the contents of the color image. In this dissertation, an algorithm was developed to segment the color image using L*a*b color space and then the spectral algorithm was applied to the data to classify it. Description of shape nor representation is an important issue both in the identification and classification of objects. The resulting color segmentation scheme has been applied to some of the images and empiricism data indicate a well-advanced segmentation algorithm if the coefficients are better configured.

Author

Dr. Edrees Ramadan Marsel Edrees Ramadan Marsel

How to Cite

Edrees Ramadan Marsel Edrees Ramadan Marsel (Master Thesis). Spektral kümeleme kullanan Renkli Görüntü bölütleme, 2017, Altınbaş University.

Keywords

License

Tüm Hakları Saklıdır

This work is shared under the specified license terms.

More theses from Altınbaş University