AN ITERATIVE SEGMENTATION METHOD FOR REGION OF INTEREST EXTRACTION

Abstract

In this paper, a method is presented for applications which include mammographic image segmentation and region of interest extraction. Segmentation is a very critical and difficult stage to accomplish in computer aided detection systems. Although the presented segmentation method is developed for mammographic images, it can be used for any medical image which resembles the same statistical characteristics with mammograms. Fundamentally, the method contains iterative automatic thresholding and masking operations which is applied to the original or enhanced mammograms. Also the effect of image enhancement to the segmentation process was observed. A version of histogram equalization was applied to the images for enhancement. Finally, the results show that enhanced version of the proposed segmentation method is preferable because of its better success rate

Keywords

• Digital Image Processing, Mammography, Image Segmentation, Image Enhancement

Kaynakça

1. Processing Techniques for Tumor Detection”, Marcel Dekker, Inc. 270 Madison Avenue, New York, (2002) Strickland,
2. “Image [2] Aided Detection System for Mammographic Masses”, PhD Thesis, Duke University (2004) [3] Processing
3. Cambridge University Press, The Edinburgh Building, Cambridge CB2 8RU, UK, (2009) [4] Research
4. http://turkkanser.org.tr/news.php?id=61 (Date of Access: 15.11.2009)
5. Control, [5] Health,
6. http://www.saglik.gov.tr/TR/Genel/BelgeGoste r.aspx?F6E10F8892433CFF7A2395174CFB32 E1BFEA5D815DAC9C9E (Date of Access: 15.11.2009)
7. Method from Gray-Level Histograms”, IEEE Transactions on Systems, Man and Cybernetics, Vol. SMC-9, No. 1, (1979)