Effects of image processing techniques on mammographic phantom images: a pilot study

Breast cancer is one of the most important diseases among females. According to the Malaysian Oncological Society (Wahid, 2007), about 4% of women who are 40 years old and above are suffering from breast cancer. Masses and microcalcifications are two important signs for breast cancer diagnosis on mammography. In this research, the effects of different image processing techniques which include enhancement, restoration, segmentation, and hybrid methods on phantom images were studied. Three different phantom images, which were obtained at 25kv (63.2 MAS), 28kv (29.8 MAS) and 35kv (9.5 MAS), were manipulated using image processing methods. The images were scored by two expert radiologists and the results were compared to explore any significant improvements. Meanwhile, the Wilcoxen Rank test was used to compare the quality of the manipulated images with the original one (alpha=0.05). Each image processing method was found to be effective on some particular criteria for image quality. Some methods were effective on just one criterion while some others were effective on a few criteria. The statistical test showed that there was an average improvement of 41 percent when the images were manipulated using the histogram modification methods. It could be concluded that different image processing methods have different effects on phantom images which generally improve radiologists’ visualization. The results confirm that the histogram stretching and histogram equation methods lead to higher improvement in image quality as compared to the original image (p < 0.05).

Preview PDF #12 Pg 67-76.pdf Download (564kB) | Preview Official URL or Download Paper: http://www.pertanika.upm.edu.my/Pertanika%20PAPERS...

Actions (login required)

View Item