年間 4 号発行
ISSN オンライン: 2162-3511
Application of a wavelet thresholding algorithm for contrast enhancement of mammograms
要約
We present a novel algorithm for enhancement and visibility of low-contrast features in breast mammograms. Contrast of mammograms is enhanced by the use of mathematical morphology and wavelet denoising with an emphasis on differentiating edge pixels from noise to enhance visible diagnostic signs like masses and microcalcifications. We present the results of our proposed method and compare it with three other contrast-enhancement algorithms based on the quantitative metrics of the contrast improvement index (CII) and the edge preservation index (EPI), as well as radiological evaluation. Our results based on these metrics indicate that our method outperforms the three other approaches. The results of the clinical evaluation are in good agreement with the comparative analyses based on CII and EPI.