Mammographic phantom images using receiver operating characteristic analysis

Abstract

Breast cancer is a form of cancer which is a leading cause of death among women worldwide. Some digital mammographic images are noisy and have low contrast. Image processing techniques have been used to improve the quality of images. The aim of this study is to perform receiver operating characteristic (ROC) analysis on mammographic phantom images subjected to two enhancement techniques in order to verify whether the enhancements improve the quality of the images. The digital mammographic images were enhanced using the morphological and wavelet transform techniques. For the morphological techniques, the images were enhanced using dilation for morphological operation and morphological closing. For wavelet transform enhancement, biorthogonal 2.8 wavelet filter with two levels of decomposition (L=2) was used. Four observers evaluated the images that contain fibres, nodules and micronodules. The detection performances of the original and enhanced images were evaluated using ROC curves. The images were also rated based on contrast visibility, sharpness and overall image quality. The ROC analysis showed that detection of micronodules gave higher area index of curves and sensitivity than the detection of nodules and fibrils in all image datasets. For evaluation of overall image quality using observers’ subjective rating scale, original images have higher mean values than the enhanced images.