Enhancement of Screening Mammograms Using Dual-Tree Complex Wavelet Transform

Breast cancer has become one of the most common cancers worldwide. It is also one of the leading causes of death in women. Early detection of breast cancer is the need of the hour as it offers increased treatment options, better survival chances, and improved quality of life. The most economical technique is screening mammograms to promote breast cancer identification at an early stage. This paper proposes an enhancement strategy for screening mammograms based on Dual-Tree Complex Wavelet Transform (DTCWT). The DTCWT yields six directionally selective high-pass sub-bands on multiple scales. The reconstruction from the decomposed bands produces a resultant image with enhancement in directional edges. Thus, edge details are enhanced, improving the visualisation of mass boundaries inside the breast tissues. The proposed method offers PSNR values of 54.3612, 46.043, 48.8719 and SSIM is 0.9977, 0.9869, 0.9916 for the MIAS, DDSM and Inbreast datasets respectively. The proposed DTCWT-based mammogram enhancement can be used as an efficient preprocessing technique in a computer-aided diagnosis system to detect breast cancer in its early stages. The comparison of the proposed method with many of the recent algorithms depicts the overall dominance of its performance, with minimum error and good signal enhancement.