An improved Unsharp Masking algorithm for enhancement of mammographic masses

Unsharp Masking (UM) is a classical and popular tool for sharpening in digital images. Mammographic masses with diameters less than 1 cm are generally benign, while those with diameters above 2 cm could prove cancerous and hence require further investigation. This paper introduces two new filter templates to catalyze the performance of UM algorithm using an improved high pass filter. The conventional UM algorithm is extremely sensitive to noise because of the presence of the linear high pass filter, on the other hand region segmentation does not prove effective when dealing with the objects having multiple discontinuities. The proposed algorithm combines the conventional UM algorithm with the region segmentation approach. Simulation results show a marked improvement in performance in terms of contrast and edge enhancement as compared to the recently published works in literature. Moreover the results are also subjectively evaluated by medical experts which validates the effectiveness of the proposed technique.