Improvement of masses detection in digital mammograms employing non-linear filtering

Computer aided detection of mammographic masses can be improved to a greater extent employing non-linear filters for image enhancement. The present work proposes a truncated Volterra filter combination to provide contrast enhancement as well as texture based processing of masses in digital mammograms. Noteworthy improvement in visualization of masses has been observed in the simulation results carried out on cases from DDSM database. The improved performance of the proposed filtering approach is well supported with calculated values of objective evaluation parameters.