Implementation of Modified Selective Median Filter in Fpga – towards the Detection of Breast Cancer

In medical image processing, noise removal is the challenging task. Removal of noise using the existing methods like Median Filter (MF), Center Weighted Median Filter (CWMF), Rank Condition Rank Selection Filter (RCRSF) and SMF Selective Median Filter (SMF) provides satisfactory results. In this paper, we have made improvements in SMF by providing the adaptive thresholding concept and called as Modified Selective Median Filter (MSMF). Mean square error (MSE) and mean absolute error (MAE) has been calculated and the performance are compared. The performance of MSMF is better than the other methods and it is also implemented in field programmable gate array (FPGA). We found this method works good for the noise removal in mammograms for the detection of breast cancer. Furthermore, the method has been tested for the sample image mdb218 taken from mammographic image analysis society (MIAS) and the results found were good.