Description

Mammography is the most effective procedure for the early detection of breast cancer. In this paper, we develop a novel algorithm to detect suspicious lesions in mammograms. For detecting the lesion here we are using the adaptive threshold method. While producing mammogram there is a chance to occur noise the image. So here we remove the noise in initial step. Then we apply the wavelet transform to decompose from the decomposition here we find the probability of the mammogram image. Then we apply the adaptive threshold method it will segment the mammogram as the coarse segmentation. The segmented result passed to the window based adaptive thresholding technique. In window adaptive threshold we separate the image as patches. For each patch here we compute the mean value for each patch. The mean value is the threshold value. By this value we segment the lesions from the mammogram images. Finally we calculating the area, perimeter, no of lesions are calculated. The adaptive thresholding detection algorithm is proved to be an effective method to detect lesions in mammogram. In these algorithms, the key step is to segment a suspicious region by a threshold. The threshold selected by the PDF curve is a global threshold for the whole image. For obvious lesion regions, the gray-level values have a global superiority in the whole images and can be easily segmented as suspicious.