The most effective method for detection of early breast cancer is mammography, which is the most reliable method for the detection of early breast cancer, Mammography is the compulsory and only option for the premature detection of breast cancer in women, of all diagnostic methods currently available for this purpose this is more reliable method.This paper present a wavelet and window method to detect breast tumours using wavelet thresholding. The compressed image is decomposed at level 4 using daubechies 6 wavelet and histograms for all components (approximation, horizontal, vertical & diagonal) is evaluated for first two level and only approximation components are calculated for remaining two level. Further the histogram of horizontal, vertical and diagonal components are decomposed at level 5 and the thresholds corresponding to minima of the histogram is found. Transferring those thresholds at components horizontal, vertical and diagonal are thresholded. Further a global threshold is found to segment reconstructed image, fine segmentation is done using window based operation & using a wavelet approach for final threshold. Depending upon the threshold values, the suspicious areas have been segmented. The works effection in an over segmented image. The related work was implemented using image processing tools, and using the MATLAB.
Loading....