International Journal of Computer Science Issues

More than a traditional journal...

Wavelets have enjoyed a widespread exposure in applications of image processing and computer vision. So mush so that wavelet is widely used in medical applications as the computer aided detection of microcalcifications in mammograms. A several types of wavelet transforms were employed in algorithms to achieve automated detection of microcalcifications. In this work we present a comparative study of wavelets to pick the better one and its optimal potential level of decomposition that give us better detection. Our algorithm involves four steps: First, delimitation of the Region of Interest. Second, extraction of microcalcifications profiles. Next, a 1-D DWT with different families of wavelet is applied on the signal up to the sixth level. Finally, comparison between details coefficients of each level is done to carry out the optimal level. To prove our result, 2-D wavelet transform decomposition and reconstruction with the list of wavelets used above and up to the optimal level of each one is applied on digital mammograms from the MIAS data base (Mini Database for Screening Mammography) to carry out microcalcifications. A comparative study based on the true positive (TP) is performed to confirm the result.