A new fast brain tumor extraction method based on Nl-means and expectation maximization

The scientific development has allowed researchers to touch on all areas of research, even the medicine field that has long been dominated by the doctors. The objective of our work is to help Radiologists in the treatment of brain MRI images, by detecting locating and giving the exact characteristics of a tumor areas. This is to cope with the very large number of images that have to be processed every day. Thus we can reduce the treatment time and the patients waiting time. The burden has become less heavy with the step automation. In this article we present a new fast and totally automatic tumor detection and extraction method, our work is divided into two parts: In the first time we will filter the image by the Nl-means filter to reduce the confusion risk and to soften the image. And in the second step we will classify the MRI into two classes (tumor classes and no tumor classes), this step will be done by the expectation maximization algorithm (EM). At the last time and to prove our method effectiveness a multiple evaluation criteria will be calculate, to compare our method to seven tumor extraction method which are the frequently extraction methods used in the literature and the tree last extraction methods proposed recently.