Novel automatic tumor extraction method based on decision Tree Classifier

Early detection of a brain tumor is much easier to cure than to find out at an advanced stage. In this article we present a tumor extraction method based on the decision tree classifier for early and rapid detection. This method can also provide the tumor characteristics, the exact location and it geometric properties. Our method will be applied to a huge number of MRI images to prove its accuracy, and speed. This method is carried out in three stages: the first step is filtering the image using the Nlmeans filter. The second step is the brain extraction of the skull, to retain only the brain, in order to avoid the influence of another substance on the detection of the tumor, and at the last step the tumor area extraction of with the detection tree classifier. KeywordBrain tumor, MRI, decision tree classifier, Nl-means filter, Dice, Jaccard.