An Intelligent Approach for Automatic Brain Tumor Detection
Manojkumar. S. Kathane, Vilas M. Thakare · 2014
In this paper we are comparing our proposed intelligent approach using Genetic Algorithm (GA) for automatic brain tumor detection in MRI brain images with the existing intelligent approaches. This uses Relative Localization Weighted F-measure (RLWF) (2), which attempts to acknowledge the worth of individual localizations made by the genetic program. The fitness function is based on precision and recall which always guaranty the best results. We use the images in the MICCAI BRAT 2012 database (3), a well-known and recent public dataset, to do our experiments. In this database, there are 30 subjects MRI brain images with T1, T1 contrast, T2 and FLAIR parameters. On observing results, it can be seen that GA based tumor segmentation algorithm is capable of detecting the tumor points.