Brain Tumor Detection and Classification Using PSO and Convolutional Neural Network

Tumor detection has been an active research topic in recent years due to the high mortality rate. Computer vision (CV) and image processing techniques have recently become popular for detecting tumors MRI images. The automated process is simpler takes less time than manual processing. In addition, difference expanding shape of brain tumor tissues complicates clinicians. We proposed a new framework as well classification into relevant categories this paper. For segmentation, employs Particle Swarm Optimization (PSO) algorithm, classification, convolutional neural network (CNN) algorithm. Popular preprocessing such noise removal, sharpening, skull stripping are used at start segmentation process. Then, PSO-based applied. step, two pre-trained CNN models, alexnet inception-V3, trained using transfer learning. Using serial approach, features extracted from both models fused final classification. variety machine learning classifiers used. Average dice values on datasets BRATS-2018 BRATS-2017 98.11 percent 98.25 percent, respectively, whereas average jaccard 96.30 96.57% (Segmentation Results). results were extended same achieved 99.0% accuracy, sensitivity 0.99, specificity precision 0.99. Finally, method compared state-of-the-art existing methods outperforms them.