Glioma Detection using Improved Artificial Neural Network in MRI Images

The physical identification of tumors may be a laborious and time-consuming process for medical professionals because of the complex nature of the tumor and the noise involution that can occur in magnetic resonance (MR) imaging information. Therefore, determining the location of the tumor at an earlier stage is quite important. The medical scan may track and prognosticate the uncontrolled proliferation of cancer pretentious regions at different levels in order to deliver a felicitous diagnosis at an early time. This is accomplished via the utilisation of segmentation in conjunction with relegation procedures. In order to recognize the tissues of a brain tumor, segmentation of the picture obtained from the MRI is a crucial and challenging step. So, the proposed work includes the tumor segmentation process using Region Growth Algorithm with Gray-Level-Run-Length-Matrix and Centre-Symmetric-Local-Binary-Patterns texture feature extraction process. The segmented images undergo feature extraction process with higher level of accuracy. The performance metrics are measured using accuracy, sensitivity and specificity. The proposed work has 0.97% sensitivity, 0.85% specificity and 99.80% accuracy.