A brain tumor is a tissue group formed by the addition of unusual cells in the brain, and it's significant to identify brain tumor through Magnetic Resonance Imaging (MRI) for treatment. Because of the different t...
详细信息
A brain tumor is a tissue group formed by the addition of unusual cells in the brain, and it's significant to identify brain tumor through Magnetic Resonance Imaging (MRI) for treatment. Because of the different tumor types, better segmentation of brain tumor is a tricky problem in MRI. To conquer the existing issues, the latest scheme is proposed for brain tumor classification and segmentation by deep structured architectures. In the first stage, the source images are garnered from online data sources. The garnered data is then subjected to a new segmentation model which is termed as Multiscale Atrous Convolution-based Adaptive ResUNet3+ (MAC-ARUNet3 + ). Here the parameters of deep learning approaches are optimally tuned by improved math optimizer accelerated-based Arithmetic Optimization Algorithm (IMOA-AOA). Finally, the tumor classification is carried out by Attentionbased Ensemble Convolution Networks (AECN), in which it is encompassed with ResNet, Inception, and MobileNet. Thus, the final results are determined by the high-ranking-based estimation. The performance and comparative analysis are done using diverse measures and distinct algorithms. Hence, the recommended model outperforms with the high segmentation and classification accuracy that aids the practitioner in diagnosing the disease rapidly.
暂无评论