Voxel based morphometry of the human brain imaging in improved convolution neural network

The accurate detection of brain tumors in humans still remains an issue. Though several works have been proposed to detect brain tumors, only a few of them consider the voxel-based morphometry (VBM) process during classification. VBM is used to detect the edges of the tumor regions by, which helps in the accurate evaluation of brain tumors. Thus, this work evaluates the combination of VBM and Convolutional Neural Network (CNN) and its feasibility for brain tumor classification. VBM detects the volume of gray matter and white matter by detecting the tumor edges. Then these regions’ volumes were calculated and used as the improved CNN features. The improved CNN model was trained and tested with ten-fold cross-validation. Also, data augmentation is done to improve the dataset size. Several comparison methods are used to test the efficiency of the proposed method. The proposed VBM and improved CNN have an area under the receiver operating characteristic curve (AUC). Thus, the proposed method with VBM and improved CNN was applicable in the brain tumor classification with high accuracy. The proposed method’s performance is identified by using different parameters such as sensitivity, precision, specificity, accuracy, F1 score and negative prediction value on the dataset from 10-time repeated tenfold cross-validation. Here, the proposed method shows a higher accuracy of more than 95% when compared to existing methods. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024.