The early stage of brain tumor detection saves millions of lives worldwide. The variability of cell growth lowers the detection ratio of brain tumors. Recent research focuses on the texture property of MRI imaging and how to detect brain tumors at an early stage. For the purpose of detecting brain tumors, this paper suggests a hybrid texture feature optimization approach. Particle swarm optimization and genetic algorithms are two swarm intelligence methods that are combined in the MKSVM algorithm. The MKSVM algorithms classify the low-intensity features of texture and detect the area of brain tumor in MRI images. The extraction of texture features is an important phase of processing. In order to extract texture features, we apply discrete wavelet transform (DWT) methods. These methods are well known as texture feature extraction approaches instead of other transform methods. The BRATS dataset 2018 is used to test and evaluate the proposed technique using MATLAB software. Various existing algorithms, including CNN and DCNN, are used to compare the performance of the proposed method. The effectiveness findings indicate that the suggested method is more effective than current algorithms for brain tumor identification

Brain Tumor, MRI Image, DWT, PSO, GA, SVM, CNN, DCNN