DL-RF: A Hybrid Deep Learning and Random Forest Approach for Enhanced Brain Tumor Detection and Classification Using MRI Images

Correct treatment strategies can be identified for patient outcome improvement in the field of neurology by improving the brain tumor detection and classification. Manual MRI analysis, however, is time-consuming and error prone; therefore, automated systems are necessary. In this study, we present a hybrid deep learning and random forest (DL-RF) methodology to improve brain tumor detection and classification from MRI images. The main aim of this research is to design an advanced hybrid model (DL-RF) for improving the accuracy and reliability of brain tumor identification and classification using MRI data. The study combines CNN based feature extraction with RF's classification strength to address the challenges of the traditional methods and obtain high performance metrics in tumor detection. This work is motivated from reducing human error, improving diagnostic accuracy and offering a faster and more reliable early tumour diagnosis method using advanced AI techniques, which utilize the DL-RF model based on convolutional neural networks (CNNs) for feature extraction to capture the complex hierarchical features along with random forest (RF) for feature selection and classification. The structure of this approach is based on data preprocessing: normalization, noise reduction, image augmentation, to produce high quality training data. In the binary classification tasks, this model outperforms traditional methods such as SVM and KNN with an accuracy of $98.7 \%$, and $\mathbf{9 4 . 5 \%}$ in multiclass classification tasks. The benefits of random forest’s robustness in classification and deep learning’s feature extraction capabilities in combination allow a more precise detection of disjoint critical features compared with other existing techniques. Both quantitative as well as qualitative evaluation of this novel brain tumor analysis method are promising and is therefore a substantial contribution to the field of brain tumor analysis.