Abstract Brain tumor (BT) involves the persistent proliferation of abnormal brain cells, often resulting in mortality. Therefore, early detection and classification are required for timely diagnosis and treatment planning. However, previous methodologies rely on analysis of manual imaging and which is a time-consuming process and leads to errors. Therefore, a sonar energy optimized distributed tetra head attention-based convolutional bidirectional network (SEnO-DTCBiNet) framework is proposed to detect and classify BTs through magnetic resonance imaging (MRI) images. The architecture leverages the precise segmentation through optimized fused attention enabled distributed W-Net based segmentation (OFA-based DW-Net) that includes tetra-head attention and SEnO algorithm to fine-tune segmented features and hyperparameters of the model. Moreover, the comprehensive multistructural feature extraction process extracts the inherent tumor characteristics, thereby enhancing the capability of the model. For the BraTS 2020 dataset, the SEnO-DTCBiNet model achieved an accuracy of 97.18%, a precision of 97.31%, a recall of 97.89%, and an F1-score of 97.26%.
Chavan et al. (Thu,) studied this question.