Comparative Research Analysis of Deep Learning Models for the Detection of Brain Stroke

Nowadays, early-stage prediction is important for optimizing neurological damage and improving patient recovery. Precisely classifying stroke in CT scan imaging remains challenging due to the difficulties of stroke-based brain abnormalities in medical imaging. Rapid and accurate classification requires advanced diagnostic techniques for reliable outcomes. The comparative research article defined an evaluation of the implemented ERDP-CNN method. It compares the different performance metrics with existing methods, like Logistic Regression (LR), Support Vector Machine (SVM), K-nearest neighbor (KNN), Decision Tree (DT), OzNet, and so on, in the area of stroke detection. The comparative result analysis demonstrates that the ERDPCNN model outperforms other methods with different performance metrics, such as Accuracy (Acc) at 98.9%, Precision (Pre) at 97.03 %, Sensitivity (SN) at 95.56 %, and F1-Score at 99.98 %. An advanced feature extraction and selection method, the method efficiently optimizes dimensionality and reduces the error rates, ensuring a reliable MATLAB simulation tool for early diagnosis. The results highlight the potential of the ERDP-CNN method to transform medical decision-making in stroke detection in CT brain images, with promising uses in other domains of medical imaging.