Comparative Analysis of VGG19 and ResNet50 V2 for Enhanced Brain Tumor Detection

This paper compares the performance of two most common deep learning models ResNet50 V2 and VGG19 when using medical imaging for brain tumour detection. Even though early and accurate diagnosis is crucial to improving the treatment and survival of patients, traditional diagnostic methods are ineffective. The purposes of this paper are, first, to evaluate the performance and capabilities of VGG19 and ResNet50 V2 with regard to current achievements in convolutional neural networks using those for brain tumour detection, and second, to compare the accuracy of the results of both models. Based on the accuracy, precision, recall, and F1-scores of both models, the study obtained a general accuracy of 60%. The obtained results demonstrate marked trade-offs between the accuracy and recall of the two models, showing their major advantages and drawbacks. The conclusion of the paper deals with the possible implications of the study for practice and further research aimed to increase the efficiency and applicability of AI in medical diagnostics. Such studies may explore hybrid models, expansion of data augmentation, and the role of interpretability in the support for medical experts.