Artificial intelligence algorithms in the digital processing of biomedical images

Abstract

Artificial intelligence has revolutionized the way we process and analyze data, including biomedical images. The present thesis investigates the application of convolutional neural networks (CNNs) for the classification of MRI brain scans, a crucial task in the medical diagnosis profession from brain images. The thesis is experimental; it comparatively evaluates the performance of two state-of-the-art CNN architectures, EfficientNetB0 and ResNet18, on a dataset of MRI brain scans. The results show that the EfficientNetB0 architecture outperforms the ResNet18 architecture in terms of accuracy and efficiency, demonstrating the potential of machine learning techniques for biomedical classification of images. The findings of this study provide useful insights into the capabilities and limitations of artificial intelligence algorithms for the digital processing of biomedical images. The results demonstrate the potential of machine learning methods to enhance the precision and efficiency of automated biomedical image classification, which has important implications for the early detection and diagnosis of brain damages. Additionally, the study highlights the importance of selecting the appropriate CNN architectures for the given task, as well as the need for large and diversified datasets to improve the generalizability of the results.