A Comparative Study of State-of-the-Art Deep Neural Network on High-Resolution Medical Images

The growing accessibility and storage of high-resolution chest X-ray images in contemporary hospitals have generated unparalleled prospects for data-driven decision-making in healthcare. Nonetheless, the efficient application of deep learning models for precise diagnosis is hindered by the complex features and fine details inherent in these images. This study aims to tackle these problems by deploying and evaluating three sophisticated deep neural networks-Inception-ResNetV2, ResNet152, and DenseNet201-utilizing a meticulously selected dataset of chest X-ray pictures. The efficacy of neural networks is evaluated by many metrics, including accuracy, precision, recall, Fl-score, and AUC-ROC, which illustrate their proficiency in picture classification tasks with high precision and accuracy. This study's findings yield significant insights into the strengths and shortcomings of each network, presenting possible opportunities for improving high-resolution image classification. Ultimately, these enhancements may provide more efficient and prompt diagnosis in healthcare environments.