AbstractBackground: Magnetic Resonance Imaging (MRI) is a widely used diagnostic tool that provides detailed images of soft tissues in the body. However, MRI scans are often time-consuming and susceptible to motion artifacts, which can reduce image quality. Recent advancements in deep learning algorithms offer the potential to significantly improve MRI image reconstruction quality by reducing scan time and enhancing image resolution and clarity.
Objective: To review and analyze the current state of research on the application of deep learning algorithms in MRI image reconstruction and their impact on image quality.
Methods: A systematic review of literature published between 2015 and 2023 was conducted using databases such as PubMed, IEEE Xplore, and Google Scholar. Studies were included if they evaluated deep learning-based approaches for MRI image reconstruction, including methods like Convolutional Neural Networks (CNNs), Generative Adversarial Networks (GANs), and Recurrent Neural Networks (RNNs). The quality of studies was assessed using the PRISMA guidelines.
Results: A total of 85 studies met the inclusion criteria. The most commonly used deep learning models were CNNs and GANs, which demonstrated significant improvements in image quality, noise reduction, and artifact suppression. These methods reduced reconstruction time by up to 50% compared to traditional techniques and improved image resolution, enabling more accurate diagnosis.
Conclusion: Deep learning algorithms have shown great promise in enhancing MRI image reconstruction quality. Future research should focus on optimizing these models for clinical use, ensuring robustness, and minimizing potential biases in reconstructed images.
