Memristive Implementation of Variable Pixel MRI Imaging Using Skywater 130nm PDK

Improving the Magnetic Resonance Imaging (MRI) scans with on-chip hardware for diagnostic precision and image quality is an emerging research topic. The spatial sampling of MRI using a range of pixel window shapes, with a particular emphasis on intensity-based sampling, is a promising approach in this field. This research implements adaptive window-based sampling techniques with square, triangle, symmetric, and irregular shapes, enabling the sampling windows to dynamically adjust based on local intensity changes and improving feature detection accuracy in MRI scans. This approach is particularly effective in identifying and classifying white and gray matter regions, using intensity thresholds to enhance contrast and improve tissue differentiation. By sharpening the clarity of boundaries and structures, the method proves ideal for medical image analysis, especially in neuroimaging applications. The classification of brain tissue is validated through comparison with the generated pixel-sampled dataset, ensuring data reliability. Results indicate that this technique enhances intensity-based classification, improving accuracy in detection and potential advancements in MRI analysis for clinical diagnostics. The pixel sampling process is integrated using ReRAM and CMOS Skywater 130nm PDK, offering a comprehensive solution that elevates MRI analysis capabilities in medical imaging and diagnostic applications beyond algorithmic innovation.