Fig. 3
From: Evaluation of deep learning MRI reconstruction for dental implant crowns in a phantom study
Schematic representation of the study. (A) Fabrication of the four crown types using a CAD-CAM system. (B) Phantom preparation containing the implant and crown embedded in a 1.5% agar solution. The crown margin is positioned at the center ( +). (C) Post-MRI acquisition image analysis on the axial slice showing the largest crown area. Image quality (noise and SNR) and visual score of metal artifact were assessed using across the entire image, and metal artifacts were analyzed within the central 5 × 5 cm square ROI (red box). Signal intensity for binarization was measured in four 1 × 1 cm ROIs (blue boxes) placed immediately outside the four corners of the central ROI. After binarization, metal artifacts appeared as black areas, and their area was automatically calculated as a percentage of the total ROI. PMMA, polymethyl methacrylate; Ni–Cr, nickel-chrome; SNR, signal-to-noise ratio; ROI, region of interest.
