Synergistic Reconstruction and Synthesis via Generative Adversarial Networks for Accelerated Multi-Contrast MRI

• DOI: 10.48550/arxiv.1805.10704
• Publication Date: 2018-01-01
• Authors: Salman Ul Hassan Dar; Mahmut Yurt; Mo Shahdloo; M. Emrullah Ildiz; Tolga Çukur
• Keywords: Generative adversarial network; Limiting

Multi-contrast MRI acquisitions of an anatomy enrich the magnitude information available for diagnosis. Yet, excessive scan times associated with additional contrasts may be a limiting factor. Two mainstream approaches enhanced efficiency are reconstruction undersampled and synthesis missing acquisitions. In reconstruction, performance decreases towards higher acceleration factors diminished sampling density particularly at high-spatial-frequencies. synthesis, absence data samples from target contrast can lead to artefactual sensitivity or insensitivity image features. Here we propose new approach synergistic reconstruction-synthesis multi-contrast based on conditional generative adversarial networks. The proposed method preserves high-frequency details by relying shared source contrast, prevents feature leakage loss contrast. Demonstrations brain datasets healthy subjects patients indicate superior compared previous state-of-the-art. help improve quality exams.