Fat-saturated image generation from multi-contrast MRIs using generative adversarial networks with Bloch equation-based autoencoder regularization

• Published in: Medical image analysis Vol. 73; p. 102198
• Main Authors: Kim, Sewon, Jang, Hanbyol, Hong, Seokjun, Hong, Yeong Sang, Bae, Won C., Kim, Sungjun, Hwang, Dosik
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.10.2021; Elsevier BV
• Subjects: Autoencoder regularization; Bloch equation; Coders; Feature extraction; Geneartive adversarial networks; Generative adversarial networks; Image contrast; Image processing; Image synthesis; Magnetic resonance; Magnetic resonance image; Magnetic resonance imaging; Medical imaging; Multi-contrast imaging; Regularization; Spine; Image synthesis; Autoencoder regularization; 41A10; 65D05; Bloch equation; 65D17; Multi-contrast imaging; Geneartive adversarial networks; 41A05; Magnetic resonance image

•Bloch equation-based autoencoder regularization GAN (BlochGAN).•BlochGAN uses multi-contrast MR images to generate other contrast images.•BlochGAN can learn the multi-contrast MR relationship based on the Bloch equation.•Performance of BlochGAN is quantitatively and qualitatively demonstrated.•Bloc...