Disentangled latent energy-based style translation: An image-level structural MRI harmonization framework

Mengqi Wu, Lintao Zhang, Pew-Thian Yap, Hongtu Zhu, Mingxia Liu

Abstract: Highlights•A novel disentangled latent energy-based model for unpaired MRI harmonization.•MRI style translation in a low-dimensional latent space with less computation cost.•Strong generalizability on new MRI domains without model retraining.•Synthesis of brain MRIs matching a given domain’s intensity distribution.•Extensive experimental validation on two datasets with 4,000+ subjects in 4 tasks.