One Shared LoRA Weight for MRI Reconstruction across Acceleration Factors
Shared LoRA is proposed, a parameter-efficient framework that freezes the pretrained SHFormer backbone and trains a single shared set of LoRA adapters together with a lightweight gating network that achieves the best or competitive PSNR and SSIM across acceleration factors.
Zhi-Wei Zhao, Wei-Kang Gong, Zhongnian Li et al.
· 0 citations