Deep Unfolded BM3D: Unrolling Non-local Collaborative Filtering into a Trainable Neural Network
This is an incremental improvement for medical imaging denoising, enhancing performance in low-dose CT scenarios.
The paper tackled the problem of denoising in low-dose CT by proposing Deep Unfolded BM3D, a hybrid framework that unrolls BM3D into a trainable neural network, resulting in higher PSNR and SSIM compared to classic BM3D and U-Net, particularly in high-noise conditions.
Block-Matching and 3D Filtering (BM3D) exploits non-local self-similarity priors for denoising but relies on fixed parameters. Deep models such as U-Net are more flexible but often lack interpretability and fail to generalize across noise regimes. In this study, we propose Deep Unfolded BM3D (DU-BM3D), a hybrid framework that unrolls BM3D into a trainable architecture by replacing its fixed collaborative filtering with a learnable U-Net denoiser. This preserves BM3D's non-local structural prior while enabling end-to-end optimization. We evaluate DU-BM3D on low-dose CT (LDCT) denoising and show that it outperforms classic BM3D and standalone U-Net across simulated LDCT at different noise levels, yielding higher PSNR and SSIM, especially in high-noise conditions.