Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
139 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
46 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Learning Fourier-Constrained Diffusion Bridges for MRI Reconstruction (2308.01096v3)

Published 2 Aug 2023 in eess.IV

Abstract: Deep generative models have gained recent traction in accelerated MRI reconstruction. Diffusion priors are particularly promising given their representational fidelity. Instead of the target transformation from undersampled to fully-sampled data required for MRI reconstruction, common diffusion priors are trained to learn a task-agnostic transformation from an asymptotic start-point of Gaussian noise onto the finite end-point of fully-sampled data. During inference, data-consistency projections are injected in between reverse diffusion steps to reach a compromise solution within the span of both the trained diffusion prior and the imaging operator for an accelerated MRI acquisition. Unfortunately, performance losses can occur due to the discrepancy between target and learned transformations given the asymptotic normality assumption in diffusion priors. To address this discrepancy, here we introduce a novel Fourier-constrained diffusion bridge (FDB) for MRI reconstruction that transforms between a finite start-point of moderately undersampled data and an end-point of fully-sampled data. We derive the theoretical formulation of FDB as a generalized diffusion process based on a stochastic degradation operator that performs random spatial-frequency removal. We propose an enhanced sampling algorithm with a learned correction term for soft dealiasing across reverse diffusion steps. Demonstrations on brain MRI indicate that FDB outperforms state-of-the-art methods including non-diffusion and diffusion priors.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (74)
  1. M. Lustig, D. Donoho, and J. M. Pauly, “Sparse mri: The application of compressed sensing for rapid mr imaging,” Magn. Reson. Med., vol. 58, no. 6, pp. 1182–1195, 2007.
  2. D. Liang, J. Cheng, Z. Ke, and L. Ying, “Deep magnetic resonance image reconstruction: Inverse problems meet neural networks,” IEEE Signal Process. Mag., vol. 37, no. 1, pp. 141–151, 2020.
  3. C. M. Sandino, J. Y. Cheng, F. Chen, M. Mardani, J. M. Pauly, and S. S. Vasanawala, “Compressed sensing: From research to clinical practice with deep neural networks: Shortening scan times for magnetic resonance imaging,” IEEE Sig. Process. Mag., vol. 37, no. 1, pp. 117–127, 2020.
  4. K. Hammernik et al., “Physics-driven deep learning for computational magnetic resonance imaging,” arXiv:2203.12215, 2022.
  5. F. Lam, X. Peng, and Z.-P. Liang, “High-dimensional mr spatiospectral imaging by integrating physics-based modeling and data-driven machine learning: Current progress and future directions,” IEEE Signal Process. Mag., vol. 40, no. 2, pp. 101–115, 2023.
  6. S. Wang et al., “Accelerating magnetic resonance imaging via deep learning,” in IEEE ISBI, 2016, pp. 514–517.
  7. J. Schlemper, J. Caballero, J. V. Hajnal, A. Price, and D. Rueckert, “A Deep Cascade of Convolutional Neural Networks for MR Image Reconstruction,” in IPMI, 2017, pp. 647–658.
  8. K. Hammernik et al., “Learning a variational network for reconstruction of accelerated MRI data,” Magn. Reson. Med., vol. 79, no. 6, pp. 3055–3071, 2017.
  9. B. Zhu, J. Z. Liu, B. R. Rosen, and M. S. Rosen, “Image reconstruction by domain transform manifold learning,” Nature, vol. 555, no. 7697, pp. 487–492, 2018.
  10. H. K. Aggarwal, M. P. Mani, and M. Jacob, “MoDL: Model-Based deep learning architecture for inverse problems,” IEEE Trans. Med. Imaging, vol. 38, no. 2, pp. 394–405, 2019.
  11. S. U. H. Dar, M. Özbey, A. B. Çatli, and T. Çukur, “A transfer-learning approach for accelerated MRI using deep neural networks,” Magn. Reson. Med., vol. 84, no. 2, pp. 663–685, 2020.
  12. K. Kwon, D. Kim, and H. Park, “A parallel MR imaging method using multilayer perceptron,” Med. Phys., vol. 44, no. 12, pp. 6209–6224, 2017.
  13. M. Mardani et al., “Deep generative adversarial neural networks for compressive sensing MRI,” IEEE Trans. Med. Imaging, vol. 38, no. 1, pp. 167–179, 2019.
  14. Z. Ramzi, C. G R, J.-L. Starck, and P. Ciuciu, “NC-PDNet: A Density-Compensated Unrolled Network for 2D and 3D Non-Cartesian MRI Reconstruction,” IEEE Trans. Med. Imaging, vol. 41, no. 7, pp. 1625–1638, 2022.
  15. G. Wang, T. Luo, J.-F. Nielsen, D. C. Noll, and J. A. Fessler, “B-spline parameterized joint optimization of reconstruction and k-space trajectories (bjork) for accelerated 2d mri,” IEEE Trans. Med. Imaging, vol. 41, no. 9, pp. 2318–2330, 2022.
  16. F. Liu, A. Samsonov, L. Chen, R. Kijowski, and L. Feng, “Santis: Sampling-augmented neural network with incoherent structure for MR image reconstruction,” Magn. Reson. Med., vol. 82, no. 5, pp. 1890–1904, 2019.
  17. Z. Wang et al., “One-dimensional deep low-rank and sparse network for accelerated MRI,” IEEE Trans. Med. Imaging, vol. 42, no. 1, pp. 79–90, 2023.
  18. A. Pramanik, S. Bhave, S. Sajib, S. D. Sharma, and M. Jacob, “Adapting model-based deep learning to multiple acquisition conditions: Ada-modl,” Magn. Reson. Med., 2023.
  19. K. C. Tezcan, C. F. Baumgartner, R. Luechinger, K. P. Pruessmann, and E. Konukoglu, “Mr image reconstruction using deep density priors,” IEEE Trans. Med. Imaging, vol. 38, no. 7, pp. 1633–1642, 2018.
  20. D. Narnhofer, K. Hammernik, F. Knoll, and T. Pock, “Inverse gans for accelerated mri reconstruction,” in Wavelets and Sparsity XVIII, vol. 11138.   SPIE, 2019, pp. 381–392.
  21. Y. Korkmaz, S. U. Dar, M. Yurt, M. Özbey, and T. Cukur, “Unsupervised mri reconstruction via zero-shot learned adversarial transformers,” IEEE Trans. Med. Imaging, vol. 41, no. 7, pp. 1747–1763, 2022.
  22. A. Jalal, M. Arvinte, G. Daras, E. Price, A. G. Dimakis, and J. Tamir, “Robust compressed sensing mri with deep generative priors,” in Adv. Neural Inf. Process. Sys., vol. 34, 2021, pp. 14 938–14 954.
  23. H. Chung and J. C. Ye, “Score-based diffusion models for accelerated mri,” Med. Image Anal., vol. 80, p. 102479, 2022.
  24. G. Luo, M. Heide, and M. Uecker, “Mri reconstruction via data driven markov chain with joint uncertainty estimation,” arXiv:2202.01479, 2022.
  25. Y. Xie and Q. Li, “Measurement-conditioned denoising diffusion probabilistic model for under-sampled medical image reconstruction,” arXiv:2203.03623, 2022.
  26. C. Peng, P. Guo, S. K. Zhou, V. Patel, and R. Chellappa, “Towards performant and reliable undersampled mr reconstruction via diffusion model sampling,” arXiv:2203.04292, 2022.
  27. C. Cao, Z.-X. Cui, S. Liu, H. Zheng, D. Liang, and Y. Zhu, “High-frequency space diffusion models for accelerated MRI,” arXiv:2208.05481, 2022.
  28. A. Güngör et al., “Adaptive diffusion priors for accelerated MRI reconstruction,” Med. Image Anal., vol. 88, p. 102872, 2023.
  29. H. Peng et al., “One-shot generative prior in hankel-k-space for parallel imaging reconstruction,” IEEE Trans. Med. Imaging, pp. 1–1, 2023.
  30. H. Chung, B. Sim, and J. C. Ye, “Come-closer-diffuse-faster: Accelerating conditional diffusion models for inverse problems through stochastic contraction,” in Comput. Vis. Pattern Recognit., 2022, pp. 12 413–12 422.
  31. Z.-X. Cui et al., “SPIRiT-Diffusion: Self-Consistency Driven Diffusion Model for Accelerated MRI,” 2023.
  32. B. Ozturkler, M. Mardani, A. Vahdat, J. Kautz, and J. M. Pauly, “Regularization by denoising diffusion process for MRI reconstruction,” in NeurIPS Workshop Deep Learn. Inv. Prob., 2023.
  33. A. Bansal et al., “Cold diffusion: Inverting arbitrary image transforms without noise,” arXiv:2208.09392, 2022.
  34. G. Daras, M. Delbracio, H. Talebi, A. G. Dimakis, and P. Milanfar, “Soft diffusion: Score matching for general corruptions,” arXiv:2209.05442, 2022.
  35. S. Lee, H. Chung, J. Kim, and J. C. Ye, “Progressive deblurring of diffusion models for coarse-to-fine image synthesis,” arXiv:2207.11192, 2022.
  36. J. Huang, A. Aviles-Rivero, C.-B. Schönlieb, and G. Yang, “Cdiffmr: Can we replace the gaussian noise with k-space undersampling for fast mri?” arXiv:2306.14350, 2023.
  37. G.-H. Liu, A. Vahdat, D.-A. Huang, E. A. Theodorou, W. Nie, and A. Anandkumar, “I22{}^{2}start_FLOATSUPERSCRIPT 2 end_FLOATSUPERSCRIPTSB: Image-to-Image Schrödinger Bridge,” arXiv:2302.05872, 2023.
  38. Z. Fabian, B. Tinaz, and M. Soltanolkotabi, “Diracdiffusion: Denoising and incremental reconstruction with assured data-consistency,” arXiv:2303.14353, 2023.
  39. M. Delbracio and P. Milanfar, “Inversion by direct iteration: An alternative to denoising diffusion for image restoration,” Tran. Mach. Learn. Res., 2023.
  40. H. Chung, J. Kim, and J. C. Ye, “Direct diffusion bridge using data consistency for inverse problems,” arXiv:2305.19809, 2023.
  41. C. M. Hyun, H. P. Kim, S. M. Lee, S. Lee, and J. K. Seo, “Deep learning for undersampled MRI reconstruction,” Phys. Med. Biol., vol. 63, no. 13, p. 135007, 2018.
  42. T. M. Quan, T. Nguyen-Duc, and W.-K. Jeong, “Compressed sensing MRI reconstruction with cyclic loss in generative adversarial networks,” IEEE Trans. Med. Imaging, vol. 37, no. 6, pp. 1488–1497, 2018.
  43. G. Yang et al., “DAGAN: Deep de-aliasing generative adversarial networks for fast compressed sensing MRI reconstruction,” IEEE Trans. Med. Imaging, vol. 37, no. 6, pp. 1310–1321, 2018.
  44. T. Eo, Y. Jun, T. Kim, J. Jang, H.-J. Lee, and D. Hwang, “KIKI-net: cross-domain convolutional neural networks for reconstructing undersampled magnetic resonance images,” Magn. Reson. Med., vol. 80, no. 5, pp. 2188–2201, 2018.
  45. J. Yoon et al., “Quantitative susceptibility mapping using deep neural network: QSMnet,” NeuroImage, vol. 179, pp. 199–206, 2018.
  46. Q. Liu, Q. Yang, H. Cheng, S. Wang, M. Zhang, and D. Liang, “Highly undersampled magnetic resonance imaging reconstruction using autoencoding priors,” Magn. Reson. Med., vol. 83, no. 1, pp. 322–336, 2020.
  47. S. U. Dar, M. Yurt, M. Shahdloo, M. E. Ildız, B. Tınaz, and T. Çukur, “Prior-guided image reconstruction for accelerated multi-contrast mri via generative adversarial networks,” IEEE J. Sel. Top. Signal Process., vol. 14, no. 6, pp. 1072–1087, 2020.
  48. P. Guo, Y. Mei, J. Zhou, S. Jiang, and V. M. Patel, “Reconformer: Accelerated mri reconstruction using recurrent transformer,” arXiv:2201.09376, 2022.
  49. B. Yaman, S. A. H. Hosseini, and M. Akcakaya, “Zero-shot physics-guided deep learning for subject-specific mri reconstruction,” in NeurIPS Workshop Deep Learn. Inv. Prob., 2021.
  50. G. Luo, N. Zhao, W. Jiang, E. S. Hui, and P. Cao, “Mri reconstruction using deep bayesian estimation,” Magn. Reson. Med., vol. 84, no. 4, pp. 2246–2261, 2020.
  51. M. Z. Darestani and R. Heckel, “Accelerated mri with un-trained neural networks,” IEEE Trans. Comput. Imaging, vol. 7, pp. 724–733, 2021.
  52. K. P. Slavkova et al., “An untrained deep learning method for reconstructing dynamic mr images from accelerated model-based data,” Magn. Reson. Med., vol. 89, no. 4, pp. 1617–1633, 2023.
  53. G. Elmas et al., “Federated learning of generative image priors for mri reconstruction,” IEEE Trans. Med. Imaging, vol. 42, no. 7, pp. 1996–2009, 2023.
  54. J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” Adv. Neural Inf. Process. Sys., vol. 33, pp. 6840–6851, 2020.
  55. Y. Song, L. Shen, L. Xing, and S. Ermon, “Solving inverse problems in medical imaging with score-based generative models,” in Int. Conf. Learn. Rep., 2022.
  56. Y. Balaji et al., “ediffi: Text-to-image diffusion models with an ensemble of expert denoisers,” arXiv:2211.01324, 2022.
  57. A. Q. Nichol and P. Dhariwal, “Improved denoising diffusion probabilistic models,” in Int. Conf. Mach. Learn., 2021, pp. 8162–8171.
  58. D. L. Ruderman and W. Bialek, “Statistics of natural images: Scaling in the woods,” Phys. Rev. Lett., vol. 73, pp. 814–817, 1994.
  59. F. Knoll et al., “fastMRI: A publicly available raw k-space and DICOM dataset of knee images for accelerated MR image reconstruction using machine learning,” Rad. Artif. Intell., vol. 2, no. 1, p. e190007, 2020.
  60. T. Zhang, J. M. Pauly, S. S. Vasanawala, and M. Lustig, “Coil compression for accelerated imaging with cartesian sampling,” Magn. Reson. Med., vol. 69, no. 2, pp. 571–582, 2013.
  61. M. Uecker et al., “Espirit—an eigenvalue approach to autocalibrating parallel mri: where sense meets grappa,” Magn. Reson. Med., vol. 71, no. 3, pp. 990–1001, 2014.
  62. J. P. Haldar and J. Zhuo, “P-loraks: low-rank modeling of local k-space neighborhoods with parallel imaging data,” Magn. Reson. Med., vol. 75, no. 4, pp. 1499–1514, 2016.
  63. H. A. Bedel and T. Çukur, “DreaMR: Diffusion-driven counterfactual explanation for functional MRI,” arXiv:2307.09547, 2023.
  64. O. Dalmaz et al., “One model to unite them all: Personalized federated learning of multi-contrast mri synthesis,” arXiv:2207.06509, 2022.
  65. L. Xiang et al., “Deep-learning-based multi-modal fusion for fast MR reconstruction,” IEEE Trans. Biomed. Eng., vol. 66, no. 7, pp. 2105–2114, 2019.
  66. B. Yaman, S. A. H. Hosseini, S. Moeller, J. Ellermann, K. Uğurbil, and M. Akçakaya, “Self-supervised learning of physics-guided reconstruction neural networks without fully sampled reference data,” Magn. Reson. Med., vol. 84, no. 6, pp. 3172–3191, 2020.
  67. J. Liu, Y. Sun, C. Eldeniz, W. Gan, H. An, and U. S. Kamilov, “Rare: Image reconstruction using deep priors learned without groundtruth,” IEEE J. Sel. Top. Signal Process., vol. 14, no. 6, pp. 1088–1099, 2020.
  68. Y. Korkmaz, T. Cukur, and V. Patel, “Self-supervised MRI reconstruction with unrolled diffusion models,” arXiv:2306.16654, 2023.
  69. Z.-X. Cui et al., “Self-Score: Self-Supervised Learning on Score-Based Models for MRI Reconstruction,” arXiv:2209.00835, 2022.
  70. A. Aali, M. Arvinte, S. Kumar, and J. I. Tamir, “Solving inverse problems with score-based generative priors learned from noisy data,” arXiv:2305.01166, 2023.
  71. T. Xiang, M. Yurt, A. B. Syed, K. Setsompop, and A. Chaudhari, “Ddm22{}^{2}start_FLOATSUPERSCRIPT 2 end_FLOATSUPERSCRIPT: Self-supervised diffusion mri denoising with generative diffusion models,” 2023.
  72. M. Özbey et al., “Unsupervised medical image translation with adversarial diffusion models,” IEEE Trans. Med. Imaging, 2023.
  73. C.-M. Feng et al., “Multimodal transformer for accelerated mr imaging,” IEEE Trans. Med. Imaging, vol. 42, no. 10, pp. 2804–2816, 2023.
  74. O. Dalmaz, M. Yurt, and T. Çukur, “ResViT: Residual vision transformers for multi-modal medical image synthesis,” IEEE Trans. Med. Imaging, vol. 41, no. 10, pp. 2598–2614, 2022.
Citations (11)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now