Unlocking the Full Potential of MRI with Advanced Quantitative Imaging

The Deep Compressed Magnetic Resonance Imaging (DCQMRI) project started in March 2023, under the leadership of Mohammad Golbabaae.

We are developing innovative computational solutions to address a critical limitation in current MRI technology—its lack of quantitative, standardized measurements. While MRI is the gold standard for noninvasive, high-resolution imaging, conventional scans cannot be easily compared across hospitals or time points, limiting their diagnostic precision. Our research focuses on advancing quantitative MRI (qMRI), which provides reproducible measurements of tissue properties and has the potential to transform MRI into a precise scientific tool. However, qMRI’s long scan times have hindered its widespread clinical adoption. To overcome this, we are collaborating with world-leading partners, including GE Healthcare, University of Zurich, IRCCS Stella Maris, and University College London. Together, we are leveraging advanced deep learning methods for compressed sensing to significantly reduce scan times, making qMRI faster and more practical for routine clinical use. By enhancing both image acquisition and reconstruction processes, we aim to turn existing MRI systems into powerful tools for standardized, quantitative imaging that can be seamlessly applied across clinical sites, patients, and time points—unlocking new possibilities for precision diagnostics and patient monitoring.

This project is funded by EPSRC grant EP/X001091/1

news

Oct 10, 2024	We launch the ML-SIP website!
Oct 06, 2024	The paper “StoDIP: Efficient 3D MRF image reconstruction with deep image priors and stochastic iterations” receives a “Best Paper Award” recognition at the 15th International Workshop of Machine Learning in Medical Imaging (MLMI) at MICCAI!
Jun 10, 2024	Perla wins second place for “Best Poster Award” at Geometric Deep Learning workshop
May 04, 2024	The paper “MRI2Qmap: compressed-sampled multiparametric quantitative MRI reconstruction using learned spatial priors from multimodal MRI datasets” is selected as an Annual Meeting Program Committee (AMPC) Highlight paper (top-1%)

selected publications

Geometry of deep learning for magnetic resonance fingerprinting

Mohammad Golbabaee, Dongdong Chen, Pedro A Gómez, and 2 more authors

In ICASSP 2019-2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2019

Oral

Oral
Deep MR Fingerprinting with total-variation and low-rank subspace priors

Mohammad Golbabaee, Carolin M Pirkl, Marion I Menzel, and 2 more authors

2019

Oral

Oral

Compressive mr fingerprinting reconstruction with neural proximal gradient iterations

Dongdong Chen, Mike E Davies, and Mohammad Golbabaee

In Medical Image Computing and Computer Assisted Intervention–MICCAI 2020: 23rd International Conference, Lima, Peru, October 4–8, 2020, Proceedings, Part II 23, 2020

Oral Bib

Oral

@inproceedings{chen2020compressive,
  title = {Compressive mr fingerprinting reconstruction with neural proximal gradient iterations},
  author = {Chen, Dongdong and Davies, Mike E and Golbabaee, Mohammad},
  booktitle = {Medical Image Computing and Computer Assisted Intervention--MICCAI 2020: 23rd International Conference, Lima, Peru, October 4--8, 2020, Proceedings, Part II 23},
  pages = {13--22},
  year = {2020},
  organization = {Springer},
}

A plug-and-play approach to multiparametric quantitative MRI: image reconstruction using pre-trained deep denoisers

Ketan Fatania, Carolin M Pirkl, Marion I Menzel, and 2 more authors

In 2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI), 2022

Oral Bib

Oral

@inproceedings{fatania2022plug,
  title = {A plug-and-play approach to multiparametric quantitative MRI: image reconstruction using pre-trained deep denoisers},
  author = {Fatania, Ketan and Pirkl, Carolin M and Menzel, Marion I and Hall, Peter and Golbabaee, Mohammad},
  booktitle = {2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI)},
  pages = {1--4},
  year = {2022},
  organization = {IEEE},
}

MRI2Qmap: compressed-sampled multiparametric quantitative MRI reconstruction using learned spatial priors from multimodal MRI datasets

Mohammad Golbabaee, Matteo Cencini, Carolin Pirkl, and 3 more authors

In 2024 ISMRM & ISMRT Annual Meeting & Exhibition. Selected as an Annual Meeting Program Committee (AMPC) Highlight paper (top 1%) , 2024

Highlight Paper (top 1%) Bib

Selected as an Annual Meeting Program Committee (AMPC) Highlight paper
@inproceedings{golbabaee2024mri2qmap, title = {MRI2Qmap: compressed-sampled multiparametric quantitative MRI reconstruction using learned spatial priors from multimodal MRI datasets}, author = {Golbabaee, Mohammad and Cencini, Matteo and Pirkl, Carolin and Menzel, Marion and Tosetti, Michela and Menze, Bjoern}, booktitle = {2024 ISMRM \& ISMRT Annual Meeting \& Exhibition}, year = {2024}, }

StoDIP: Efficient 3D MRF image reconstruction with deep image priors and stochastic iterations

Perla Mayo, Matteo Cencini, Carolin M Pirkl, and 4 more authors

arXiv preprint arXiv:2408.02367, 2024

Best Paper Award Bib

Recognised as Best Paper Award at the 15th International Workshop in Machine Learning in Medical Imaging (MLMI) 2024

@article{mayo2024stodip,
  title = {StoDIP: Efficient 3D MRF image reconstruction with deep image priors and stochastic iterations},
  author = {Mayo, Perla and Cencini, Matteo and Pirkl, Carolin M and Menzel, Marion I and Tosetti, Michela and Menze, Bjoern H and Golbabaee, Mohammad},
  journal = {arXiv preprint arXiv:2408.02367},
  year = {2024},
}