{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/convolutional-recurrent-neural-networks-for-7","title":"Convolutional Recurrent Neural Networks for Dynamic MR Image Reconstruction","arxiv_id":"1712.01751","date":"2017-12-05","proceeding":null,"authors":["Chen Qin","Jo Schlemper","Jose Caballero","Anthony Price","Joseph V. Hajnal","Daniel Rueckert"],"abstract":"Accelerating the data acquisition of dynamic magnetic resonance imaging (MRI)\nleads to a challenging ill-posed inverse problem, which has received great\ninterest from both the signal processing and machine learning community over\nthe last decades. The key ingredient to the problem is how to exploit the\ntemporal correlation of the MR sequence to resolve the aliasing artefact.\nTraditionally, such observation led to a formulation of a non-convex\noptimisation problem, which were solved using iterative algorithms. Recently,\nhowever, deep learning based-approaches have gained significant popularity due\nto its ability to solve general inversion problems. In this work, we propose a\nunique, novel convolutional recurrent neural network (CRNN) architecture which\nreconstructs high quality cardiac MR images from highly undersampled k-space\ndata by jointly exploiting the dependencies of the temporal sequences as well\nas the iterative nature of the traditional optimisation algorithms. In\nparticular, the proposed architecture embeds the structure of the traditional\niterative algorithms, efficiently modelling the recurrence of the iterative\nreconstruction stages by using recurrent hidden connections over such\niterations. In addition, spatiotemporal dependencies are simultaneously learnt\nby exploiting bidirectional recurrent hidden connections across time sequences.\nThe proposed algorithm is able to learn both the temporal dependency and the\niterative reconstruction process effectively with only a very small number of\nparameters, while outperforming current MR reconstruction methods in terms of\ncomputational complexity, reconstruction accuracy and speed.","url_abs":"http://arxiv.org/abs/1712.01751v3","url_pdf":"http://arxiv.org/pdf/1712.01751v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"convolutional-recurrent-neural-networks-for-7","repo_url":"https://github.com/cjandrioli/Deep-MRI-Reconstruction","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"convolutional-recurrent-neural-networks-for-7","repo_url":"https://github.com/js3611/Deep-MRI-Reconstruction","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"convolutional-recurrent-neural-networks-for-7","repo_url":"https://github.com/myyaqubpython/https-github.com-cq615-Deep-MRI-Reconstruction","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"convolutional-recurrent-neural-networks-for-7","repo_url":"https://github.com/sainzmac/Deep-MRI-Reconstruction-master","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"image-reconstruction","task_name":"Image Reconstruction"},{"task_slug":"temporal-sequences","task_name":"Temporal Sequences"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1712.01751","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}