Authors

Xinwen Liu, Jing Wang, Cheng Peng, Shekhar S. Chandra, Feng Liu, S. Kevin Zhou

Abstract

Magnetic resonance (MR) images exhibit various contrasts and appearances based on factors such as different acquisition protocols, views, manufacturers, scanning parameters, etc. This generally accessible appearance-related side information affects deep learning-based undersampled magnetic resonance imaging (MRI) reconstruction frameworks, but has been overlooked in the majority of current works. In this paper, we investigate the use of such side information as normalisation parameters in a convolutional neural network (CNN) to improve undersampled MRI reconstruction. Specifically, a Side Information-Guided Normalisation (SIGN) module, containing only few layers, is proposed to efficiently encode the side information and output the normalisation parameters. We examine the effectiveness of such a module on two popular reconstruction architectures, D5C5 and OUCR. The experimental results on both brain and knee images under various acceleration rates demonstrate that the proposed method improves on its corresponding baseline architectures with a significant margin.

Link to paper

DOI: https://link.springer.com/chapter/10.1007/978-3-031-16446-0_31

SharedIt: https://rdcu.be/cVRTq

Link to the code repository

N/A

Link to the dataset(s)

N/A

Reviews

Review #1

Please describe the contribution of the paper

This paper investigates the use of MRI acquisition protocols, views, scanning parameters/manufacturers as embedding to improve image reconstruction quality in both knee and brain datasets.
Please list the main strengths of the paper; you should write about a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.

This paper is novel and addresses an important area of research in the field to further improve image reconstruction accuracy.

The proposed formulation is sound and experiments are solid.
Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.

One major weakness is that the evalution of the methods should include comparison with other training mechanisms.
Please rate the clarity and organization of this paper

Very Good
Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance

I encourage the authors provide open access to their code.
Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://conferences.miccai.org/2022/en/REVIEWER-GUIDELINES.html

As per weakness point.
Rate the paper on a scale of 1-8, 8 being the strongest (8-5: accept; 4-1: reject). Spreading the score helps create a distribution for decision-making

7
Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

The paper is strong with novelty of the methods and comprehensive datasets for testing.
Number of papers in your stack

4
What is the ranking of this paper in your review stack?

2
Reviewer confidence

Very confident
[Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed

N/A
[Post rebuttal] Please justify your decision

N/A

Review #2

Please describe the contribution of the paper

This paper uses side information, which is normally accessible but overlooked, as the normalisation parameter to improve undersampled MRI reconstruction. The proposed SIGN encodes MRI acquisition information, attributes/imaging regions into a vector, which are mapped to a parameter space in the normalisation layers. The simple and effective design of SIGN improved performance considerably on two backbones: D5C5 and OUCR.
Please list the main strengths of the paper; you should write about a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.

(1) The idea of using side information to improve undersampled MRI reconstruction is novel and interesting. The motivation of the design is also clearly presented. (2) The proposed SIGN is simple and effective. The SIGN module only contains a few embedding and fully-connected layers. They can be inserted in the reconstruction backbones easily. (3) The improvement is significant, as demonstrated in the Table1 and Fig.5. (4) The paper is really clear and easy to follow.
Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.

The author should mention the ways of how to insert SIGN in other backbones.
Please rate the clarity and organization of this paper

Very Good
Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance

This paper uses public dataset and reports the details of implementations. It would relatively easy to reproduce the results.
Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://conferences.miccai.org/2022/en/REVIEWER-GUIDELINES.html

(1) For the OUCR backbone, the parameters of convolutional layers are shared in the recurrent design. Are the SIGN modules shared as well? (2) The author could mention a more general case of using SIGN. i.e., How to insert the SIGN modules in other backbone networks. Is it after the convolutional layer as well?
Rate the paper on a scale of 1-8, 8 being the strongest (8-5: accept; 4-1: reject). Spreading the score helps create a distribution for decision-making

7
Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

Overall it is an novel and interesting paper, with a simple but effective design leading to superior results. I would like to see it presenting in the conference.
Number of papers in your stack

5
What is the ranking of this paper in your review stack?

1
Reviewer confidence

Very confident
[Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed

N/A
[Post rebuttal] Please justify your decision

N/A

Review #3

Please describe the contribution of the paper

The authors propose a Side Information-Guided Normalisation (SIGN) module to encode the different acquisition parameters in a heterogenous dataset to generate the normalization parameters for the feature maps in the CNN reconstruction network. The SIGN module is tested in two popular CNN reconstruction networks and is demonstrated to be effective in improving the reconstruction performance.
Please list the main strengths of the paper; you should write about a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.

The SIGN module is a novel way of providing prior information to the network to improve the learning performance Solid investigation of the SIGN module by inputting the wrong side information
Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.

If the reconstruction network has enough capacity, it should be able to model the heterogeneity in the dataset Possible redundancy in the selected side information
Please rate the clarity and organization of this paper

Good
Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance

The proposed method is clearly described though the code is not made available
Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://conferences.miccai.org/2022/en/REVIEWER-GUIDELINES.html

1: The image contrast (T1, T2 or the proton-density) is determined by the key imaging parameters of TR and TE. The question is if the scanning parameters are already encoded by the SIGN, the categorial variables related to the different contrasts are still necessary? 2: It seems the simpler network of D5C5 benefit much more from the SIGN module than the OUCR network. Does it mean that if the capacity of the reconstruction network is high enough, the network itself can model the heterogeneity in the dataset without explicitly inputting the side information? 3: For the baseline methods, even without the SIGN module, is the instance normalization kept? If not, the comparisons are not fair.
Rate the paper on a scale of 1-8, 8 being the strongest (8-5: accept; 4-1: reject). Spreading the score helps create a distribution for decision-making

5
Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

If the reconstruction network has enough capacity, the benefit of adding the side information seems marginal.
Number of papers in your stack

5
What is the ranking of this paper in your review stack?

4
Reviewer confidence

Very confident
[Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed

N/A
[Post rebuttal] Please justify your decision

N/A

Primary Meta-Review

Please provide your assessment of this work, taking into account all reviews. Summarize the key strengths and weaknesses of the paper and justify your recommendation. In case you deviate from the reviewers’ recommendations, explain in detail the reasons why. In case of an invitation for rebuttal, clarify which points are important to address in the rebuttal.

All reviewers agree that this work is novel and interesting. It is suggested to justify that the method can work with larger reconstruction models and also to make the code publicly available.
What is the ranking of this paper in your stack? Use a number between 1 (best paper in your stack) and n (worst paper in your stack of n papers). If this paper is among the bottom 30% of your stack, feel free to use NR (not ranked).

1

Author Feedback

N/A

back to top

Undersampled MRI Reconstruction with Side Information-Guided Normalisation