Special Issue in Medical Image Analysis

We are happy that Medical Image Analysis, a top journal in the field, will publish a Special Issue based on contributions to MIDL.

The best accepted full conference papers will be invited for submission to this special issue. Authors will be invited after acceptance of their paper for the conference. They will be requested to adapt their paper for Medical Image Analysis based on comments by the Special Issue Guest Editors and reviewers.

Timeline for the Special Issue

The papers will be prepared following the Elsevier guidelines and submitted online via the submission system (EES). Detailed instructions for authors are available on the journal website.

Review process for the Special Issue

We choose an open review process for MIDL, and in the same spirit, we would like to openly communicate on the choices we made for selecting and reviewing papers for the Special Issue. We do this on this page, which is updated continuously during the process.

The Editors of MedIA asked us to select the very best papers only and would like that the authors augment their content significantly from the conference version. As Guest Editors, we should apply the regular and thorough reviewing process of MedIA before acceptance or reject.

We had to indicate how many papers we expected to be in the special issue. This was a hard call, as we are organizing this conference for the first time, so we made a conservative estimation of 10 papers. We were positively surprised by the large number of papers submitted to MIDL and by the excellent quality of many papers.

We consider only full papers eligible for the special issue. We asked the reviewers (PC members and others reviewers recruited by the PC members) to indicate with their reviews if they believed the paper should be eligible for the special issue. We started the selection process by excluding all papers that were not recommended by any reviewer, and papers that were not accepted to MIDL. This left us with a large number of papers. It was obvious we needed to be highly selective. We have therefore decided to remove papers that

1. made only a small contribution (something that is fine for a conference paper but not for a journal publication);
2. had substantial and in our view justified criticism from the reviewers (usually the ones that did not nominate the paper for the special issue);
3. still required a lot of additional work and possibly substantial changes to the method in order to warrant publication in MedIA.

Especially because of the latter criterion, we are convinced that we have removed some very promising papers that will find their way to journals eventually.

With these criteria, we have compiled a longlist of 25 papers for the special issue.

Mid May we have notified the authors whose papers have been selected via a comment on the OpenReview website. We ask them to answer three questions:

1. To indicate if they are interested to be eligible for the special issue. This would mean they are willing to augment the content of their paper significantly and submit the full manuscript before August 1, 2018.
2. To confirm that the paper, or any paper with overlap with the contents of the MIDL paper, is and will not be under review or under consideration elsewhere, following the Elsevier rules on “Multiple, redundant or concurrent publication”.
3. To send us any related publications of the author group, for example, other papers that apply a similar methodological idea to another data set, to help reviewers judge the novel contribution of the work.

For all papers for which we received a positive reply, we asked one or two new reviewers and one or more of the MIDL reviewers to advise us if they believe an augmented version of the paper should be considered for the special issue and recommend what additional material to include. Please note the MedIA papers have on average twice as many words as the full papers submitted to MIDL.

After discussions with the Editors of MedIA, we received an upper limit for the number of papers we could invite to submit, and we made a shortlist of 14 papers. We provided fairly detailed advice on what should be changed and added to the MIDL paper. We hope that these recommended changes will work as a ‘major revision’ review.