MELBA (The Journal of Machine Learning for Biomedical Imaging) encourages the submission of manuscripts on the general topic of “generative models for biomedical imaging and image analysis”. MELBA is a web-based journal devoted to the free and unrestricted access of high quality articles in the broad field that bridges machine learning and biomedical imaging.
In recent years, there has been a flurry of developments in machine learning (including Variational Auto-Encoders or VAEs, Generative Adversarial Networks or GANs, Normalizing Flows or NFs, and lately, Diffusion Models) that enable us to generate high-quality, realistic synthetic data such as high-dimensional images, volumes, or tensors.
These techniques have numerous applications in medical imaging and analysis, including: educational purposes, dataset augmentation and imputation, correction of dataset biases, image reconstruction and synthesis problems. The Special Issue aims to collect new contributions in this general area.
Topics of interest can include, but are not limited to:
new generative modeling methods that are motivated by or applicable to known problems in biomedical imaging, with strong theoretical justification and appropriate empirical evaluation;
empirical accounts that thoroughly evaluate, and/or compare existing generative modeling methods for biomedical imaging applications of interest; and
presentation of a novel biomedical imaging application, where existing generative models offer a key solution.
We kindly invite researchers to contribute their high-quality original articles on these topics to our Special Issue. We will be considering submissions between now and July 1, 2023. When submitting your manuscript, the authors should use the cover letter to indicate their intention to be considered in the special issue. The editorial team will aim to streamline the reviews and return a decision within 4-6 weeks from submission. Final accepted papers will be highlighted in a special issue that will be collectively published in the fall of 2023.
Information Processing in Medical Imaging (IPMI) is one of the longest-running conference series in medical imaging, founded in 1969. The conference has a number of traditions, ranging from unrestricted discussion time, via reading groups and a conference choir, to a soccer game. Due to the COVID-19 pandemic, the 2021 conference was held virtually, still holding on to as many traditions as possible. The conference had 150 attendees, and featured both the traditional reading groups and a far less traditional virtual conference venue to recreate the community feeling known from physical IPMI conferences as much as possible (including the soccer game!).
IPMI 2021 received 200 valid submissions, of which 59 were accepted for publication at the conference. From these, 29 papers were invited to submit an extended journal version to the first special edition of the MELBA journal associated with an IPMI conference.
Out of the invited papers, we received 11 submissions, which all underwent a new peer-review process. All 11 papers were accepted for the final special issue. These papers cover the topics discussed at the conference well, ranging from uncertainty estimation, via learning hyperparameter tuning, to designing and utilizing geometric priors. We are excited and thankful to present these papers in openly available form to the community through MELBA, including the recorded conference talks, which can be found from the conference website.
The guest editors,
Aasa Feragen, Technical University of Denmark;
Mads Nielsen, University of Copenhagen;
Stefan Sommer, University of Copenhagen;
Julia Schnabel, Helmholtz Munich,Technical University of Munich, and King's College London.
RSS (Really Simple Syndication) is a simple and convenient way to receive automatically updates and news from a website, without having to check it manually. Many e-mail clients (including Thunderbird and Outlook) can support it out-of-the-box, and many alternative RSS clients exist.
The MELBA Journal is excited to announce a new initiative that aims to increase engagement between authors and readers, and spotlight high quality papers.
We plan to have an online (virtual) symposium, multiple times a year, where two or three papers that were previously published in MELBA, will be presented by the authors and interactively discussed in detail. Each paper will have a dedicated reading group that might include editor(s) and reviewer(s) involved in the evaluation of the paper and other expert researchers in the field. This group will lead the live discussion with the author(s) of the paper during the symposium, where audience members will be welcome to participate too.
The choice of papers will be made based on feedback from readers, reviewers, and editors. The timing of the symposium and selected papers will be posted later.
We are excited to unveil our new, in-house MELBA website, which will enable more information and context about each published paper. Authors are invited to provide extra links, videos, and presentation as companion documents to the PDF. The publishing editor can be contacted to add or modify this extra content, even after publication.
The submission and reviewing process is still hosted on Scholastica.
We are pleased to announce the first MELBA special issue on selected papers from the Medical Imaging with Deep Learning (MIDL) conference held virtually, also for the first time, in Montreal from July 6 to 9, 2020. The annual MIDL conference attracts world-class researchers, engineers, as well as clinicians, who develop novel algorithms to solve medical imaging problems using deep learning. The virtual conference had 3,575 registered participants. Two submission tracks were received through the OpenReview system, 148 full papers and 106 short papers. The conference selection had 65 full papers and 40 short papers. A smaller subset of 20 papers were invited to submit an extended journal version to this MELBA special issue, based on careful examination of the submission, AC, reviewer and rebuttal comments. These submissions have gone through a new thorough review process. The final selection of 8 papers covers various topics, including learning under uncertainty, image generation and reconstruction, now presented in this MELBA special issue. All papers can be found on the MELBA website: https://new.melba-journal.org/issues/midl20.html
We are pleased to announce that MELBA has published its first article: “An Uncertainty-Driven GCN Refinement Strategy for Organ Segmentation”. This paper presents an extension of the authors’ MIDL 2020 paper and is included in the MIDL 2020 Special Issue.