Description & Relevance
Big and complex data is fuelling diverse research directions in both medical image analysis and computer vision research fields. These can be divided into two main categories: (1) analytical methods, and (2) predictive methods. While analytical methods aim to efficiently analyse, represent and interpret data (static or longitudinal), predictive methods leverage the data currently available to predict observations at later time-points (i.e., forecasting the future) or predicting observations at earlier time-points (i.e., predicting the past for missing data completion). For instance, a method which only focuses on classifying patients with mild cognitive impairment (MCI) and patients with Alzheimer’s disease (AD) is an analytical method, while a method which predicts if a subject diagnosed with MCI will remain stable or convert to AD over time is a predictive method. Similar examples can be established for various neurodegenerative or neuropsychiatric disorders, degenerative arthritis or in cancer studies, in which the disease/disorder develops over time.
Why predictive intelligence?
It would constitute a stunning progress in the MICCAI research community if, in a few years, we contribute to engineering a ‘predictive intelligence’ which can map both low-dimensional and high-dimensional medical data onto the future with high precision. This workshop is the first endeavor to drive the field of ‘high-precision predictive medicine’, where late medical observations are predicted with high precision, while providing explanation via machine and deep learning, and statistically, mathematically- or physically-based models of healthy, disordered development and ageing. Despite the terrific progress that analytical methods have made in the last twenty years in medical image segmentation, registration or other related applications, efficient predictive intelligent models/methods are somewhat lagging behind. As such predictive intelligence develops and improves —and this is likely to do so exponentially in the coming years— this will have far-reaching consequences for the development of new treatment procedures and novel technologies. These predictive models will begin to shed light on one of the most complex healthcare and medical challenges we have ever encountered, and, in doing so, change our basic understanding of who we are.
What kind of research problems we aim to solve?
The main aim of PRIME-MICCAI is to propel the advent of predictive models in a broad sense, with application to medical data. Particularly, the workshop will admit 12-page papers describing new cutting-edge predictive models and methods that solve challenging problems in the medical field. We hope that PRIME workshop becomes a nest for high-precision predictive medicine, one that is set to transform multiple fields of healthcare technologies in unprecedented ways.
Topics of interests include but are not limited to predictive methods dedicated to the following topics:
- Modeling and predicting disease development or evolution from a limited number of observations;
- Computer-aided prognostic methods (e.g., for brain diseases, prostate cancer, cervical cancer, dementia, acute disease, neurodevelopmental disorders);
- Forecasting disease/cancer progression over time;
- Predicting low-dimensional data (e.g., behavioral scores, clinical outcome, age, gender);
- Predicting the evolution or development of high-dimensional data (e.g., shapes, graphs, images, patches, abstract features, learned features);
- Predicting high-resolution data from low-resolution data;
- Prediction methods using 2D, 2D+t, 3D, 3D+t, ND and ND+t data;
- Predicting image modality from a different modality (e.g., data synthesis);
- Predicting lesion evolution;
- Predicting missing data (e.g., data imputation or data completion problems).
- Predicting clinical outcome from medical data (genomic, imaging data, etc).
This workshop will mediate ideas from both machine learning and mathematical/statistical/physical modeling research directions in the hope to provide a deeper understanding of the foundations of predictive intelligence developed for medicine, as well as to where we currently stand and what we aspire to achieve through this field. PRIME-MICCAI 2023 will feature a single-track workshop with keynote speakers with deep expertise in high-precision predictive medicine using machine learning and other modeling approaches —which are believed to stand at opposing directions. Our workshop will also include technical paper presentations, poster sessions, and demonstrations. Eventually, this will help steer a wide spectrum of MICCAI publications from being ‘only analytical’ to being ‘jointly analytical and predictive’.
Short bio: Dr Wenjia Bai is Senior Lecturer (Associate Professor) in AI in Medicine at Imperial College London, with joint appointments at Department of Computing and Department of Brain Sciences. His research is at the interface between machine learning and medical imaging, with a focus on developing computational and machine learning algorithms to understand the structure, motion and function of anatomical organs from medical images. In the recent couple of years, he has been focusing on developing algorithms that are deployable for population-level imaging data analysis, such as for the UK Biobank project that consists of imaging scans for 100,000 subjects across the UK. The developed algorithms greatly facilitate clinical research in leveraging large-scale datasets for understanding phenotypic traits in human health and diseases.
Short bio: Dr. Xiaoxiao Li is an Assistant Professor at the Department of Electrical and Computer Engineering (ECE) at the University of British Columbia starting August 2021. Before joining UBC, Dr. Li was a Postdoc Research Fellow in the Computer Science Department at Princeton University. Dr. Li obtained her PhD degree from Yale University in 2020 and received Yale Advanced Graduate Leadership Fellowship. Dr. Li is leading the Trusted and Efficient AI Lab. Her research interest ranges across the interdisciplinary files of deep learning and biomedical data analysis, aiming to improve the trustworthiness of AI systems for healthcare. In the recent few years, Dr. Li has published in leading machine learning conferences and journals, including NeurIPS, ICML, ICLR, MICCAI, IPMI, IEEE Transactions on Medical Imaging, Medical Image Analysis, Nature Methods. Her work has been recognized with the several best paper award in international conferences..
Short bio: Dr Dalca is an Assistant Professor, A.A. Martinos Center for Biomedical Imaging Massachusetts General Hospital, Harvard Medical School. He is a Research Scientist at the Computer Science and Artificial Intelligence Lab EECS, Massachusetts Institute of Technology. His research focuses on machine learning techniques and probabilistic models with a frequent focus on medical image analysis, computer vision, and healthcare. He was a postdoctoral fellow at CSAIL, MIT and MGH, Harvard Medical School, working with Mert Sabuncu and John Guttag. He completed my PhD in the Medical Vision Group, CSAIL, EECS, MIT, advised by Polina Golland.
Program Committee Members
Ahmed Nebli, Forschungszentrum Jülich, Germany
Gang Li, University of North Carolina,USA
Hao Chen, The Hong Kong University of Science and Technology, China
Ilwoo Lyu, Vanderbilt University, USA
Jaeil Kim, Kyungpook National University (KNU), Korea
Jiahong Ouyang, Stanford University, USA
Lequan Yu, The University of Hong Kong, China
Li Wang, UNC, USA
Lichi Zhang, Shanghai Jiao Tong University, China
Manhua Liu, Shanghai Jiao Tong University, China
Maria A. Zuluaga, EURECOM, France
Moti Freiman, Technion – Israel Institute of Technology, Israel
Pew-Thian Yap, University of North Carolina (UNC), USA
Philip Chikontwe, DGIST (Daegu Gyeongbuk Institute of Science and Technology), Korea
Qian Tao, Delft University of Technology, Netherlands
Qingyu Zhao, Stanford University, USA
Reza Azad, RWTH University, Germany
Seung Yeon Shin, National Institutes of Health, USA
Soochahn Lee, Kookmin University, Korea
Su Ruan, Laboratoire LITIS, France
Tolga Tasdizen, University of Utah, USA
Won-Ki Jeong, Korea University, Korea
Yalin Wang, Arizona State University, USA
Yonggang Shi, USC, USA
Ziga Spiclin, University of Ljubljana, Slovenia
The number of pages can range between 8 and 12 pages including references. Papers should be formatted in Springer LNCS style. PRIME reviewing is double-blind.
The accepted papers will be published in the PRIME LNCS Springer Proceedings.
For paper submission, please use the following link: https://cmt3.research.microsoft.com/PRIME2023
To download PRIME 2023 flyer, click here.
Full Paper Deadline:
June 16, 2023; 11:59 PM EST; extended to June 30, 2023 at 11:59 PM EST
Notification of Acceptance: July 25, 2023
Camera-ready Version: August 5, 2023, 11:59 PM PST
Workshop date: 8 October 2023 – Sunday PM
Papers should be submitted electronically following the guidelines for authors and LaTeX and MS Word templates available at Lecture Notes in Computer Science, double-blind review). Manuscripts should be up to 12 pages and submitted via the PRIME submission website. No modifications to the templates are permitted. Failure to abide by the formatting guidelines will result in immediate rejection of the paper. The papers will be evaluated by three external reviewers or potential inclusion in the scientific program of PRIME-MICCAI.
We have also included a checklist for paper reproducibility.
SPECIAL PRIME 2023: Talented minority scholarship to register accepted PRIME papers by students in low-middle income countries (LMIC).
-We offer scholarships supported by BASIRA Lab to register accepted PRIME papers if the first author is a student at a University in a low-middle-income country. If the rank of your country is larger or equal to 50 based on expenditure on R&D sorted in decreasing order, then you are eligible:
Instructions for submitting your camera-ready paper:
Please make sure to upload a zip file including the following documents:
1. The final camera-ready PDF of your paper.
2. All original files are required to generate the final PDF. If you are using Latex, please make sure to include .tex, .bib, figures, and any other files for compiling the tex file. If you are using Word, please upload the docx file. Name the main file using your submission ID number (example: 32.tex or 32.docx).
3. Please download and fill out the Springer copyright form which can be found using this link: https://shorturl.at/gjSU2
Include the signed PDF form of the PRIME LNCS Copyright document.
Please make sure that the copyright forms have been filled out correctly before uploading the zip file.
The paper length can vary between 8 and 12 pages.
Finally, please upload only one zip (not rar) file with everything included (using your paper submission ID, for example: PRIME22-32.zip, where 32 is your paper ID number).