Efficient multi-scale 3D CNN with fully connected CRF for accurate brain lesion segmentation
- Submitting institution
-
Imperial College of Science, Technology and Medicine
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 2404
- Type
- D - Journal article
- DOI
-
10.1016/j.media.2016.10.004
- Title of journal
- Medical Image Analysis
- Article number
- -
- First page
- 61
- Volume
- 36
- Issue
- 1
- ISSN
- 1361-8423
- Open access status
- Compliant
- Month of publication
- October
- Year of publication
- 2016
- URL
-
-
- Supplementary information
-
10.1016/j.media.2016.10.004
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- Yes
- Number of additional authors
-
7
- Research group(s)
-
-
- Citation count
- 976
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- DeepMedic is a neural network architecture outperforming state-of-the-art image segmentation systems. The publicly available source code (https://github.com/Kamnitsask/deepmedic; >400 stars, >200 forks) has been widely used, e.g.in large clinical studies (UPenn, Mayo Clinic), industry (Philips, Microsoft), and adopted in popular toolkits such as NiftyNet (UCL), CapTk (UPenn) and DLTK (Imperial). DeepMedic has won several international computational challenges (ISLES 2015, BRATS 2017) and received honorary mention for the 2016 NVIDIA Global Impact Award. DeepMedic has been instrumental in grant awards, including an ERC Starting Grant (MIRA; €1.5M), 2 MRC/NIHR grants (MALIBO, MALIMAR), and 3 EPSRC grants. Amongst most downloaded MIA articles (https://www.journals.elsevier.com/medical-image-analysis/most-downloaded-articles).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
