Abstract | ||
---|---|---|
•We propose a deep learning method for stroke lesion segmentation from multimodal MRI.•We use a combination of patches and a weighted loss function to handle class imbalance.•We pre-process the data to exploit the brain symmetry between hemispheres.•We analyse its performance using two tasks from the ISLES challenge.•We rank first in both tasks without any dataset specific training parameter tuning |
Year | DOI | Venue |
---|---|---|
2020 | 10.1016/j.cmpb.2020.105521 | Computer Methods and Programs in Biomedicine |
Keywords | DocType | Volume |
Brain,MRI,Ischemic stroke,Automatic lesion segmentation,Convolutional neural networks,Deep learning | Journal | 194 |
ISSN | Citations | PageRank |
0169-2607 | 1 | 0.37 |
References | Authors | |
0 | 6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Albert Clèrigues | 1 | 1 | 0.37 |
Sergi Valverde | 2 | 228 | 24.13 |
José Bernal | 3 | 20 | 3.29 |
Jordi Freixenet | 4 | 887 | 64.40 |
Arnau Oliver | 5 | 1034 | 83.82 |
Xavier Llado | 6 | 578 | 40.04 |