TY - CONF
T1 - Large-Scale Deep Learning Medical Image Methodology and Applications using Multiple GPUs
AU - Krishnasamy, Ezhilmathi
AU - Boudissa, Selma
AU - Kanli, Georgia
AU - Marzal-Abarca, Francesc Xavier
AU - Keunen, Olivier
AU - Bouvry, Pascal
N1 - The first author is funded by the EuroCC (U-AGR-8200-00-C), and the second author is funded by the T ́el ́evie grant F5/20/5–TLV/BC. We also thank the computational resources provided through the EHPC-DEV-2022D11-022 EuroHPC Development Access Call. The authors declare that there is no conflict of interes
PY - 2023/10/12
Y1 - 2023/10/12
N2 - Digitalization of medical images has become firmly established in research labs and clinical settings worldwide, and its numerous benefits have helped drive its rapid expansion, which is predicted to continue momentum over the coming years. Its advantages of cutting costs, improving analysis, productivity and patient outcomes, reducing errors, and enhancing imaging and innovation will likely propel its widespread adoption. In recent years digitalization of medical images especially relied on high performance computing. This work illustrates how a large volume of digital radiology data can be analysed in a parallel heterogeneous architecture. We have used TensorFlow (including Keras) and the Horovod API to speed up the learning of Deep Learning models, particularly using GPUs in single and multiple compute nodes. We conclude that using Horovod API with data parallelism shows tremendous speed gains compared to traditional CPU when training the 2D UNet model.
AB - Digitalization of medical images has become firmly established in research labs and clinical settings worldwide, and its numerous benefits have helped drive its rapid expansion, which is predicted to continue momentum over the coming years. Its advantages of cutting costs, improving analysis, productivity and patient outcomes, reducing errors, and enhancing imaging and innovation will likely propel its widespread adoption. In recent years digitalization of medical images especially relied on high performance computing. This work illustrates how a large volume of digital radiology data can be analysed in a parallel heterogeneous architecture. We have used TensorFlow (including Keras) and the Horovod API to speed up the learning of Deep Learning models, particularly using GPUs in single and multiple compute nodes. We conclude that using Horovod API with data parallelism shows tremendous speed gains compared to traditional CPU when training the 2D UNet model.
KW - Deep Learning
KW - Digital Radiology
KW - UNet
KW - Keras
KW - Horovod
KW - GPU
KW - HPC
UR - http://www.scopus.com/inward/record.url?scp=85178616243&partnerID=8YFLogxK
U2 - 10.1109/ICABME59496.2023.10293105
DO - 10.1109/ICABME59496.2023.10293105
M3 - Paper
AN - SCOPUS:85178616243
SP - 17
EP - 22
T2 - 7th International Conference on Advances in Biomedical Engineering, ICABME 2023
Y2 - 12 October 2023 through 13 October 2023
ER -