DeepVoxNet2
A Python library to implement deep learning pipelines for medical applications using convolutional neural networks (CNNs).
Description:
DeepVoxNet2 (DVN2) is a Python library to make it easier to implement deep learning pipelines for medical applications using convolutional neural networks (CNNs). The library can be used as an add-on to Tensorflow/Keras, Pytorch or any other deep learning framework in Python. Currently, the use with Tensorflow/Keras is simplest, with readily available CNN architectures, metrics, losses and the DvnModel class to group your entire pipeline and, e.g., bypass the use of Keras' fit function, etc.
More Information:
Originates from the AI-Driven Data Science research challenge (FAIR 2019–2023).
Responsible Research Lead: Frederik Maes
Authors: Jeroen Bertels, Lotte Huysmans
Programming Language: Python
Categories: Supervised learning
Publications:
Bertels, J., Robben, D., Lemmens, R., & Vandermeulen, D. (2022). DeepVoxNet2: Yet another CNN framework. ArXiv, 1–15. http://arxiv.org/abs/2211.09569