neural networks for brain image analysis
This toolbox is currently in development, with the goal providing a set of tools with infrastructure for medical image analysis with neural network. While the tools are somewhat general, neuron will generally run with keras on top of tensorflow.
callbacks: a set of callbacks during keras training to help with understanding your fit, such as Dice measurements and volume-segmentation overlaps
generators: generators for medical image volumes and various combinations of volumes, segmentation, categorical and other output
dataproc: a set of tools for processing medical imaging data for preparation for training/testing
metrics: metrics (most of which can be used as loss functions), such as dice or weighted categorical crossentropy.
models: a set of flexible models (many parameters to play with...) particularly useful in medical image analysis, such as a U-net/hourglass model and a standard classifier.
Other utilities and a few jupyter notebooks are also provided.
- tensorflow
- keras and all of it's requirements (e.g. hyp5)
- numpy, scipy
- tqdm
- python libraries from @adalca github account
Please contact Adrian Dalca, [email protected] for question related to neuron