eil-nn is an image processing toolkit to analyse tomography data for battery electrodes. It uses neural networks to segment the images and successively identify and classify particles.
Follow the steps given below to install the eilnn
Python package. The package must be installed to run the included examples. It is recommended to create a virtual environment for the installation.
# Clone the repository
$ git clone https://github.com/sdaemi/eil-nn.git
# Got to the root directory
$ cd eil-nn
# Install the eilnn package from within the repository
$ pip install -e .
The two example notebooks in the "examples" fodler indicate the two main functionalities of the package, namely automated COCO-style annotation generation and segmentation + classification. Please note for this iteration of package, Mask R-CNN (preferably GPU enabled) must be previously installed.
Currently, features such as the removal of partially truncated particles are tailored to data collected using a lab-based Zeiss Ultra 810 X-ray computed tomography instrument. If you need to analyse other types of data, please feel free to get in touch.
Both the pre-trained classifier model and Mask R-CNN weights are provided in the model/ folder for use with these tools or others.