Interested in creating a plugin for dataset preparation

[RafeMcBeth]

I'm interested in creating a plugin that helps me to prepare/standardize a large number of patient datasets and convert them to numpy arrays for use in our models. I have created a standalone jupyter notebook that does exactly what I want that heavily uses the dicompyler-core library but it might be more useful to add it as a plugin. Here are my requirements.

• Select a subset of structures that are mapped to a standardized naming convention. It would be useful to be able to visualize the structure (scroll through slices, toggle) and also look at the corresponding DVH. The goal is to rename the dicom structures or just export the name mappings.
  -Create binary masks for each standardized structure at a specified resolution
  -Export dosewash, ct and binary mask of structures at a specified resolution, ensuring that they are all correctly align, to 3D numpy arrays.

I'm not sure the best place to start based on the structure of the code and would be interested in any feedback.

[bastula]

Hi sorry for the delay in response. I don't really check this repository often. I'll try to look at it more in the future.

Do you think this would be useful to other people in the community? If so, I would post a message to the dicompyler google groups thread as that has more active users than the issues list here. 😄

If so, then we can explore and discuss how to translate your notebook to a plugin. Another person who has experience with this kind of data processing work may be @cutright.

[cutright]

Sounds like maybe a combination of DVH Analytics and dicompyler would be a fruitful direction?

@bastula Maybe I could add a method to launch an instance of the dicompyler GUI and feed it file paths to be automatically loaded?

@RafeMcBeth I've been meaning to add plug-in support to DVHA and I already have methods to build an ROI naming map. Presumably your script would just need file paths (or dicom datasets) and the roi map?