search

chrischute / numpy2roi

Import a 3D NumPy array as an ROI in OsiriX/Horos.
2 stars 0 forks source link

npy shape #1

Open yilei-wu opened 5 years ago

yilei-wu commented 5 years ago

Hi, May I ask what is the acceptable numpy array shape for this plugin?

chrischute commented 5 years ago

Hi Yilei,

This repo was developed for a 3D medical imaging project, where there is only one channel in the input images (the entries were sigmoid outputs, i.e. floats between 0 and 1).

The expected shape is (depth, height, width).

yilei-wu commented 5 years ago

Hi Chris, Thank you so much for your reply. I assume your repo is trying to feed the result from deep learning to the dicom viewer. I am testing your plugin with a predicted 3D segmentation mask on osirix. My segmentation is originally mat file but I can rewrite them into .npy. One other question is how many region of interest you expect for one npy? My predicted mask is a 3d contour and it shows normal when I open it as the Image. But it shows as scattered points when I import it to osirix thourgh the numpy2roi plugin. Any input is appreciated. Thank you.