Issue with Generating Membrane Segmentations

[ashishsdalvi]

Hi DeePiCt team,

I ran the tool on my data and made sure to preprocess it according to your instructions. I used your config.yaml but turned the particle picking and segmentation evaluation to false since I am not looking to evaluate right now. I made sure to use the memb model pth file. in post_processing_clustering i set the threshold to 0.2 and min_cluster_size to 10 to be more lenient but even with those parameters it doesn't seem to segment the membrane. The tomogram I am processed has a voxel size of 9 angstroms. Any ideas on what I can try? Btw as a side note I am doing this on my server but it appears that the colab doesn't work because the np.int in clustering.py and the model loading function in segment.py are outdated. Thank you for your time and I appreciate your open source tool!

Sincerely,
Ashish Dalvi

The first 2 images are the post_processed_predictions.mrc at 2 similar z slices and the last image is the tomogram before any processing which clearly shows membranes. I would also like to segment ribosomes after I get it to work for membranes if you have any advice for this as well!