Try to visualize the intermediate states in ThaumatoAnakalyptor pipeline. Currently for point cloud step only.
Install dependency
pip install numpy torch tifffile scipy open3d opencv-python
Choose a cell block here and run the script. It's slightly modify from ThaumatoAnakalyptor point cloud step. And will generate an output
folder to save intermediate info as torch tensors. Checkout here to learn more. Here's an example command if you place the data in root folder.
python surface_detection.py cell_yxz_006_008_004.tif
Now, let's visualize those tensors 🙌
python visualize.py
I think it's a great starting point to learn ThaumatoAnakalyptor. Tweak values with visual feedback. Try to figure out which parts could be done better. We can go even further by putting those output videos on an endless whiteboard. Here's a prototype demo. This part has not been started yet, but may be implemented via VC Whiteboard.
In addition, I will also discuss with VC community how to integrate these python scripts into ThaumatoAnakalyptor itself to avoid maintaining too much identical code.
Let's visualize the entire pipeline 🙌