MitoEM-challenge
starter code for mitoEM ISBI2021 Challenge
Data Visualization
Here's the neuroglancer link for [MitoEM-R], [MitoEM-H]
Auxiliary code to generate H5 files
If you have mitochondria predictions images in a folder, sorted by name, you can convert them into a H5 file following the code.
Notice some points of the script:
- The volume is created with a downsampled (by 2) version of the original images
- Connected components is used to label the different mitochondria instances. However, another labelling method are also allowed (please describe it in the submission comments or in supplementary files)
- LZF compression is applied in order to save space and submission time