N5¶

WEBKNOSSOS can read N5 datasets.

N5 datasets can both be uploaded to WEBKNOSSOS through the web uploader or streamed from a remote server or the cloud.

WEBKNOSSOS reads N5 but does not write it. To produce N5 datasets, use the external N5 tooling from the N5 ecosystem (e.g. the n5 Java or Python libraries). If you want WEBKNOSSOS to create a dataset for you, consider the natively supported Zarr or WKW formats instead.

Examples¶

You can try the N5 support with the following datasets. Load them in WEBKNOSSOS as a remote dataset:

Interphase HeLa cell EM data hosted on AWS S3
- s3://janelia-cosem-datasets/jrc_hela-3/jrc_hela-3.n5/em/fibsem-uint16
- Source: Open Organelle Project

N5 folder structure¶

WEBKNOSSOS expects the following file structure for N5 datasets:

my_dataset.n5             # One root folder per dataset
├─ attributes.json        # Dataset metadata
└─ my_EM                  # One N5 group per data layer. In WK directly link to a N5 group.
   ├─ attributes.json                
   ├─ s0                  # Chunks in a directory hierarchy that enumerates their positive integer position in the chunk grid. (e.g. 0/4/1/7 for chunk grid position p=(0, 4, 1, 7)).
   │  ├─ 0
   │  │  ├─ <chunks>
   │  ├─ ...
   │  └─ n
   ...                
   └─ sn

For details see the N5 spec.

Performance Considerations¶

To get the best streaming performance for N5 datasets consider the following settings.

Use chunk sizes of 32 - 128 voxels^3