Eager to try something COOL?

Interested in improving the largest connectome to date by yourself? Here are the instructions!

For any questions, please contact Ting Zhao ([email protected])


  • Active DVID database that supplies the hemibrain dataset. Visit dvid.io for more instructions.

  • A modern Linux/Mac machine with a mainstream Graphics card

    • Make sure the machine can connect to the database

Install NeuTu

curl -X GET https://raw.githubusercontent.com/janelia-flyem/NeuTu/master/neurolabi/shell/setup_neutu.sh > setup_neutu.sh
bash setup_neutu.sh <install_dir> neutu

Note: <install_dir> is the path to where NeuTu will be installed. It is preferred to be a non-existing or empty directory.

Start NeuTu


Load Dataset from DVID

You can follow the instructions from Get Started to open a dataset in NeuTu. You can also copy and paste the following settings to NeuTu after replacing <grayscale_uuid> with the actual grayscale UUID, assuming that the segmentation/synapse data are hosted on localhost:8000 with and the grayscale data are hosted on localhost:8001 with uuid <grayscale_uuid>. For different server addresses, you need to replace the address settings accordingly.

"address": "localhost",
"port": 8000,
"name": "local_test",
"default": false,
"segmentation": "segmentation",
"uuid": "52a1",
"proofreading": true,
"synapse": "synapses",
"supervised": false,
"@grayscale": [
"address": "localhost",
"port": 8001,
"background": 0,
"gray_scale": "grayscale",
"default": false,
"uuid": "<grayscale_uuid>"
Paste Settings to Load DVID Dataset