Pipeline Creation

In order to create a new Pipeline you’re advised to use NEFI2 pipeline panel. This will guarantee that you won’t make any syntax errors. However if you still want to do it manually here is a quick overview.

The pipeline files are stored in default_pipelines/ directory. You can keep your pipeline json files anywhere you want, but if you put them in default_pipelines/ they will be loaded automatically at startup.

Pipeline Syntax Example

[
  ["Watershed DE Adaptive",{"Foreground Iteration":0,
                            "Background Iteration":0,
                            "Threshold Block Size":11,
                            "Threshold Constant":1,
                            "type":"Segmentation",
                            "store_image":false}],

  ["Median Blur",{"channel1":true,
                  "channel2":true,
                  "channel3":true,
                  "kernelsize":2,
                  "type":"Preprocessing",
                  "store_image":false}],

  ["Guo Hall",{"store_image":false,
               "type":"Graph Detection"}],

  ["Keep only LCC",{"store_image":false,
                    "type":"Graph Filtering"}],

  ["Simple Cycle",{"store_image":false,
                   "type":"Graph Filtering"}]
]

The order of the Algorithms ["Guo Hall", ...], ["Keep only LCC", ...] is important, however the order of the settings slot-value pairs "kernelsize":2 is not. If the pipeline is correct it will be loaded, otherwise nothing will be added to the Pipeline panel and you’ll have need to revise your json file.