AIW-I_Task_Performer_Capabilities

Overview

The AI Workflow for Imaging (AIW-I) Profile specifies how to request, manage, perform, and monitor AI Inference on digital image data.  

Both the sequence of transactions in AIW-I and the content of the workitem(s) created by the Task Requester depend on the AI inferences and workflows implemented on the AIW-I Task Performer actor.  Therefore, the purpose of this Preparatory test is to gather information from the Task Performer which will influence how it will interact with its test partners during the Connectathon.   The Task Performer will describe:

  • AIW-I workflows it supports. See AIW-I Section 50.1.1.3, Section 50.4.1.5, and Section 50.4.2 .  One or more of:
    • Pull workflow
    • Triggered pull workflow
    • Push workflow
  • the AI algorithms (inferences) it has implemented
  • the inputs each algorithm needs when it is triggered.  See AIW-I Section 50.4.1.1 and 50.4.1.2.

This description will help the Task Requester ensure that the workitems it creates are adequately populated for you, and that you test the workflow(s) you support with your partners at the Connectathon.

Instructions for Task Performers

For this test you (the Task Performer) will produce a short document describing your implementation in the context of the AIW-I Profile specification.  According to AIW-I, Section 50.4.1.1, a DICOM Conformance Statement is the ideal home for these details.  If you have one, great!  But, for the purpose of this preparatory test, the format of the document is not important.  It may be a PDF, a Word or google doc, or some other narrative format.

Your document shall include the following content:

  1. Your system name in Gazelle Test Management (eg. OTHER_XYZ-Medical)
  2. Technical Contact name/email - this is someone who can be contacted if there are questions about what you provide below
  3. The AIW-I Workflow(s) you support:   Pull Workflow,  Triggered-pull Workflow, and/or Push Workflow
  4. AI Algorithm Description - this should be a sentence or two describing what your algorithm does (e.g. detect lung nodules)
  5. The Workitem Code that will trigger this AI Algorithm - refer to AIW-I Section 4.1.2.   You may use a code in Table 50.4.1.2-1 if there is one that applies.  Otherwise, suggest a code that is appropriate for your AI algorithn
  6. The Input Parameters  & Values required by your AI Algorithm - you will need to be very specific in answering this.  Please refer to AIW-I Section 50.4.1.1 and Table 50.4.1.1-1.  Identify the UPS attribute(s) your algorithm relies on, and the value(s) you expect for each.
  7. The Input Information Sequence content your AI altorithm requires - Please refer to Section 50.4.1.3 and identify the DICOM images (if any) that your AI algorithm expects to see in the Input Information Sequence (0040,4021) of the UPS Workitem that triggers your AI algorithm.
  8. Any other information that will help the reader understand your algorithm and how it is triggered.
  9. REPEAT 3 - 8 for each AI Algorithm that can be triggered on your Task Performer
  10. Finally, in order to share it with your test partners, upload your document as a Sample in Gazelle Test Management.  On the 'List of Samples' page, use the dropdowns to find your test system, and on the 'Samples to share' tab, find the "AIW-I_Performer_Capabilities" entry and upload your document there.   When you save your sample, it will be visible to your test partners.

 

Instructions for Task Requesters, Managers, Watchers

You will find and read the document provided by the Task Performer above.

  1. In Gazelle Test Management, on the 'List of Samples' page, use the dropdowns to find your test system, and on the 'Samples available for rendering' tab, find the "AIW-I_Performer_Capabilities" entry with the document provided by your Task Performer partner(s) above.  
  2. Download that document and use it to configure your system for items such as workitem codes, input parameters, etc, that you will need in order to create a UPS workitem [RAD-80] for that Performer

Evaluation

There is no "pass/fail" for this test.  However, you must complete it because it is a prerequisite for several Connectathon tests.  Your AIW-I test partners, plus the Connectathon monitor, will be looking for the document produced here.