Been playing around with Computer Vision to automate a process that’s running on Dynamics Axapta on a Virtual Desktop (Terminal Server).
Now, from the main app I define a CV scope and I use CV click to open up another window. However I want to return to the first window at some later stage, but how do I define this? Can I not create a variable for this? I’m missing a function similar to “Output” for the “Attach Window” activity.
I don’t find a reusable CV Scope either. Maybe it is already in the works for enhancement, also, I have given it as an idea so the product team might consider it, but, there might be more pressing enhancements that are in the pipeline.
Could you explain the reasoning behind this ask? Like, why is it better for you to use a variable that you have to type in each scope, instead of clicking on the app once (indicate)?
Say I use CV Scope for a window, then my understanding is that it “image scans” the “stream” and with AI/ML it detects the UI elements. Now if I open another window another CV scope has to learn this new window and so on. When I close this window/scope and come back to the “first” window it would be elegant if I could somehow refer to it by a variable name.
It seems very cumbersome working with all these scopes. Also: How does CV know which window it is working in?
So what’s the advantage of using a variable that you have to type in each scope, instead of clicking on the app once (indicate) to have it analysed? I guess execution time, since you have one analysis less, but other than that?
How are scopes cumbersome? Maybe we can improve them.
What I have seen is every-time you indicate the scope during development it will take some time to create it, So might as well store it, so that repeated calls aren’t made and dev time is reduced if you need to come back to the same screen again and again
Regards
You’ll be happy to hear that we’ve added this improvement to our backlog.
We intend to enable this passing of CV Scopes both for design-time and runtime, but for designtime it will have some limitations: you won’t be able to reuse the detections done in another workflow, so a new detection will be performed. It will reuse scopes if they are in the same workflow, and of course at runtime it will always reuse detections.