The latest release of ESM (12.5) has new features regarding image recognition, including this parameter:
In the Citrix monitoring script I have, I've noticed that after upgrading from 12.4.5 to 12.5.1 all my image verifications ALWAYS succeed. Which is of course not the preferred scenario, if the image isn't actually there. When I disable that AdvancedImageProcessing parameter (and with it the "edge detection" feature), the image verifications begin to work as intended again. I.e. if the image is there, the WaitForImage function returns true, and if it isn't, WaitForImage returns false.
So according to what I've seen so far, if the AdvancedImageProcessing parameter is set to true, WaitForImage always returns true. I'm curious to hear other people's experience with this feature thus far, and if there are any best practises I could consider when configuring the monitoring?
I noticed that the "Locate" button on the Agent Recorder (after opening an image map definition) appears to work correctly regardless of the AdvancedImageProcessing value, so perhaps the "edge detection" isn't used there? I mention this because I usually test the image maps from there before adding them to script, and that's why I was so surprised to see my script just run forward even during situations when the image should not have been found.
"Significantly reduced overhead of the OCR and IR pattern matching"
"Highly improved IR precision"
-> I would also like to know if these improvements regarding IR described in the 12.5 release notes are only active if this AdvancedImageProcessing = True? So are these performance and reliability improvements thanks to "edge detection", or has the IR functionality also changed in other ways since 12.4.5?
Solved! Go to Solution.
Just to be more accurate: there are actually _some_ cases when the AdvancedImageProcessing doesn't find the image when it's not there (i.e. it works correctly). But in most cases it seems to find a nonexistent image. Many of my image maps are pieces of text, and even though I tried to only capture the mid/center part of the text (in order to leave as little as possible of the background there), AdvancedImageProcessing "finds" the text even though it's not there. So I suppose it would help if there's some information on how the edge detection works; to me it appears that it accepts the image as found even there are very small fractions of it visible.
There are situations where using advanced image processing against textual images can cause false positives . To help with your particular issues it would be best if you open a support ticket where you can provide more detail so that we can better diagnose the issue(s) .