Unable to Find Text/String from the Page

Hello Folks,

I’m trying to read the IP Address appearing in the Page , refer Snapshot, from the List of available IP Addresses. I tried to extract the IP " 10.112.58.91" from the list. The eggplant IDE is unable to recognize/found the string “10.112.58.91” . I Implemented in code by

click(text: “10.112.58.91”)

Error: Selection Failed, No Image Found on Screen.

However the above code is working for “Add” (refer Snapshot , top-left corner)
click (text: “Add”) : Successfull: able to click over “Add”.

The difference between “10.112.58.91” and “Add” is only of Font Size and boldness.

Group, Any help in this Scenario will be highly appreciable.

Thanks,
S

You need to be aware that in order to perform OCR, every pixel has to be examined and a determination made whether the pixel is part of a character and what that character is. Based on what the human eye sees on the screen, this doesn’t seem like much of a challenge, but at screen resolutions and with antialiasing thrown into the mix, it’s harder than it seems.

For example, this is what the word “Accounts” looks like at the pixel level when ClearType font smoothing is enabled. The characters are a bit run together and the two c’s could just as easily be o’s or a “co” or “oc”. How does the OCR decide what they actually are? One way is by comparing the possible permutations against a dictionary; it doesn’t find “aooounts”, “acoounts”, or “aocounts”, but “accounts” is a word, so that’s likely what it will return. If for some reason your screen actually contained one of the other permutations, it’s probably going to get it “wrong”.

In our tests it’s actually been quite reliable at reading a high percentage of text, even when in “English” terms the text is gibberish, but there are things that can trip it up. We’ll continue to try to improve the results that are returned.