← Back to team overview

sikuli-driver team mailing list archive

Re: [Question #233608]: Efficiency for Chess

 

Question #233608 on Sikuli changed:
https://answers.launchpad.net/sikuli/+question/233608

    Status: Answered => Open

jay dee is still having a problem:
Mr. RaiMan, let us jump further down this rabbit hole... I have a few more questions for you regarding this potential pathway. 
1.) How fast does sikuli capture a 1920x1080 desktop screen capture ? No processing, just constant grabbing of screen and dumping the data into the same object over and over. What is the best speed one can achieve (perhaps in fps?) ? 

The reason for this would be to have real-time piece detection. I have
reason to believe this is the slowest implementation.

2.) Otherwise, would having a simple change-was-detected method that
triggers the pieceDetection method be better? I believe this might be so
because the simple change method will not take as many cpu cycles. It's
only when it detects a change on one of the squares will it call the
more intensive method.

3.) I will be using sikuli for another portion of the project which will
not be as crazy as a chess game. I am trying to implement it anywhere I
can because the API is so easy. If this truly is too slow to be
implemented then I will look at other options. Are there any other image
processing libraries that may fit my purpose better for chess piece
recognition? Sikuli has the best similarity compare with the fuzzy
logic, its hard to beat.

4.) What is the difference between the regular sikuli and the exclusive
sikuli java API ???  How is the java api 1.0.2 when the regular sikuli
is not even 1.0.1 yet the java was released in 2012? Very confusing,
this warrants an explanation please. What do I want to use for just
strict java?

Thank you for your time sir.

-- 
You received this question notification because you are a member of
Sikuli Drivers, which is an answer contact for Sikuli.