← Back to team overview

sikuli-driver team mailing list archive

Re: [Question #150551]: Which makes faster scans, increasing or decreasing similarity

 

Question #150551 on Sikuli changed:
https://answers.launchpad.net/sikuli/+question/150551

    Status: Open => Answered

RaiMan proposed the following answer:
I made this test:

x = 50
for s in [0.2, 0.99]:
	start = time.time()
	for i in range(20):
		m = find(Pattern("some_image.png").similar(s))
	print '***', s, (time.time()-start)*1000/x

        m.highlight(2)

In general, it seems to be, that searching with 0.2 is faster than
searching with 0.99 (the match always has a similarity of 1.0). The
difference increases with increased size of the searched image.

e.g. searching on a screen of 1280x800 in average an image of 600 x 500
is matched at 0.2 after 350 msec and with 0.99 after 750 msec. - with
100 x 150 it is about 130 to 150.

But be aware, lowering the min similarity increases the risk to match
the wrong object.

one more thing:
the best way to speed up searches is to make the region of interest as small as possible.

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