Using DIS dense optical flow in conda OpenCV 3.4 [closed]

asked 2018-11-22 06:49:32 -0600

kdx2 gravatar image

Hello OpenCV community, I am working on an application where speed performance is a critical parameter and I found out that the Dense Optical Flow I am using is causing the biggest slow-down in the processing. I found further out that OpenCV 4 implements Dense Inverse Search Optical Flow which is supposed to be faster and I have some hopes on it. I was wondering if there is any way I can use DIS in the Python opencv 3.4 ran in Anaconda Spyder? If so, what are the steps shall I take? (I am not good with linking libraries and sorting out dependencies manually)

Besides, do you plan on releasing OpenCV 4 for Anaconda, if so - when? Do you plan adding the gpu-accelerated open cv to Anaconda, too?

Thank you in advance, Konstantin

edit retag flag offensive reopen merge delete

Closed for the following reason question is off-topic or not relevant by berak
close date 2018-11-22 09:29:39.960509

Comments

opencv does not maintain any anacoonda releases

berak gravatar imageberak ( 2018-11-22 07:12:30 -0600 )edit

Ok, even so, is there a way for us to run DIS OF in OpenCV 3.4? If yes, what is it?

kdx2 gravatar imagekdx2 ( 2018-11-22 07:36:08 -0600 )edit

for 3.4 you would need the contrib modules , no idea, if conda offers this (or build from src, which is the recommended method anyway)

for 4.0 it is in the man trunk, but again no idea (or congtrol) over conda from here.

(i think, you have to discuss all of this with the package maintainers, not us)

berak gravatar imageberak ( 2018-11-22 07:39:27 -0600 )edit