Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Improving performance #1

Open
eried opened this issue Oct 23, 2020 · 2 comments
Open

Improving performance #1

eried opened this issue Oct 23, 2020 · 2 comments

Comments

@eried
Copy link

eried commented Oct 23, 2020

Thanks for the code! I had the idea to synchronize 30 or so dashcam videos I have with different weather. What approach would you recommend for improving the performance besides your recommendations?

It has been running for 10 hours :D in my surface pro and it is on the second frame:
image

So I was thinking on making the videos extremely small to use them as "media cache" and outputting just the frame numbers.

@owang
Copy link
Owner

owang commented Oct 23, 2020

Try to lower the fps, depending on the amount of motion in your videos you can likely get away with processing the videos at a much lower fps (e.g., something like 5). The second main thing that you should do is parallelize the sift feature matching, this should give a mostly linear speed improvement. Also, if your videos are very high res, you can resize them before processing. Finally, you can set the maximum number of SIFT features to find (nfeatures=500) or so.

@owang
Copy link
Owner

owang commented Oct 23, 2020

You might also want to try: https://github.com/cgtuebingen/LearningToSynchronizeVideos, which was designed specifically for the use case you describe.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants