You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
scikit-video should not be used for reading the videos as the development/maintenance of the library seems to be quite stagnant (based on the repository's activity)
Instead, after considering a few alternatives, ImageIO seems to be suitable for usage, since most of the functionality provided by scikit-video is available in ImageIO (not directly though). This would help in making the transition simple and more efficient (hopefully).
Further, the reason for this change will also be justified by relevant tests and performance comparisons.
Any suggestions/improvements are welcome 😄
The text was updated successfully, but these errors were encountered:
As of now, ImageIO is not able to provide better performance results (particularly when resizing videos and converting them to grayscale). ImageIO also uses ffmpeg as their backend, and are planning to improve their plugin/create a new one for reading/writing videos.
So, work on this issue will begin after this plugin is introduced in the upcoming release of ImageIO.
The new backend used would be ffmpeg-python.
Basically, mydia will act as a wrapper around the "filter" functionality of ffmepg-python for frame extraction, grayscale conversion, and frame resizing.
Also, after conducting few internal tests, this approach is ~6x faster that scikit-video (especially when the number of frames to be extracted is relatively small).
The next release (mydia 2.0.0) will use ffmpeg-python as its backend.
scikit-video should not be used for reading the videos as the development/maintenance of the library seems to be quite stagnant (based on the repository's activity)
Instead, after considering a few alternatives, ImageIO seems to be suitable for usage, since most of the functionality provided by scikit-video is available in ImageIO (not directly though). This would help in making the transition simple and more efficient (hopefully).
Further, the reason for this change will also be justified by relevant tests and performance comparisons.
Any suggestions/improvements are welcome 😄
The text was updated successfully, but these errors were encountered: