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
In this folder there are three scripts (3 steps) to preprocess zoom recordings where a single speaker is sharing their screen and webcam so that the recorded video file shows the screenshare+webcam on the same frame.
4
+
5
+
- Step1 trims the video (e.g. to remove introduction and post-talk discussions)
6
+
- Step2 extracts pngs from the video to identify the coordinates for the screenshare and the webcam view of the speaker.
7
+
- Step3 splits the zoom video into two sub-videos: the screenshare view and the webcam view, and then overlaps the webcam view on top of the screenshare on the top right corner.
8
+
9
+
Note: if the speaker is sharing something important in the top right corner, the webcam view will hide it.
# This step is needed to identify the coordinates of the screen-share and of the "webcam" of the speaker. It basically saves pngs for one frame for each video and then manually one has to identify the coordinates for step3
0 commit comments