On my Yi cameras when a motion is detected both a six second low resolution 640x360 mp4 video clip and a 640x360 jpg picture, which seems to be from the second key frame which is 2 seconds into the clip, are uploaded to Yi server providing there’s a connection. This picture doesn’t always catch the action. If the video clip is used for AI detection would it not be better for Yi to scrap the firmware generated jpg, upload the video clip as usual and use the AI to produce a picture frame which best shows the action that cause the alert? Just a thought.
1 Like
Thanks for the feedback @anyone ! I have really appreciated your contribution to the forum this week. We appreciate you sharing your knowledge for our community