no keypoints in the frame

[simonetortorella]

I was testing the package with different types of videos. The entire structure fails if it does not detect at least one keypoint, such as when there is no subject in the video or when the subject temporarily leaves the camera's view. I tried modifying some lines, but it seems that not only the output videos but also the output data are corrupted. What changes should I make?

[chingswy]

Hello, you should make sure that all views have at least one people.

[simonetortorella]

Ok, but if the subject goes passes through the camera , like opening a door and entering the room, easymocap does not work properly, crashes instantaneously.

[raflitp]

Of course, that's why you need a bunch of cameras if there's some props blocking one of the cameras. If there's no person it will crash obviously because there's no person to detect and predict

[simonetortorella]

Ok, but I expected that in a 30 second video in which the subject is visible for 10 seconds, the entire video would be analyzed with consequent restitution of the 30 seconds and rendering only when the keypoints are found (i.e. in the 10 seconds in which the subject is present), without giving an error from the beginning, as soon as the keypoints are not identified in frame 0. But if it's an easymocap specification, that's fine, thanks.