mediapipe

mediapipe copied to clipboard

MediaPipe Solution (you are using)

FaceLandmarker

Programming language

Python

Are you willing to contribute it

None

Describe the feature and the current behaviour/state

Currently, whenever I want to analyze to videos in video mode I need to recreate the FaceLandmarker object from scratch (as the timestamps of each video need to be monotonically increasing). This have the price of spending time on reloading the model back from memory. I suggest having a reset function (or some other interface) that will let me reset the internal status of the FaceLandmarker without reloading the whole model.

Will this change the current API? How?

No response

Who will benefit with this feature?

No response

Please specify the use cases for this feature

When analyzing multiple videos one after the other, while wanting to spend minimal time at reloading data from memory.

Any Other info

No response