Mediapipe infer multiple images at the same time

[aleshem]

MediaPipe Solution (you are using)

Selfie Segmentor tflite

Programming language

Android

Are you willing to contribute it

No

Describe the feature and the current behaviour/state

Currenly the API supports inference of a single image at a time. I am getting multiple video streams and would like to infer them with a batch as an input

Will this change the current API? How?

instead of 
{height, width, channels_preserved} 
wil change to 
{n, height, width, channels_preserved}

Who will benefit with this feature?

Stereo computer vision, Multi video computer vision, multi input channels

Please specify the use cases for this feature

Anyone treating multiple video streams (stereo, different sources, etc.)
May improve running speed (depending on CPU/GPU/platform)

Any Other info

No response

[kuaashish]

Hello @aleshem,

Thank you for detailed explanation about the feature request. We will discuss it internally and based on discussion will prioritise it in our roadmap. However, For now, We can not provide the any specific timeline for this request.