Mediapipe: Hello, in the android gesture recognition, I want to use your own video YUV data is passed to the mediapipe, instead of using the data of mobile phone camera, I need how to modify com.google.mediapipe.com ponents package of files. Please help me, thank you.

[Sanerly]

[eknight7]

What is the source of the YUV data? Is it from a different camera or a video file?

[afsaredrisy]

I am also looking for similar solution, Its looking they added support for YUV frames in image_transformation_calculator.cc calculator but there is no method available in AndroidPacketCreator or in PacketCreator classes to create YUVBuffer. So I convert YUV2RGB using shader and render to a texture of FrameBuffer. I used GL_TEXTURE_2D instead of GL_TEXTURE_EXTERNAL_OES in TextureRenderer, Now I can pass this texture to FrameProcessor.onNewFrame(final TextureFrame frame) method.
Other things will be similar now.

[Sanerly]

I am also looking for similar solution, Its looking they added support for YUV frames in image_transformation_calculator.cc calculator but there is no method available in AndroidPacketCreator or in PacketCreator classes to create YUVBuffer. So I convert YUV2RGB using shader and render to a texture of FrameBuffer. I used GL_TEXTURE_2D instead of GL_TEXTURE_EXTERNAL_OES in TextureRenderer, Now I can pass this texture to FrameProcessor.onNewFrame(final TextureFrame frame) method.
Other things will be similar now.

Thanks for your help. I am busy with other things these days. I will try the way you said.

[Sanerly]

What is the source of the YUV data? Is it from a different camera or a video file?

Hello, I used my own camera to transmit the data, which is currently in YUV format