search

google-ai-edge / mediapipe

Cross-platform, customizable ML solutions for live and streaming media.
https://ai.google.dev/edge/mediapipe
Apache License 2.0
27.8k stars 5.18k forks source link

Convert Existing 2D TF Object Detection API Model into 2-stage Objectron Graph #1295

Closed kells1986 closed 3 years ago

kells1986 commented 4 years ago

I have an object detection model trained on custom classes using the tensorflow object detection api.

Is there any way to convert that model into the 2-stage objectron format? or is further training needed specifically for the 3d fitting?

It seems to me that those two things are distinct processes, and I wondered if there was a tool that could create a 2-stage TFLite file using a custom TFLite / frozen inference graph 2d object detection model?

matteo-filippini commented 4 years ago

I'm also interested in some documentation about that. I guess this is not so straightforward. Browsing the objectron example folders you can find some obj.uuu files and textures. These are 3d models of sneakers/cups/cameras etc, maybe they are used to fit the 2d image to 3d model and infer the volume of the object? would be great get some advices from developers!

sgowroji commented 3 years ago

Hi @kells1986, Could you please go through our latest documentation regarding your above query and let us know, if you have a different use case.

google-ml-butler[bot] commented 3 years ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you.

google-ml-butler[bot] commented 3 years ago

Closing as stale. Please reopen if you'd like to work on this further.

google-ml-butler[bot] commented 3 years ago

Are you satisfied with the resolution of your issue? Yes No