google-ai-edge / mediapipe

Cross-platform, customizable ML solutions for live and streaming media.
https://ai.google.dev/edge/mediapipe
Apache License 2.0
27.07k stars 5.11k forks source link

Pose landmarks on Android #3164

Closed vidheyoza closed 1 year ago

vidheyoza commented 2 years ago

Please make sure that this is a solution issue.

System information (Please provide as much relevant information as possible)

  • Have I written custom code (as opposed to using a stock example script provided in Mediapipe): Yes
  • OS Platform and Distribution (e.g., Linux Ubuntu 16.04, Android 11, iOS 14.4): Android 11
  • MediaPipe version: latest git clone (v0.8.9)
  • Bazel version: 4.2.1
  • Solution (e.g. FaceMesh, Pose, Holistic): Pose
  • Programming Language and version ( e.g. C++, Python, Java): Java and Kotlin

Describe the expected behavior:

I am trying to use the pose landmarks model in my custom Android project, but the landmarks model has multiple output nodes on the network, and it is not clear how to get the right output stream in the Android project as the TF Lite support on Android just mentions the input and output as Object instances, and I could not find documentation on using specific models on the Android platform. There are sample projects provided for Face Mesh recently on Maven, but couldn't use similar techniques for Pose without significant changes.

sureshdagooglecom commented 2 years ago

Hi @vidheyoza , can you explain use case w.r.t above issue.

vidheyoza commented 2 years ago

I have a simple app that gets a bitmap image data from a video stream, and I have a simple TF Lite model that gives output the keypoints matrix directly, which I then use to show it on the screen and perform different operations like finding angles between vectors.

I want to use BlazePose from MediaPipe similarly, and the documentation of the model itself and the paper shows a similar output matrix. But the model itself has multiple output streams and it is not clear how that should be used to get the pose matrix for a provided image.

google-ml-butler[bot] commented 2 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.

vidheyoza commented 2 years ago

@sureshdagooglecom can you please answer this question? Seems like the repo bot will mark it closed soon due to no activity.

tulanglaoli commented 2 years ago

you can learn from android example from source.

vidheyoza commented 2 years ago

Could you please link that to me? I've been trying to find that but there's a lot of models and there's no source example I could find for pose estimation.

google-ml-butler[bot] commented 2 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.

vidheyoza commented 2 years ago

@ivan-grishchenko can you please answer this question? Seems like the repo bot will mark it closed soon due to no activity.

kuaashish commented 1 year ago

Hello @vidheyoza, We are upgrading the MediaPipe Legacy Solutions to new MediaPipe solutions However, the libraries, documentation, and source code for all the MediapPipe Legacy Solutions will continue to be available in our GitHub repository and through library distribution services, such as Maven and NPM.

You can continue to use those legacy solutions in your applications if you choose. Though, we would request you to check new MediaPipe solutions which can help you more easily build and customize ML solutions for your applications. These new solutions will provide a superset of capabilities available in the legacy solutions.

github-actions[bot] commented 1 year ago

This issue has been marked stale because it has no recent activity since 7 days. It will be closed if no further activity occurs. Thank you.

github-actions[bot] commented 1 year ago

This issue was closed due to lack of activity after being marked stale for past 7 days.

google-ml-butler[bot] commented 1 year ago

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