Documentation for building pose-tracking iOS app using mediapipe library with swift

[diyoraharshitN]

Description of issue (what needs changing)

Hi Team, I couldn't find any guidelines/documentation for building pose-tracking iOS app using mediapipe library with swift

Clear description

I am looking for Documentation for building pose-tracking iOS app using mediapipe library with swift.

Correct links

No response

Parameters defined

No response

Returns defined

No response

Raises listed and defined

No response

Usage example

https://developers.google.com/mediapipe/framework/getting_started/ios

Request visuals, if applicable

No response

Submit a pull request?

No response

[kuaashish]

Hi @diyoraharshitN,

At the moment, we don't have the pose landmarker available on the iOS platform, and there's no official documentation for its use in Swift. However, we do have iOS documentation for legacy pose tracking, and the iOS target is mentioned here. It's important to note that we are not offering support for legacy solutions, as our primary focus is on the new Task APIs. However, We are actively working to provide iOS-Swift support for other Task APIs as well.

Thank you

[diyoraharshitN]

Hi @kuaashish Thank you for your response Kindly look into this link it's not working iOS target that you have mentioned

also can you give me rough timeline so that we can plan accordingly for the pose tracking api documentation?

[diyoraharshitN]

There isn't any demo available for iOS click here

[dprichards-carespace]

Hi @diyoraharshitN,

At the moment, we don't have the pose landmarker available on the iOS platform, and there's no official documentation for its use in Swift. However, we do have iOS documentation for legacy pose tracking, and the iOS target is mentioned here. It's important to note that we are not offering support for legacy solutions, as our primary focus is on the new Task APIs. However, We are actively working to provide iOS-Swift support for other Task APIs as well.

Thank you

Do you have any updated timeline on this documentation?

[dprichards-carespace]

We want to use pose landmarker on the iOS platform, and there's no official documentation for its use in Swift. Do you have any updated timeline for this documentation?

[schmidt-sebastian]

@ktonthat

[kuaashish]

Hi @diyoraharshitN,

We have completed the code example for the new Pose Landmarker for iOS. You can view the example here. Feel free to try implementing it for now. However, please note that we are still working on the documentation, which will be included soon.

Thank you!!

[diyoraharshitN]

Hi @kuaashish,

Thank you for the update about example. Do you have any updated timeline on documentation?

[diyoraharshitN]

Hi @kuaashish

i have searched in the example but didn't find any seperetion for Left Body landmarks and Right Body landmarks points. where in web version in the node_modules @mediapipe>holistic>index.d.ts there is const declared for Left side landmark for pose as POSE_LANDMARKS_LEFT and for Right side landmark for pose as POSE_LANDMARKS_RIGHT.

is there any way to separate Left and Right landmarks from the MPPNormalizedLandmark array that we get as result array from the scan?

Thank you Harshit

[schmidt-sebastian]

cc @priankakariatyml

[diyoraharshitN]

Hi @kuaashish

I have found that there are 4 points 24 and 26 (right knee and right hip), 25 and 27(left foot and left knee) are missing from the ios Version of example in mediapipe.

Thank you

[diyoraharshitN]

Hi @kuaashish

i have searched in the example but didn't find any seperetion for Left Body landmarks and Right Body landmarks points. where in web version in the node_modules @mediapipe>holistic>index.d.ts there is const declared for Left side landmark for pose as POSE_LANDMARKS_LEFT and for Right side landmark for pose as POSE_LANDMARKS_RIGHT.

is there any way to separate Left and Right landmarks from the MPPNormalizedLandmark array that we get as result array from the scan?

Thank you Harshit

Hi @kuaashish @priankakariatyml

This is gentle reminder regarding clarification for the Ride side and Left side landmark points.

Thanks Harshit

[diyoraharshitN]

Hi @kuaashish

I have found that there are 4 points 24 and 26 (right knee and right hip), 25 and 27(left foot and left knee) are missing from the ios Version of example in mediapipe.

Thank you

This issue is resolve by using connection. Click Here

[kuaashish]

Hi @diyoraharshitN,

Thank you for confirming. Can we now proceed with closing the issue and marking it as resolved?

[diyoraharshitN]

Hi @diyoraharshitN,

Thank you for confirming. Can we now proceed with closing the issue and marking it as resolved?

Hi @kuaashish

1. Documentation is still not there for mediapipe iOS Pose landmark detection Click Here.

2. There is no seperation for Left Body Points and Right Body Points.I have already comment in this issue only Click Here

[diyoraharshitN]

Hi,

This is gentle reminder for the issue i have raised regarding

Documentation is still not there for mediapipe iOS Pose landmark detection Click Here.

There is no seperation for Left Body Points and Right Body Points.I have already comment in this issue only https://github.com/google/mediapipe/issues/4895#issuecomment-2051796753

[diyoraharshitN]

Hi @kuaashish @schmidt-sebastian,

any update regarding documentation and about these Comment?

[kuaashish]

Hi @diyoraharshitN,

Apologies for the delayed response. Currently, there is no update on the documentation. We are sharing this issue with our team again and marking it as a feature request to ensure it remains a priority. However, we cannot provide a timeline for this at the moment.

Thank you!!

[priankakariatyml]

Hi,

This is gentle reminder for the issue i have raised regarding

Documentation is still not there for mediapipe iOS Pose landmark detection Click Here.

There is no seperation for Left Body Points and Right Body Points.I have already comment in this issue only #4895 (comment)

Separation for left and right body points is not supported by the newer MediaPipe task APIs on any platforms. Only the legacy example for holistic landmarker has the provision for this.
You can consider defining this separation in your app based on this reference image for pose landmarks:

[diyoraharshit]

Hi @priankakariatyml

i have searched in the example but didn't find any seperetion for Left Body landmarks and Right Body landmarks points. where in web version in the node_modules @mediapipe>holistic>index.d.ts there is const declared for Left side landmark for pose as POSE_LANDMARKS_LEFT and for Right side landmark for pose as POSE_LANDMARKS_RIGHT.

is there any way to separate Left and Right landmarks from the MPPNormalizedLandmark array that we get as result array from the scan?

Thank you Harshit

[kuaashish]

Hi @diyoraharshit,

Documentation for implementing our Pose Landmarker in iOS is now available. Could you please try it and let us know if you encounter any further issues?

Thank you!!

[diyoraharshitN]

Hi @kuaashish

Thank you for the information. i will check it and let you know if there is any issue.

Thank you.

[kuaashish]

Hi @diyoraharshit,

Could you please confirm if we can mark this issue as resolved and close it, as the iOS documentation for the pose landmarker is now available?

Thank you!!

[github-actions[bot]]

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.