Fix broken tracking issue in Pose3DInferencer — previous results neve…

[isjackwild]

…r being updated

Motivation

I am looking into 3D pose detection in video, and pose-sequence tracking over time. I came across an issue in the Pose3DInferencer whereby poses were always assigned a new track_id on every frame. This is because self._buffer['results_pose2d_last'] was set to the same value as their input to the tracking function, and never updated with the current frames results at the end of the loop

Modification

After tracking has been completed in the preprocess_single method, I saved the results of the current frame to self._buffer['results_pose2d_last'] = results_pose2d. Now pose tracking is working between frames.

BC-breaking (Optional)

I do not believe this affects backwards compatibility

Use cases (Optional)

N/A

Checklist

Before PR:

• I have read and followed the workflow indicated in the CONTRIBUTING.md to create this PR.
• Pre-commit or linting tools indicated in CONTRIBUTING.md are used to fix the potential lint issues.
• Bug fixes are covered by unit tests, the case that causes the bug should be added in the unit tests.
• New functionalities are covered by complete unit tests. If not, please add more unit tests to ensure correctness.
• The documentation has been modified accordingly, including docstring or example tutorials.

After PR:

• CLA has been signed and all committers have signed the CLA in this PR.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.

---

Jack Wild seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}