I’m building an AR application where I need to estimate the pose of a physical object and overlay a 3D model of the same object on top of it, with tracking.

I’m considering using SIFT (Scale-Invariant Feature Transform) for pose estimation, but I’m unsure how to properly overlay the 3D model onto the object and keep it aligned as the object moves. I’m also aware that object detection doesn’t directly provide pose information, which I need for this task.

Here are the challenges I’m facing:

1. Pose Estimation: How can I use SIFT (or an alternative method) to accurately estimate the object’s 3D pose relative to the camera?

2. Overlaying the 3D Model: Once I have the pose, how can I project a 3D model onto the object and ensure it stays aligned, even as the object moves?

3. Tracking: How can I track the object so the 3D overlay remains stable in real-time?

I’m looking for any techniques, libraries, or frameworks that can help with pose estimation, 3D overlay, and tracking.