Skip to content
All projects

Case study

SmoothPose

Built an end-to-end 6-DoF object-pose tracking pipeline for four rigid objects across 1,404 frames using monocular point clouds, segmentation masks, camera intrinsics, and template meshes.

Mesh-rendered pose overlay · Suitcase
Contribution
Built and evaluated an end-to-end pose-tracking pipeline using monocular point clouds, segmentation masks, camera intrinsics, and template meshes.

Context

The problem

Track the full 6-DoF pose of four rigid objects across a monocular sequence despite occlusion, symmetry, and reconstruction outliers.

System

Architecture

  • 1,404-frame monocular sequence
  • Point clouds and segmentation masks
  • Camera intrinsics and template meshes
  • Geometric and temporal pose optimization
  • Mesh-rendered overlays and per-frame evaluation

Engineering judgment

Technical decisions

Compare geometric and temporal tracking approaches

Implemented and evaluated six approaches, including 24-start ICP, robust sequential tracking, quaternion smoothing, dynamic programming, and joint optimization.

Design around difficult object geometry

The pipeline explicitly addresses occlusion, symmetry, and reconstruction outliers that can destabilize object-pose estimates.

Evaluate stability as well as geometric fit

Measured silhouette IoU, Chamfer distance, and rotation-jump behavior alongside mesh-rendered overlay inspection.

Constraints

Challenges

  • Occlusion can remove discriminative geometry from an individual frame.
  • Symmetric objects can produce plausible but incorrect rotations.
  • Reconstruction outliers can destabilize direct geometric alignment.

Result

Outcome

Reached 0.693 mean silhouette IoU and 3.12 cm mean Chamfer distance without rotation jumps above 60°.