Iterative Optimisation with an Innovation CNN for Pose Refinement

TL;DR

This paper introduces an Innovation CNN that refines object pose estimation iteratively without needing textured 3D models, achieving state-of-the-art results on LINEMOD datasets.

Contribution

The proposed Innovation CNN enables pose refinement without textured 3D models, improving accuracy through iterative stochastic gradient descent.

Findings

Achieves state-of-the-art accuracy on LINEMOD datasets.

Does not require textured 3D models for pose refinement.

Improves initial pose estimates progressively through iterative refinement.

Abstract

Object pose estimation from a single RGB image is a challenging problem due to variable lighting conditions and viewpoint changes. The most accurate pose estimation networks implement pose refinement via reprojection of a known, textured 3D model, however, such methods cannot be applied without high quality 3D models of the observed objects. In this work we propose an approach, namely an Innovation CNN, to object pose estimation refinement that overcomes the requirement for reprojecting a textured 3D model. Our approach improves initial pose estimation progressively by applying the Innovation CNN iteratively in a stochastic gradient descent (SGD) framework. We evaluate our method on the popular LINEMOD and Occlusion LINEMOD datasets and obtain state-of-the-art performance on both datasets.