RGB-D-Based Categorical Object Pose and Shape Estimation: Methods, Datasets, and Evaluation

TL;DR

This paper reviews current RGB-D-based methods for object pose and shape estimation at the category level, critically evaluates existing evaluation protocols, and introduces new metrics, annotations, and a benchmarking toolbox for fair comparison.

Contribution

It provides a comprehensive overview, proposes improved evaluation metrics, adds new dataset annotations, and offers a standardized toolbox for assessing state-of-the-art methods.

Findings

Existing methods struggle with unconstrained orientations.

Methods are biased towards upright objects.

New evaluation metrics improve assessment accuracy.

Abstract

Recently, various methods for 6D pose and shape estimation of objects at a per-category level have been proposed. This work provides an overview of the field in terms of methods, datasets, and evaluation protocols. First, an overview of existing works and their commonalities and differences is provided. Second, we take a critical look at the predominant evaluation protocol, including metrics and datasets. Based on the findings, we propose a new set of metrics, contribute new annotations for the Redwood dataset, and evaluate state-of-the-art methods in a fair comparison. The results indicate that existing methods do not generalize well to unconstrained orientations and are actually heavily biased towards objects being upright. We provide an easy-to-use evaluation toolbox with well-defined metrics, methods, and dataset interfaces, which allows evaluation and comparison with various state-of-the-art approaches (https://github.com/roym899/pose_and_shape_evaluation).