On the Evaluation of RGB-D-based Categorical Pose and Shape Estimation
This work addresses evaluation challenges in computer vision for researchers, providing a toolbox for fair comparison, but it is incremental as it focuses on improving assessment rather than developing new methods.
The authors tackled the problem of evaluating RGB-D-based categorical pose and shape estimation methods by critically analyzing existing protocols and proposing new metrics and annotations for the Redwood dataset, finding that current methods generalize poorly to unconstrained orientations and are biased toward upright objects.
Recently, various methods for 6D pose and shape estimation of objects have been proposed. Typically, these methods evaluate their pose estimation in terms of average precision, and reconstruction quality with chamfer distance. In this work we take a critical look at this predominant evaluation protocol including metrics and datasets. 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. We find that existing methods do not generalize well to unconstrained orientations, and are actually heavily biased towards objects being upright. We contribute an easy-to-use evaluation toolbox with well-defined metrics, method and dataset interfaces, which readily allows evaluation and comparison with various state-of-the-art approaches (see https://github.com/roym899/pose_and_shape_evaluation ).