SSD-6D: Making RGB-based 3D detection and 6D pose estimation great again

TL;DR

This paper introduces SSD-6D, a fast and effective RGB-based method for 3D object detection and 6D pose estimation, trained solely on synthetic data, outperforming many RGB-D methods.

Contribution

It extends SSD to cover 6D pose estimation from RGB images and trains exclusively on synthetic data, achieving competitive accuracy and high speed.

Findings

Outperforms state-of-the-art RGB-D methods on multiple datasets

Operates at around 10Hz, enabling real-time applications

Achieves high accuracy using only synthetic training data

Abstract

We present a novel method for detecting 3D model instances and estimating their 6D poses from RGB data in a single shot. To this end, we extend the popular SSD paradigm to cover the full 6D pose space and train on synthetic model data only. Our approach competes or surpasses current state-of-the-art methods that leverage RGB-D data on multiple challenging datasets. Furthermore, our method produces these results at around 10Hz, which is many times faster than the related methods. For the sake of reproducibility, we make our trained networks and detection code publicly available.