DenseFusion: 6D Object Pose Estimation by Iterative Dense Fusion
This addresses the problem of accurate and efficient object pose estimation for robotics and augmented reality applications, representing a strong specific gain rather than a foundational breakthrough.
The paper tackles 6D object pose estimation from RGB-D images by introducing DenseFusion, a framework that processes RGB and depth data individually and fuses them with a dense network, integrating iterative refinement for improved accuracy and near real-time performance. It outperforms state-of-the-art methods on YCB-Video and LineMOD datasets, as demonstrated in experiments and real robot deployment.
A key technical challenge in performing 6D object pose estimation from RGB-D image is to fully leverage the two complementary data sources. Prior works either extract information from the RGB image and depth separately or use costly post-processing steps, limiting their performances in highly cluttered scenes and real-time applications. In this work, we present DenseFusion, a generic framework for estimating 6D pose of a set of known objects from RGB-D images. DenseFusion is a heterogeneous architecture that processes the two data sources individually and uses a novel dense fusion network to extract pixel-wise dense feature embedding, from which the pose is estimated. Furthermore, we integrate an end-to-end iterative pose refinement procedure that further improves the pose estimation while achieving near real-time inference. Our experiments show that our method outperforms state-of-the-art approaches in two datasets, YCB-Video and LineMOD. We also deploy our proposed method to a real robot to grasp and manipulate objects based on the estimated pose.