SurfEmb: Dense and Continuous Correspondence Distributions for Object Pose Estimation with Learnt Surface Embeddings
This addresses the challenge of accurate object pose estimation for robotics and computer vision, offering a novel unsupervised approach that outperforms methods trained on real data.
The paper tackles the problem of 6D object pose estimation by learning dense, continuous 2D-3D correspondence distributions from data without prior knowledge of visual ambiguities like symmetry, and it significantly improves state-of-the-art on the BOP Challenge using only synthetic data.
We present an approach to learn dense, continuous 2D-3D correspondence distributions over the surface of objects from data with no prior knowledge of visual ambiguities like symmetry. We also present a new method for 6D pose estimation of rigid objects using the learnt distributions to sample, score and refine pose hypotheses. The correspondence distributions are learnt with a contrastive loss, represented in object-specific latent spaces by an encoder-decoder query model and a small fully connected key model. Our method is unsupervised with respect to visual ambiguities, yet we show that the query- and key models learn to represent accurate multi-modal surface distributions. Our pose estimation method improves the state-of-the-art significantly on the comprehensive BOP Challenge, trained purely on synthetic data, even compared with methods trained on real data. The project site is at https://surfemb.github.io/ .