LatentFusion: End-to-End Differentiable Reconstruction and Rendering for Unseen Object Pose Estimation
This addresses the scalability issue in robotics and AR/VR by enabling pose estimation for new objects without retraining.
The paper tackles the problem of 6D object pose estimation for unseen objects without requiring 3D models or additional training per object, achieving competitive performance to supervised methods on datasets like MOPED, ModelNet, and LINEMOD.
Current 6D object pose estimation methods usually require a 3D model for each object. These methods also require additional training in order to incorporate new objects. As a result, they are difficult to scale to a large number of objects and cannot be directly applied to unseen objects. We propose a novel framework for 6D pose estimation of unseen objects. We present a network that reconstructs a latent 3D representation of an object using a small number of reference views at inference time. Our network is able to render the latent 3D representation from arbitrary views. Using this neural renderer, we directly optimize for pose given an input image. By training our network with a large number of 3D shapes for reconstruction and rendering, our network generalizes well to unseen objects. We present a new dataset for unseen object pose estimation--MOPED. We evaluate the performance of our method for unseen object pose estimation on MOPED as well as the ModelNet and LINEMOD datasets. Our method performs competitively to supervised methods that are trained on those objects. Code and data is available at https://keunhong.com/publications/latentfusion/.