DeepDeform: Learning Non-rigid RGB-D Reconstruction with Semi-supervised Data

TL;DR

DeepDeform introduces a semi-supervised method and a large dataset for non-rigid RGB-D reconstruction, enabling a neural network to accurately handle large deformations and improve reconstruction quality.

Contribution

It presents a novel semi-supervised strategy for dense inter-frame correspondence and a neural network for robust non-rigid feature matching in reconstruction.

Findings

Outperforms existing non-rigid reconstruction methods.

Creates a large dataset of 400 scenes with over 390,000 frames.

Achieves more accurate reconstructions under large deformations.

Abstract

Applying data-driven approaches to non-rigid 3D reconstruction has been difficult, which we believe can be attributed to the lack of a large-scale training corpus. Unfortunately, this method fails for important cases such as highly non-rigid deformations. We first address this problem of lack of data by introducing a novel semi-supervised strategy to obtain dense inter-frame correspondences from a sparse set of annotations. This way, we obtain a large dataset of 400 scenes, over 390,000 RGB-D frames, and 5,533 densely aligned frame pairs; in addition, we provide a test set along with several metrics for evaluation. Based on this corpus, we introduce a data-driven non-rigid feature matching approach, which we integrate into an optimization-based reconstruction pipeline. Here, we propose a new neural network that operates on RGB-D frames, while maintaining robustness under large non-rigid deformations and producing accurate predictions. Our approach significantly outperforms existing non-rigid reconstruction methods that do not use learned data terms, as well as learning-based approaches that only use self-supervision.