Leveraging Cycle-Consistent Anchor Points for Self-Supervised RGB-D Registration
This addresses scene geometric reasoning for applications like robotics or AR, but it is incremental as it builds on existing self-supervised methods.
The paper tackles RGB-D registration by using cycle-consistent keypoints and a novel pose block to improve correspondence accuracy, achieving state-of-the-art results on ScanNet and 3DMatch datasets.
With the rise in consumer depth cameras, a wealth of unlabeled RGB-D data has become available. This prompts the question of how to utilize this data for geometric reasoning of scenes. While many RGB-D registration meth- ods rely on geometric and feature-based similarity, we take a different approach. We use cycle-consistent keypoints as salient points to enforce spatial coherence constraints during matching, improving correspondence accuracy. Additionally, we introduce a novel pose block that combines a GRU recurrent unit with transformation synchronization, blending historical and multi-view data. Our approach surpasses previous self- supervised registration methods on ScanNet and 3DMatch, even outperforming some older supervised methods. We also integrate our components into existing methods, showing their effectiveness.