Learning 3D-3D Correspondences for One-shot Partial-to-partial Registration
This addresses the challenge of registering incomplete 3D point clouds for applications like robotics and computer vision, offering a more efficient and accurate solution than existing iterative methods.
The paper tackles the problem of partial-to-partial 3D-3D registration by proposing a one-shot learning-based method that reduces network complexity and increases accuracy, outperforming state-of-the-art techniques on standard benchmarks with better robustness and generalization.
While 3D-3D registration is traditionally tacked by optimization-based methods, recent work has shown that learning-based techniques could achieve faster and more robust results. In this context, however, only PRNet can handle the partial-to-partial registration scenario. Unfortunately, this is achieved at the cost of relying on an iterative procedure, with a complex network architecture. Here, we show that learning-based partial-to-partial registration can be achieved in a one-shot manner, jointly reducing network complexity and increasing registration accuracy. To this end, we propose an Optimal Transport layer able to account for occluded points thanks to the use of outlier bins. The resulting OPRNet framework outperforms the state of the art on standard benchmarks, demonstrating better robustness and generalization ability than existing techniques.