NAR-*ICP: Neural Execution of Classical ICP-based Pointcloud Registration Algorithms

TL;DR

This paper introduces NAR-*ICP, a neural network framework that learns to execute classical ICP-based point cloud registration algorithms, improving performance and generalization in robotics applications.

Contribution

It presents a novel GNN-based approach that learns to perform classical ICP algorithms, bridging neural networks and traditional robotics methods.

Findings

Outperforms baseline methods in real-world and synthetic datasets.

Demonstrates ability to generalize beyond traditional ICP algorithms.

Achieves superior accuracy in point cloud registration tasks.

Abstract

This study explores the intersection of neural networks and classical robotics algorithms through the Neural Algorithmic Reasoning (NAR) blueprint, enabling the training of neural networks to reason like classical robotics algorithms by learning to execute them. Algorithms are integral to robotics and safety-critical applications due to their predictable and consistent performance through logical and mathematical principles. In contrast, while neural networks are highly adaptable, handling complex, high-dimensional data and generalising across tasks, they often lack interpretability and transparency in their internal computations. To bridge the two, we propose a novel Graph Neural Network (GNN)-based framework, NAR-*ICP, that learns the intermediate computations of classical ICP-based registration algorithms, extending the CLRS Benchmark. We evaluate our approach across real-world and synthetic datasets, demonstrating its flexibility in handling complex inputs, and its potential to be used within larger learning pipelines. Our method achieves superior performance compared to the baselines, even surpassing the algorithms it was trained on, further demonstrating its ability to generalise beyond the capabilities of traditional algorithms.