# Nearest Neighbors Meet Deep Neural Networks for Point Cloud Analysis

## TL;DR

This paper introduces SN-Adapter, a plug-and-play module that enhances 3D point cloud neural networks by leveraging nearest neighbor retrieval of spatial prototypes, improving performance without redesigning models.

## Contribution

The paper proposes SN-Adapter, a novel non-parametric method that boosts existing 3D neural networks using spatial prototypes and k-NN retrieval, applicable across multiple tasks.

## Key findings

- Improves 3D network performance without extra parameters
- Effective across shape classification, segmentation, and detection
- Demonstrates robustness and generalization across tasks

## Abstract

Performances on standard 3D point cloud benchmarks have plateaued, resulting in oversized models and complex network design to make a fractional improvement. We present an alternative to enhance existing deep neural networks without any redesigning or extra parameters, termed as Spatial-Neighbor Adapter (SN-Adapter). Building on any trained 3D network, we utilize its learned encoding capability to extract features of the training dataset and summarize them as prototypical spatial knowledge. For a test point cloud, the SN-Adapter retrieves k nearest neighbors (k-NN) from the pre-constructed spatial prototypes and linearly interpolates the k-NN prediction with that of the original 3D network. By providing complementary characteristics, the proposed SN-Adapter serves as a plug-and-play module to economically improve performance in a non-parametric manner. More importantly, our SN-Adapter can be effectively generalized to various 3D tasks, including shape classification, part segmentation, and 3D object detection, demonstrating its superiority and robustness. We hope our approach could show a new perspective for point cloud analysis and facilitate future research.