NPNet: A Non-Parametric Network with Adaptive Gaussian-Fourier Positional Encoding for 3D Classification and Segmentation

TL;DR

NPNet introduces a non-parametric 3D point-cloud classification and segmentation method using adaptive Gaussian-Fourier positional encoding, achieving competitive results with efficient memory and inference time, especially in few-shot scenarios.

Contribution

It proposes a novel non-parametric approach with adaptive Gaussian-Fourier encoding for stable, scalable 3D point-cloud analysis without learned weights.

Findings

Strong performance on ModelNet40, ModelNet-R, ScanObjectNN, and ShapeNetPart.

Effective in few-shot learning scenarios.

Favorable memory and inference efficiency.

Abstract

We present NPNet, a fully non-parametric approach for 3D point-cloud classification and part segmentation. NPNet contains no learned weights; instead, it builds point features using deterministic operators such as farthest point sampling, k-nearest neighbors, and pooling. Our key idea is an adaptive Gaussian-Fourier positional encoding whose bandwidth and Gaussian-cosine mixing are chosen from the input geometry, helping the method remain stable across different scales and sampling densities. For segmentation, we additionally incorporate fixed-frequency Fourier features to provide global context alongside the adaptive encoding. Across ModelNet40/ModelNet-R, ScanObjectNN, and ShapeNetPart, NPNet achieves strong performance among non-parametric baselines, and it is particularly effective in few-shot settings on ModelNet40. NPNet also offers favorable memory use and inference time compared to prior non-parametric methods