# Revisiting Radar Perception With Spectral Point Clouds

arXiv: https://arxiv.org/abs/2604.08282 · HTML: https://mlsanity.com/paper/2604.08282/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 28 of its 37 references that are in the map.
- First posted on arXiv 9 Apr 2026.

## Builds on (most cited first, 28 of 28)

- [PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation](https://mlsanity.com/paper/1612.00593/) (2016)
- [Gaussian Error Linear Units (GELUs)](https://mlsanity.com/paper/1606.08415/) (2016)
- [PointPillars: Fast Encoders for Object Detection from Point Clouds](https://mlsanity.com/paper/1812.05784/) (2018)
- [Point Transformer V3: Simpler, Faster, Stronger](https://mlsanity.com/paper/2312.10035/) (2023)
- [PIXOR: Real-time 3D Object Detection from Point Clouds](https://mlsanity.com/paper/1902.06326/) (2019)
- [K-Radar: 4D Radar Object Detection for Autonomous Driving in Various Weather Conditions](https://mlsanity.com/paper/2206.08171/) (2022)
- [A Tutorial on Kernel Density Estimation and Recent Advances](https://mlsanity.com/paper/1704.03924/) (2017)
- [TJ4DRadSet: A 4D Radar Dataset for Autonomous Driving](https://mlsanity.com/paper/2204.13483/) (2022)
- [Raw High-Definition Radar for Multi-Task Learning](https://mlsanity.com/paper/2112.10646/) (2021)
- [CARRADA Dataset: Camera and Automotive Radar with Range-Angle-Doppler Annotations](https://mlsanity.com/paper/2005.01456/) (2020)
- [RADDet: Range-Azimuth-Doppler based Radar Object Detection for Dynamic Road Users](https://mlsanity.com/paper/2105.00363/) (2021)
- [SMURF: Spatial Multi-Representation Fusion for 3D Object Detection with 4D Imaging Radar](https://mlsanity.com/paper/2307.10784/) (2023)
- [MAN TruckScenes: A multimodal dataset for autonomous trucking in diverse conditions](https://mlsanity.com/paper/2407.07462/) (2024)
- [Dual Radar: A Multi-modal Dataset with Dual 4D Radar for Autonomous Driving](https://mlsanity.com/paper/2310.07602/) (2023)
- [Probabilistic Oriented Object Detection in Automotive Radar](https://mlsanity.com/paper/2004.05310/) (2020)
- [T-FFTRadNet: Object Detection with Swin Vision Transformers from Raw ADC Radar Signals](https://mlsanity.com/paper/2303.16940/) (2023)
- [Rethinking of Radar's Role: A Camera-Radar Dataset and Systematic Annotator via Coordinate Alignment](https://mlsanity.com/paper/2105.05207/) (2021)
- [RadarPillars: Efficient Object Detection from 4D Radar Point Clouds](https://mlsanity.com/paper/2408.05020/) (2024)
- [MVFAN: Multi-View Feature Assisted Network for 4D Radar Object Detection](https://mlsanity.com/paper/2310.16389/) (2023)
- [Deep Open Space Segmentation using Automotive Radar](https://mlsanity.com/paper/2004.03449/) (2020)
- [A Deep Automotive Radar Detector using the RaDelft Dataset](https://mlsanity.com/paper/2406.04723/) (2024)
- [TransRadar: Adaptive-Directional Transformer for Real-Time Multi-View Radar Semantic Segmentation](https://mlsanity.com/paper/2310.02260/) (2023)
- [TransRAD: Retentive Vision Transformer for Enhanced Radar Object Detection](https://mlsanity.com/paper/2501.17977/) (2025)
- [Towards Foundational Models for Single-Chip Radar](https://mlsanity.com/paper/2509.12482/) (2025)
- [SparseRadNet: Sparse Perception Neural Network on Subsampled Radar Data](https://mlsanity.com/paper/2406.10600/) (2024)
- [Vision meets mmWave Radar: 3D Object Perception Benchmark for Autonomous Driving](https://mlsanity.com/paper/2311.10261/) (2023)
- [ERASE-Net: Efficient Segmentation Networks for Automotive Radar Signals](https://mlsanity.com/paper/2209.12940/) (2022)
- [Transmit Beamspace DDMA Based Automotive MIMO Radar](https://mlsanity.com/paper/2103.13892/) (2021)

## Abstract (arXiv)

Radar perception models are trained with different inputs, from range-Doppler spectra to sparse point clouds. Dense spectra are assumed to outperform sparse point clouds, yet they can vary considerably across sensors and configurations, which hinders transfer. In this paper, we provide alternatives for incorporating spectral information into radar point clouds and show that, point clouds need not underperform compared to spectra. We introduce the spectral point cloud paradigm, where point clouds are treated as sparse, compressed representations of the radar spectra, and argue that, when enriched with spectral information, they serve as strong candidates for a unified input representation that is more robust against sensor-specific differences. We develop an experimental framework that compares spectral point cloud (PC) models at varying densities against a dense range-Doppler (RD) benchmark, and report the density levels where the PC configurations meet the performance of the RD benchmark. Furthermore, we experiment with two basic spectral enrichment approaches, that inject additional target-relevant information into the point clouds. Contrary to the common belief that the dense RD approach is superior, we show that point clouds can do just as well, and can surpass the RD benchmark when enrichment is applied. Spectral point clouds can therefore serve as strong candidates for unified radar perception, paving the way for future radar foundation models.