mlsanityData to 14 Oct 2026

RadarDistill: Boosting Radar-based Object Detection Performance via Knowledge Distillation from LiDAR Features

arXiv 2403.05061 ↗First posted 8 Mar 2024cs.CVcited by 28

On the map: Knowledge distillation → Detection and dense prediction distillation

Abstract (arXiv)

The inherent noisy and sparse characteristics of radar data pose challenges in finding effective representations for 3D object detection. In this paper, we propose RadarDistill, a novel knowledge distillation (KD) method, which can improve the representation of radar data by leveraging LiDAR data. RadarDistill successfully transfers desirable characteristics of LiDAR features into radar features using three key components: Cross-Modality Alignment (CMA), Activation-based Feature Distillation (AFD), and Proposal-based Feature Distillation (PFD). CMA enhances the density of radar features by employing multiple layers of dilation operations, effectively addressing the challenge of inefficient knowledge transfer from LiDAR to radar. AFD selectively transfers knowledge based on regions of the LiDAR features, with a specific focus on areas where activation intensity exceeds a predefined threshold. PFD similarly guides the radar network to selectively mimic features from the LiDAR network within the object proposals. Our comparative analyses conducted on the nuScenes datasets demonstrate that RadarDistill achieves state-of-the-art (SOTA) performance for radar-only object detection task, recording 20.5% in mAP and 43.7% in NDS. Also, RadarDistill significantly improves the performance of the camera-radar fusion model.

Key facts. Cited by 28 papers in the map (data through 14 Oct 2026). Placed in Detection and dense prediction distillation, one of 12 directions in Knowledge distillation. Builds on 40 of its 45 references that are in the map. First posted on arXiv 8 Mar 2024.

Builds on

40 of 45 references.

Closest ancestors · 12

Same direction, newest first

  1. 2023DistillBEV: Boosting Multi-Camera 3D Object Detection with Cross-Modal Knowledge Distillation26 Sep
  2. 2023Representation Disparity-aware Distillation for 3D Object Detection20 Aug
  3. 2023UniDistill: A Universal Cross-Modality Knowledge Distillation Framework for 3D Object Detection in Bird's-Eye View27 Mar
  4. 2023X3KD: Knowledge Distillation Across Modalities, Tasks and Stages for Multi-Camera 3D Object Detection3 Mar
  5. 2022BEVDistill: Cross-Modal BEV Distillation for Multi-View 3D Object Detection17 Nov
  6. 2022Cross-Modality Knowledge Distillation Network for Monocular 3D Object Detection14 Nov
  7. 2022Towards Efficient 3D Object Detection with Knowledge Distillation30 May
  8. 2022PointDistiller: Structured Knowledge Distillation Towards Efficient and Compact 3D Detection23 May
  9. 2022Prediction-Guided Distillation for Dense Object Detection10 Mar
  10. 2021Focal and Global Knowledge Distillation for Detectors23 Nov
  11. 2021Instance-Conditional Knowledge Distillation for Object Detection25 Oct
  12. 2021Distilling Object Detectors via Decoupled Features26 Mar

Other references · 28

  1. 2015Distilling the Knowledge in a Neural Networkcited by 10,324
  2. 2019nuScenes: A multimodal dataset for autonomous drivingcited by 4,523
  3. 2014FitNets: Hints for Thin Deep Netscited by 1,653
  4. 2017Deformable Convolutional Networkscited by 1,488
  5. 2019KPConv: Flexible and Deformable Convolution for Point Cloudscited by 1,205
  6. 2019Objects as Pointscited by 1,175
  7. 2020Center-based 3D Object Detection and Trackingcited by 1,144
  8. 2022BEVFusion: Multi-Task Multi-Sensor Fusion with Unified Bird's-Eye View Representationcited by 850
  9. 2023ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoderscited by 712
  10. 2021FCOS3D: Fully Convolutional One-Stage Monocular 3D Object Detectioncited by 447
  11. 2019Class-balanced Grouping and Sampling for Point Cloud 3D Object Detectioncited by 350
  12. 2020Point-GNN: Graph Neural Network for 3D Object Detection in a Point Cloudcited by 335
16 more
  1. 2018Knowledge Transfer via Distillation of Activation Boundaries Formed by Hidden Neuronscited by 238
  2. 2022Unifying Voxel-based Representation with Transformer for 3D Object Detectioncited by 202
  3. 2019Distilling Object Detectors with Fine-grained Feature Imitationcited by 189
  4. 2020CenterFusion: Center-based Radar and Camera Fusion for 3D Object Detectioncited by 161
  5. 2022PillarNet: Real-Time and High-Performance Pillar-based 3D Object Detectioncited by 99
  6. 2023CRN: Camera Radar Net for Accurate, Robust, Efficient 3D Perceptioncited by 82
  7. 2021LIGA-Stereo: Learning LiDAR Geometry Aware Representations for Stereo-based 3D Detectorcited by 70
  8. 2022MonoDistill: Learning Spatial Features for Monocular 3D Object Detectioncited by 68
  9. 2022CRAFT: Camera-Radar 3D Object Detection with Spatio-Contextual Fusion Transformercited by 67
  10. 2020RadarNet: Exploiting Radar for Robust Perception of Dynamic Objectscited by 48
  11. 2022Bridging the View Disparity Between Radar and Camera Features for Multi-modal Fusion 3D Object Detectioncited by 46
  12. 2020Distilling Object Detectors with Task Adaptive Regularizationcited by 36
  13. 2022Improved Orientation Estimation and Detection with Hybrid Object Detection Networks for Automotive Radarcited by 33
  14. 2023Bi-LRFusion: Bi-Directional LiDAR-Radar Fusion for 3D Dynamic Object Detectioncited by 28
  15. 2023RCM-Fusion: Radar-Camera Multi-Level Fusion for 3D Object Detectioncited by 23
  16. 2022Boosting 3D Object Detection by Simulating Multimodality on Point Cloudscited by 22

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 7; 2025: 15; 2026: 6.

28 papers.

Sort:

By area:

  1. 20253DRS: MLLMs Need 3D-Aware Representation Supervision for Scene Understandingcited by 70
  2. 2024LiCROcc: Teach Radar for Accurate Semantic Occupancy Prediction using LiDAR and Cameracited by 19
  3. 2024CRT-Fusion: Camera, Radar, Temporal Fusion Using Motion Information for 3D Object Detectioncited by 18
  4. 2024DSRC: Learning Density-insensitive and Semantic-aware Collaborative Representation against Corruptionscited by 13
  5. 2024Radar and Camera Fusion for Object Detection and Tracking: A Comprehensive Surveycited by 9
  6. 2025RICCARDO: Radar Hit Prediction and Convolution for Camera-Radar 3D Object Detectioncited by 9
  7. 2024LIX: Implicitly Infusing Spatial Geometric Prior Knowledge into Visual Semantic Segmentation for Autonomous Drivingcited by 8
  8. 2024SpaRC: Sparse Radar-Camera Fusion for 3D Object Detectioncited by 8
  9. 2024Radar Spectra-Language Model for Automotive Scene Parsingcited by 6
  10. 2025CORENet: Cross-Modal 4D Radar Denoising Network with LiDAR Supervision for Autonomous Drivingcited by 4
  11. 2025TARS: Traffic-Aware Radar Scene Flow Estimationcited by 3
  12. 2025LRFusionPR: A Polar BEV-Based LiDAR-Radar Fusion Network for Place Recognitioncited by 3
16 more
  1. 2025RCTDistill: Cross-Modal Knowledge Distillation Framework for Radar-Camera 3D Object Detection with Temporal Fusioncited by 3
  2. 2025CAE-DFKD: Bridging the Transferability Gap in Data-Free Knowledge Distillationcited by 2
  3. 2025SFGFusion: Surface Fitting Guided 3D Object Detection with 4D Radar and Camera Fusioncited by 2
  4. 2025RISE: Single Static Radar-based Indoor Scene Understandingcited by 2
  5. 2025Enhanced 3D Object Detection via Diverse Feature Representations of 4D Radar Tensorcited by 1
  6. 2025PAN: Pillars-Attention-Based Network for 3D Object Detectioncited by 1
  7. 2025REOcc: Camera-Radar Fusion with Radar Feature Enrichment for 3D Occupancy Predictioncited by 1
  8. 2025MORDA: A Synthetic Dataset to Facilitate Adaptation of Object Detectors to Unseen Real-target Domain While Preserving Performance on Real-source Domaincited by 0
  9. 2025MSDNet: Efficient 4D Radar Super-Resolution via Multi-Stage Distillationcited by 0
  10. 2025CF-Net: A Cross-Feature Reconstruction Network for High-Accuracy 1-Bit Target Classificationcited by 0
  11. 2026SkeFi: Cross-Modal Knowledge Transfer for Wireless Skeleton-Based Action Recognitioncited by 0
  12. 2026CD-FKD: Cross-Domain Feature Knowledge Distillation for Robust Single-Domain Generalization in Object Detectioncited by 0
  13. 20264DRaL: Bridging 4D Radar with LiDAR for Place Recognition using Knowledge Distillationcited by 0
  14. 2026CLLAP: Contrastive Learning-based LiDAR-Augmented Pretraining for Enhanced Radar-Camera Fusioncited by 0
  15. 2026Spatially Stratified Distillation for Heterogeneous Radar Place Recognitioncited by 0
  16. 2026If It Moves, Radar Knows: A Physics-Aware Radar Transformer for Class-Agnostic Moving-Object Detectioncited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025IMKD: Intensity-Aware Multi-Level Knowledge Distillation for Camera-Radar Fusion
  2. 2024SCKD: Semi-Supervised Cross-Modality Knowledge Distillation for 4D Radar Object Detection
  3. 2024CRKD: Enhanced Camera-Radar Object Detection with Cross-modality Knowledge Distillation
  4. 2024LEROjD: Lidar Extended Radar-Only Object Detection
  5. 2024LabelDistill: Label-guided Cross-modal Knowledge Distillation for Camera-based 3D Object Detection
  6. 2023Efficient Object Detection in Optical Remote Sensing Imagery via Attention-based Feature Distillation
  7. 2022BEV-LGKD: A Unified LiDAR-Guided Knowledge Distillation Framework for BEV 3D Object Detection
  8. 2022LiDAR Distillation: Bridging the Beam-Induced Domain Gap for 3D Object Detection