# FreqKD: Frequency-Decoupled Cross-Modal Knowledge Distillation for Infrared Object Detection

arXiv: https://arxiv.org/abs/2606.11572 · HTML: https://mlsanity.com/paper/2606.11572/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Detection and dense prediction distillation, one of 12 directions in Knowledge distillation.
- Builds on 38 of its 48 references that are in the map.
- First posted on arXiv 10 Jun 2026.

## Placed in

- Knowledge distillation / [Detection and dense prediction distillation](https://mlsanity.com/distill/detection-and-dense-prediction-distillation/)

## Builds on (most cited first, 30 of 38)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks](https://mlsanity.com/paper/1506.01497/) (2015)
- [Segment Anything](https://mlsanity.com/paper/2304.02643/) (2023)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Sigmoid Loss for Language Image Pre-Training](https://mlsanity.com/paper/2303.15343/) (2023)
- [Domain-Adversarial Training of Neural Networks](https://mlsanity.com/paper/1505.07818/) (2015)
- [SAM 2: Segment Anything in Images and Videos](https://mlsanity.com/paper/2408.00714/) (2024)
- [Similarity of Neural Network Representations Revisited](https://mlsanity.com/paper/1905.00414/) (2019)
- [FitNets: Hints for Thin Deep Nets](https://mlsanity.com/paper/1412.6550/) (2014)
- [DINO: DETR with Improved DeNoising Anchor Boxes for End-to-End Object Detection](https://mlsanity.com/paper/2203.03605/) (2022)
- [SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features](https://mlsanity.com/paper/2502.14786/) (2025)
- [ImageBind: One Embedding Space To Bind Them All](https://mlsanity.com/paper/2305.05665/) (2023)
- [Return of Frustratingly Easy Domain Adaptation](https://mlsanity.com/paper/1511.05547/) (2015)
- [How Do Vision Transformers Work?](https://mlsanity.com/paper/2202.06709/) (2022)
- [Global Filter Networks for Image Classification](https://mlsanity.com/paper/2107.00645/) (2021)
- [LLVIP: A Visible-infrared Paired Dataset for Low-light Vision](https://mlsanity.com/paper/2108.10831/) (2021)
- [FcaNet: Frequency Channel Attention Networks](https://mlsanity.com/paper/2012.11879/) (2020)
- [AM-RADIO: Agglomerative Vision Foundation Model -- Reduce All Domains Into One](https://mlsanity.com/paper/2312.06709/) (2023)
- [Focal and Global Knowledge Distillation for Detectors](https://mlsanity.com/paper/2111.11837/) (2021)
- [Masked Generative Distillation](https://mlsanity.com/paper/2205.01529/) (2022)
- [Cross-Modality Fusion Transformer for Multispectral Object Detection](https://mlsanity.com/paper/2111.00273/) (2021)
- [Improving Multispectral Pedestrian Detection by Addressing Modality Imbalance Problems](https://mlsanity.com/paper/2008.03043/) (2020)
- [Multimodal Object Detection via Probabilistic Ensembling](https://mlsanity.com/paper/2104.02904/) (2021)
- [Multispectral Pedestrian Detection via Simultaneous Detection and Segmentation](https://mlsanity.com/paper/1808.04818/) (2018)
- [Wavelet Knowledge Distillation: Towards Efficient Image-to-Image Translation](https://mlsanity.com/paper/2203.06321/) (2022)
- [FreeKD: Knowledge Distillation via Semantic Frequency Prompt](https://mlsanity.com/paper/2311.12079/) (2023)
- [Frequency Attention for Knowledge Distillation](https://mlsanity.com/paper/2403.05894/) (2024)
- [PHI-S: Distribution Balancing for Label-Free Multi-Teacher Distillation](https://mlsanity.com/paper/2410.01680/) (2024)
- [Unveiling the Potential of Segment Anything Model 2 for RGB-Thermal Semantic Segmentation with Language Guidance](https://mlsanity.com/paper/2503.02581/) (2025)

## Abstract (arXiv)

Transfer learning from large-scale RGB foundation models to infrared (IR) imagery through knowledge distillation (KD) remains challenging due to fundamental differences in image formation physics. We investigate the spectral structure of the RGB--IR modality gap and observe that feature divergence is not uniform across spatial frequencies: low-frequency components (shape, layout) show greater cross-modal alignment than high-frequency components (texture, fine edges), which reflect modality-specific characteristics. Based on this analysis, we propose FreqKD, a frequency-decoupled distillation framework that applies asymmetric supervision adapted to each band's cross-modal consistency. The method employs strict mean squared error (MSE) on the low-frequency band to preserve shared structural information and a relaxed log-MSE loss (weighted at 0.1) on the high-frequency band to provide edge guidance while tolerating texture differences. Spectral divergence analysis on 500 paired samples shows that high-frequency divergence exceeds low-frequency divergence by a factor of 2.4x on average across all analysed transformer layers. On KAIST multispectral pedestrian detection, FreqKD achieves 64.1 mAP50, improving 2.4 points over the DINOv2 baseline. The learned representation transfers across datasets (FLIR ADAS, +2.1 mAP50), tasks (MFNet segmentation, +1.85 mean intersection-over-union), and architectures (ResNet-50, +1.0 mAP50). Code is available at: https://anonymous.4open.science/r/freq_decoupled_kd-5E5A