Few-Shot LoRA Adaptation of a Flow-Matching Foundation Model for Cross-Spectral Object Detection

Maxim Clouser, Kia Khezeli, John Kalantari

Keywords: cross-spectral image translation, infrared imaging, synthetic aperture radar, flow-matching diffusion, rectified flow transformer, foundation models, LoRA fine-tuning, few-shot adaptation, synthetic data, object detection, LPIPS, sim-to-real gap, low-data regime

TL;DR: We adapt a flow-matching diffusion model with LoRA to translate RGB to pixel-aligned IR and SAR, and show LPIPS selects adapters that give the best synthetic-data detection gains on KAIST and M4SAR in low-data regimes.

Abstract: Foundation models for vision are predominantly trained on RGB data, while many safety-critical applications rely on non-visible modalities such as infrared (IR) and synthetic aperture radar (SAR). We study whether a single flow-matching foundation model pre-trained primarily on RGB images can be repurposed as a cross-spectral translator using only a few co-measured examples, and whether the resulting synthetic data can enhance downstream detection. Starting from FLUX.1 Kontext, we insert low-rank adaptation (LoRA) modules and fine-tune them on just 100 paired images per domain for two settings: RGB→IR on the KAIST dataset and RGB→SAR on the M4-SAR dataset. The adapted model translates RGB images into pixel-aligned IR/SAR, enabling us to reuse existing bounding boxes and train object detection models purely in the target modality. Across a grid of LoRA hyperparameters, we find that LPIPS computed on only 50 held-out pairs is a strong proxy for downstream performance: lower LPIPS consistently predicts higher mAP for YOLOv11n on both IR and SAR, and for DETR on KAIST IR test data. Using the best LPIPS-selected LoRA adapter, synthetic IR from external RGB datasets (LLVIP, FLIR ADAS) improves KAIST IR pedestrian detection, and synthetic SAR significantly boosts infrastructure detection on M4-SAR when combined with limited real SAR. Our results suggest that few-shot LoRA adaptation of flow-matching foundation models is a promising path toward foundation-style support for non-visible modalities.

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 2