# ReRAW: RGB-to-RAW Image Reconstruction via Stratified Sampling for Efficient Object Detection on the Edge

arXiv: https://arxiv.org/abs/2503.03782 · HTML: https://mlsanity.com/paper/2503.03782/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 4 Mar 2025.

## Built on by (most cited first, 8 of 8)

- [RAW Image Reconstruction from RGB on Smartphones. NTIRE 2025 Challenge Report](https://mlsanity.com/paper/2506.01947/) (2025)
- [RAWild: Sensor-Agnostic RAW Object Detection via Physics-Guided Curve and Grid Modeling](https://mlsanity.com/paper/2605.05941/) (2026)
- [SemiISP/SemiIE: Semi-Supervised Image Signal Processor and Image Enhancement Leveraging One-to-Many Mapping sRGB-to-RAW](https://mlsanity.com/paper/2504.02345/) (2025)
- [Raw-JPEG Adapter: Efficient Raw Image Compression with JPEG](https://mlsanity.com/paper/2509.19624/) (2025)
- [Edit-aware RAW Reconstruction](https://mlsanity.com/paper/2512.05859/) (2025)
- [SpiralDiff: Spiral Diffusion with LoRA for RGB-to-RAW Conversion Across Cameras](https://mlsanity.com/paper/2603.14885/) (2026)
- [RawGen: Learning Camera Raw Image Generation](https://mlsanity.com/paper/2604.00093/) (2026)
- [Making the Discrete Continuous: Synthetic RAW Augmentations for Fine-Grained Evaluation of Person Detection Performance in Low Light](https://mlsanity.com/paper/2605.22455/) (2026)

## Abstract (arXiv)

Edge-based computer vision models running on compact, resource-limited devices benefit greatly from using unprocessed, detail-rich RAW sensor data instead of processed RGB images. Training these models, however, necessitates large labeled RAW datasets, which are costly and often impractical to obtain. Thus, converting existing labeled RGB datasets into sensor-specific RAW images becomes crucial for effective model training. In this paper, we introduce ReRAW, an RGB-to-RAW conversion model that achieves state-of-the-art reconstruction performance across five diverse RAW datasets. This is accomplished through ReRAW's novel multi-head architecture predicting RAW image candidates in gamma space. The performance is further boosted by a stratified sampling-based training data selection heuristic, which helps the model better reconstruct brighter RAW pixels. We finally demonstrate that pretraining compact models on a combination of high-quality synthetic RAW datasets (such as generated by ReRAW) and ground-truth RAW images for downstream tasks like object detection, outperforms both standard RGB pipelines, and RAW fine-tuning of RGB-pretrained models for the same task.