mlsanityData to 7 Oct 2026

A Learnable Color Correction Matrix for RAW Reconstruction

arXiv 2409.02497 ↗First posted 4 Sep 2024eess.IV cs.CVcited by 3

Not on the map. Why?

Abstract (arXiv)

Autonomous driving algorithms usually employ sRGB images as model input due to their compatibility with the human visual system. However, visually pleasing sRGB images are possibly sub-optimal for downstream tasks when compared to RAW images. The availability of RAW images is constrained by the difficulties in collecting real-world driving data and the associated challenges of annotation. To address this limitation and support research in RAW-domain driving perception, we design a novel and ultra-lightweight RAW reconstruction method. The proposed model introduces a learnable color correction matrix (CCM), which uses only a single convolutional layer to approximate the complex inverse image signal processor (ISP). Experimental results demonstrate that simulated RAW (simRAW) images generated by our method provide performance improvements equivalent to those produced by more complex inverse ISP methods when pretraining RAW-domain object detectors, which highlights the effectiveness and practicality of our approach.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 32 of its 47 references that are in the map. First posted on arXiv 4 Sep 2024.

Builds on

32 of 47 references.

References · 32

  1. 2015Distilling the Knowledge in a Neural Networkcited by 10,076
  2. 2015You Only Look Once: Unified, Real-Time Object Detectioncited by 4,829
  3. 2019nuScenes: A multimodal dataset for autonomous drivingcited by 4,485
  4. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,427
  5. 2019Scalability in Perception for Autonomous Driving: Waymo Open Datasetcited by 2,436
  6. 2018BDD100K: A Diverse Driving Dataset for Heterogeneous Multitask Learningcited by 1,282
  7. 2019Argoverse: 3D Tracking and Forecasting with Rich Mapscited by 728
  8. 2018Learning to See in the Darkcited by 459
  9. 2021NeRF in the Dark: High Dynamic Range View Synthesis from Noisy Raw Imagescited by 336
  10. 2018Unprocessing Images for Learned Raw Denoisingcited by 250
  11. 2020CycleISP: Real Image Restoration via Improved Data Synthesiscited by 212
  12. 2018The ApolloScape Open Dataset for Autonomous Driving and its Applicationcited by 188
20 more
  1. 2017Burst Denoising with Kernel Prediction Networkscited by 156
  2. 2020A Physics-based Noise Formation Model for Extreme Low-light Raw Denoisingcited by 138
  3. 2020Single-Image HDR Reconstruction by Learning to Reverse the Camera Pipelinecited by 131
  4. 2019Zoom To Learn, Learn To Zoomcited by 114
  5. 2021Invertible Image Signal Processingcited by 94
  6. 2018DeepISP: Towards Learning an End-to-End Image Processing Pipelinecited by 88
  7. 2023Instance Segmentation in the Darkcited by 45
  8. 2022Model-Based Image Signal Processors via Learnable Dictionariescited by 43
  9. 2023RawHDR: High Dynamic Range Image Reconstruction from a Single Raw Imagecited by 31
  10. 2020CIE XYZ Net: Unprocessing Images for Low-Level Computer Vision Taskscited by 29
  11. 2023ParamISP: Learned Forward and Inverse ISPs using Camera Parameterscited by 23
  12. 2022GenISP: Neural ISP for Low-Light Machine Cognitioncited by 22
  13. 2022Learning sRGB-to-Raw-RGB De-rendering with Content-Aware Metadatacited by 22
  14. 2022Efficient Visual Computing with Camera RAW Snapshotscited by 18
  15. 2023Raw Image Reconstruction with Learned Compact Metadatacited by 18
  16. 2023Raw or Cooked? Object Detection on RAW Imagescited by 14
  17. 2023Beyond Learned Metadata-based Raw Image Reconstructioncited by 9
  18. 2023Self-Supervised Reversed Image Signal Processing via Reference-Guided Dynamic Parameter Selectioncited by 5
  19. 2021Raw Bayer Pattern Image Synthesis for Computer Vision-oriented Image Signal Processing Pipeline Designcited by 4
  20. 2022Reversing Image Signal Processors by Reverse Style Transferringcited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 2; 2026: 1.

3 papers.

  1. 2025γ-Quant: Towards Learnable Quantization for Low-bit Pattern Recognitioncited by 2
  2. 2026RAWDet-7: A Multi-Scenario Benchmark for Object Detection and Description on Quantized RAW Imagescited by 1
  3. 2025SemiISP/SemiIE: Semi-Supervised Image Signal Processor and Image Enhancement Leveraging One-to-Many Mapping sRGB-to-RAWcited by 0