A Learnable Color Correction Matrix for RAW Reconstruction
This addresses the challenge of limited RAW image availability for autonomous driving research, offering a lightweight solution for RAW-domain perception, though it is incremental as it builds on existing inverse ISP methods.
The paper tackles the problem of reconstructing RAW images from sRGB for autonomous driving perception by introducing a learnable color correction matrix (CCM) that uses a single convolutional layer to approximate the inverse image signal processor (ISP), resulting in simulated RAW images that provide performance improvements equivalent to more complex methods when pretraining object detectors.
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.