Problem
Research questions and friction points this paper is trying to address.
RAW Restoration
RGB Restoration
Image Signal Processor (ISP)
Benchmarking
Benchmarking RAW and RGB Restoration in Image Signal Processors
🤖 AI Summary
研究通过对比RAW和sRGB域的图像恢复方法，发现ISP转换意识下的RGB恢复模型表现最佳，强调了恢复模型与成像管道匹配的重要性。
📝 Abstract
Modern cameras transform RAW sensor measurements into sRGB images through an image signal processor (ISP). We benchmark two placements for blind restoration around a fixed ISP: (A) pre-ISP restoration in the RAW domain and (B) post-ISP restoration in the sRGB domain. The benchmark covers four smartphone device groups, two learned ISPs, three degradation regimes--noise, blur, and joint noise and blur--, and several representative RAW and RGB restoration models. Our results show that placement alone does not determine performance. The RAW restoration strategy outperforms the best generic RGB restoration models. However, RGB restoration models trained considering the ISP transformations, achieve the best overall performance. Our novel benchmark demonstrates that the image reconstruction performance strongly depends on the alignment between the restoration model and the target imaging pipeline. We consequently recommend reporting restoration placement and ISP-aware supervision as key experimental factors. Our code is available at https://github.com/mv-lab/AISP
Innovation
Methods, ideas, or system contributions that make the work stand out.
benchmark
RAW restoration
RGB restoration
ISP-aware supervision
🔎 Similar Papers
No similar papers found.
