Research map / Vision-language and multimodal / Human and embodied modeling
Efficient Visual Computing With Camera RAW Snapshots
Zhihao Li, Ming Lu, Xu Zhang, Xin Feng, M. Salman Asif, Zhan Ma
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use the same ISP for training and testing instead of mismatched ISPs (ablation) | RGB-Vision YOLOv8-S · MultiRAW detection | mAP: 0.397 -> 0.529 | Improved |
| use expert-calibrated iPhone ISP instead of tuned AX620A ISP (ablation) | RGB-Vision YOLOv8-S · MultiRAW detection | mAP: 0.529 -> 0.546 | Improved |
| tune the AX620A ISP instead of default settings (ablation) | RGB-Vision YOLOv8-S · MultiRAW detection | mAP: 0.173 -> 0.397 | Improved |
| replace ISP + RGB detector with RAW-domain detector | YOLOv8-S · MultiRAW detection | mAP: 0.546 -> 0.546 | No clear effect |
| replace ISP + RGB detector with RAW-domain detector | YOLOv8-S · AX620A hardware | latency: 72% latency reduction | Improved |
| replace ISP + RGB classifier with RAW-domain classifier | MobileNet-V1 · Low-light classification | Top-1 accuracy: 0.0 -> 19.8 | Improved |
| replace ISP + RGB classifier with RAW-domain classifier | MobileNet-V1 · Low-light classification | energy: 0.128 J -> 0.006 J | Improved |
| use AX620A AI-ISP instead of standard ISP (ablation) | RGB-Vision MobileNet-V1 · Low-light classification | Top-1 accuracy: 0.0 -> 0.0 | No clear effect |
| replace ISP + RGB detector with RAW-domain detector | YOLOv8-S · HDR detection | mAP: 56.8 -> 63.3 | Improved |
| use Unpaired CycleR2R simRAW instead of Mosaic RAW | MobileNet-V1 · Low-light classification | Top-5 accuracy: 52.5 -> 72.1 | Improved |
| use Unpaired CycleR2R simRAW instead of CycleISP | HRNetv2 · Segmentation on iPhone RAW | mIoU: 43.6 -> 50.6 | Improved |
| train on simRAW instead of feeding real RAW to RGB model | HRNetv2 · Segmentation on iPhone RAW | mIoU: 11.1 -> 50.6 | Improved |
Closest papers by meaning
- ReRAW: RGB-to-RAW Image Reconstruction via Stratified Sampling for Efficient Object Detection on the Edge (CVPR 2025 · similarity 0.84)
- Logarithmic Lenses: Exploring Log RGB Data for Image Classification (CVPR 2024 · similarity 0.83)
- Guiding a Harsh-Environments Robust Detector via RAW Data Characteristic Mining (AAAI 2024 · similarity 0.82)
- RGB-to-Polarization Estimation: A New Task and Benchmark Study (NeurIPS 2025 · similarity 0.82)
- RGB-Only Supervised Camera Parameter Optimization in Dynamic Scenes (NeurIPS 2025 · similarity 0.82)
- Vision-RWKV: Efficient and Scalable Visual Perception with RWKV-Like Architectures (ICLR 2025 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
