Research map / Detection and segmentation / CNN features
ReconfigISP: Reconfigurable Camera Image Processing Pipeline
Ke Yu, Zexian Li, Yue Peng, Chen Change Loy, Jinwei Gu
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 |
|---|---|---|---|
| replace default camera ISP with searched ReconfigISP | Sony α7 II pipeline · SID Sony (RGB) | PSNR: 15.69 -> 25.65 | Improved |
| replace default camera ISP with searched ReconfigISP | Samsung S7 pipeline · S7 ISP (RGB) | PSNR: 21.08 -> 23.31 | Improved |
| replace Software ISP with ReconfigISP for detection | YOLOv3 on OnePlus · OnePlus detection | mAP: 0.515 -> 0.601 | Improved |
| use searched architecture instead of random search (ablation) | ReconfigISP · S7 ISP | PSNR: 23.13 -> 23.31 | Improved |
| use searched architecture instead of random search (ablation) | ReconfigISP · OnePlus detection | mAP: 0.566 -> 0.601 | Improved |
| use task-specific architecture instead of one from detection task (ablation) | ReconfigISP on S7 ISP · S7 ISP | PSNR: 20.55 -> 23.31 | Improved |
| use task-specific architecture instead of one from restoration task (ablation) | ReconfigISP on OnePlus · OnePlus detection | mAP: 0.352 -> 0.601 | Improved |
| add proxy tuning (ablation) | ReconfigISP architecture search · S7 ISP | PSNR: 21.01 -> 23.31 | Improved |
| increase latency penalty to beta=0.14 (Fast) (ablation) | ReconfigISP · SID | PSNR: 25.65 -> 23.72 (1.16 s -> 0.63 s) | Mixed |
| increase latency penalty to beta=0.28 (Faster) (ablation) | ReconfigISP · SID | PSNR: 25.65 -> 20.55 (1.16 s -> 0.049 s) | Mixed |
| increase latency penalty to beta=0.28 (Faster) (ablation) | ReconfigISP · S7 ISP | PSNR: 23.31 -> 20.42 (1.53 s -> 0.031 s) | Mixed |
| use ReconfigISP instead of U-Net with few training patches | SID · SID, 100 training patches | PSNR: 18.43 -> 22.73 | Improved |
Closest papers by meaning
- AdaptiveISP: Learning an Adaptive Image Signal Processor for Object Detection (NeurIPS 2024 · similarity 0.86)
- RL-SeqISP: Reinforcement Learning-Based Sequential Optimization for Image Signal Processing (AAAI 2024 · similarity 0.85)
- Learning Degradation-Independent Representations for Camera ISP Pipelines (CVPR 2024 · similarity 0.82)
- ParamISP: Learned Forward and Inverse ISPs using Camera Parameters (CVPR 2024 · similarity 0.82)
- RGB-Event ISP: The Dataset and Benchmark (ICLR 2025 · similarity 0.82)
- ISP-Teacher:Image Signal Process with Disentanglement Regularization for Unsupervised Domain Adaptive Dark Object Detection (AAAI 2024 · similarity 0.80)
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.
