Research map / Detection and segmentation / CNN features
Learning To Exploit the Sequence-Specific Prior Knowledge for Image Processing Pipelines Optimization
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 sequential parameter prediction instead of one-step prediction (ablation) | ISP parameter prediction · IMX766 image quality | PSNR: 39.52 -> 45.29 | Helped |
| use five rather than four prediction steps (ablation) | ISP parameter prediction · IMX766 image quality | PSNR: 45.29 -> 45.23 | No clear effect |
| group parameters by learned similarity rather than by ISP blocks (ablation) | Four-step ISP parameter prediction · IMX766 image quality | PSNR: 43.57 -> 45.29 | Helped |
| use ISP-prior sequence order rather than random group order (ablation) | Similarity-grouped ISP parameter prediction · IMX766 image quality | PSNR: 42.62 -> 45.29 | Helped |
| use similarity-based groups and ISP-prior sequence order rather than random groups and order (ablation) | ISP parameter prediction · IMX766 image quality | PSNR: 39.30 -> 45.29 | Helped |
| use Sequential-tuned rather than Attention-ware ISP tuning | COCO object detection · COCO | mAP 0.5: 61.0 -> 62.8 | Helped |
| use Sequential-tuned rather than Attention-ware ISP tuning | COCO image segmentation · COCO | mAP 0.5: 52.3 -> 54.2 | Helped |
| use Sequential-tuned rather than Attention-ware ISP tuning | ISP parameter prediction · IMX766 image quality | PSNR: 40.08 -> 45.29 | Helped |
Closest papers by meaning
- RL-SeqISP: Reinforcement Learning-Based Sequential Optimization for Image Signal Processing (AAAI 2024 · similarity 0.89)
- AdaptiveISP: Learning an Adaptive Image Signal Processor for Object Detection (NeurIPS 2024 · similarity 0.88)
- Learning Degradation-Independent Representations for Camera ISP Pipelines (CVPR 2024 · similarity 0.83)
- ParamISP: Learned Forward and Inverse ISPs using Camera Parameters (CVPR 2024 · similarity 0.83)
- Principled Architecture-aware Scaling of Hyperparameters (ICLR 2024 · similarity 0.83)
- RGB-Event ISP: The Dataset and Benchmark (ICLR 2025 · similarity 0.83)
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.