Research map / Detection and segmentation / CNN features
Instance-Aware Image Colorization
Check 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 |
|---|---|---|---|
| add instance-aware feature extraction and fusion | Zhang et al. colorization backbone · ImageNet ctest10k | PSNR: 26.482 -> 26.980 | Helped |
| add instance-aware feature extraction and fusion | Zhang et al. colorization backbone · COCO-Stuff validation split | PSNR: 27.251 -> 27.777 | Helped |
| add instance-aware feature extraction and fusion | Zhang et al. colorization backbone · Places205 validation split | PSNR: 25.720 -> 27.167 | Helped |
| add instance-aware feature extraction and fusion | Zhang et al. colorization backbone · COCO-Stuff validation split (instance level) | PSNR: 26.675 -> 28.339 | Helped |
| add feature fusion (ablation) | Instance-aware colorization model · COCO-Stuff validation split | PSNR: 27.251 -> 28.592 | Helped |
| add instance-feature fusion in encoder only (ablation) | Instance-aware colorization model · COCO-Stuff validation split | PSNR: 27.251 -> 28.146 | Helped |
| add instance-feature fusion in decoder only (ablation) | Instance-aware colorization model · COCO-Stuff validation split | PSNR: 27.251 -> 27.959 | Helped |
| fuse instance features in both encoder and decoder rather than encoder only (ablation) | Instance-aware colorization model · COCO-Stuff validation split | PSNR: 28.146 -> 28.592 | Helped |
| select the top 8 detector boxes by confidence rather than randomly selecting 8 (ablation) | Instance-aware colorization model · COCO-Stuff validation split | PSNR: 28.386 -> 28.592 | Helped |
| select the top 8 detector boxes by confidence rather than using a threshold (ablation) | Instance-aware colorization model · COCO-Stuff validation split | PSNR: 28.139 -> 28.592 | Helped |
| use top-confidence detector boxes rather than ground-truth boxes (ablation) | Instance-aware colorization model · COCO-Stuff validation split | PSNR: 28.470 -> 28.592 | No clear effect |
| use learned fusion weights rather than detected box masks (ablation) | Instance-aware colorization model · COCO-Stuff validation split | PSNR: 26.456 -> 28.592 | Helped |
Closest papers by meaning
- Learning a Deep Convolutional Network for Colorization in Monochrome-Color Dual-Lens System (AAAI 2019 · similarity 0.89)
- Colorization Transformer (ICLR 2021 · similarity 0.89)
- L-CoIns: Language-Based Colorization With Instance Awareness (CVPR 2023 · similarity 0.87)
- Color-Oriented Redundancy Reduction in Dataset Distillation (NeurIPS 2024 · similarity 0.86)
- Enabling Region-Specific Control via Lassos in Point-Based Colorization (AAAI 2025 · similarity 0.86)
- Automatic Controllable Colorization via Imagination (CVPR 2024 · similarity 0.85)
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.