Research map / Detection and segmentation / Open-set and few-shot detection
Cut, Paste and Learn: Surprisingly Easy Synthesis for Instance Detection
Debidatta Dwibedi, Ishan Misra, Martial Hebert
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 |
|---|---|---|---|
| add Gaussian blurring (ablation) | synthetic data pipeline · GMU Kitchen test | mAP: 65.9 -> 68.9 | Improved |
| add Poisson blending alone (ablation) | synthetic data pipeline · GMU Kitchen test | mAP: 65.9 -> 58.4 | Hurt |
| combine all blending modes (ablation) | synthetic data pipeline · GMU Kitchen test | mAP: 65.9 -> 72.4 | Improved |
| generate the same scene with different blending modes (ablation) | synthetic data pipeline · GMU Kitchen test | mAP: 72.4 -> 73.7 | Improved |
| add 2D rotation augmentation (ablation) | synthetic data pipeline · GMU Kitchen test | mAP: 69.7 -> 73.7 | Improved |
| add 3D rotation augmentation (ablation) | synthetic data pipeline · GMU Kitchen test | mAP: 68.3 -> 73.7 | Improved |
| add truncation (ablation) | synthetic data pipeline · GMU Kitchen test | mAP: 71.8 -> 73.7 | Improved |
| add occlusion (ablation) | synthetic data pipeline · GMU Kitchen test | mAP: 63.1 -> 73.7 | Improved |
| add distractor objects (ablation) | synthetic data pipeline · GMU Kitchen test | mAP: 73.7 -> 76.2 | Improved |
| add synthetic data to real GMU training data | Faster R-CNN · GMU Kitchen test | mAP: 86.3 -> 88.8 | Improved |
| add synthetic images to 10% real data | Faster R-CNN · Active Vision | mAP: 15.8 -> 43.2 | Improved |
| add synthetic images to full real data | Faster R-CNN · Active Vision | mAP: 41.9 -> 51.1 | Improved |
Closest papers by meaning
- The Devil is in the Object Boundary: Towards Annotation-free Instance Segmentation using Foundation Models (ICLR 2024 · similarity 0.89)
- Synthetic Object Compositions for Scalable and Accurate Learning in Detection, Segmentation, and Grounding (CVPR 2026 · similarity 0.88)
- Sampling Techniques for Large-Scale Object Detection From Sparsely Annotated Objects (CVPR 2019 · similarity 0.88)
- Many-shot from Low-shot: Learning to Annotate using Mixed Supervision for Object Detection (ECCV 2020 · similarity 0.87)
- Rethinking ImageNet Pre-Training (ICCV 2019 · similarity 0.87)
- A Simple Background Augmentation Method for Object Detection with Diffusion Model (ECCV 2024 · similarity 0.87)
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.
