Research map / Detection and segmentation / Open-set and few-shot detection
A High-Resolution Dataset for Instance Detection with Multi-View Object Capture
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 transformer detector DINO instead of FasterRCNN | end-to-end cut-paste-learn detector · InsDet | AP: 19.54 -> 27.99 | Helped |
| use non-learned SAM+DINOv2f instead of end-to-end DINO | InsDet method · InsDet | AP: 27.99 -> 41.61 | Helped |
| use DINOv2f features instead of DINOf | SAM + DINO non-learned method · InsDet | AP: 36.97 -> 41.61 | Helped |
| remove background pixels from SAM crops (ablation) | SAM+DINOv2f · InsDet | AP: 36.04 -> 39.12 | Helped |
| use Stable Matching instead of Rank & Select (ablation) | SAM+DINOv2f · InsDet | AP: 38.62 -> 39.12 | Helped |
| raise SAM input resolution from 768x1024 to 1536x2048 (ablation) | SAM+DINOv2f · InsDet | AP: 36.46 -> 39.12 | Helped |
| raise SAM input resolution from 1536x2048 to 3072x4096 (ablation) | SAM+DINOv2f · InsDet | AP: 39.12 -> 39.17 | No clear effect |
| raise SAM input resolution from 3072x4096 to 6144x8192 (ablation) | SAM+DINOv2f · InsDet | AP: 39.17 -> 38.74 | Hurt |
| increase DINOv2f proposal crop size from 224x224 to 448x448 (ablation) | SAM+DINOv2f · InsDet | AP: 39.12 -> 41.61 | Helped |
| decrease DINOv2f proposal crop size from 224x224 to 112x112 (ablation) | SAM+DINOv2f · InsDet | AP: 39.12 -> 26.46 | Hurt |
| evaluate on hard instead of easy scenes | SAM+DINOv2f · InsDet | AP: 47.57 -> 28.03 | Hurt |
| evaluate small instead of large objects | SAM+DINOv2f · InsDet | AP: 69.14 -> 14.58 | Hurt |
Closest papers by meaning
- Solving Instance Detection from an Open-World Perspective (CVPR 2025 · similarity 0.93)
- InstructDET: Diversifying Referring Object Detection with Generalized Instructions (ICLR 2024 · similarity 0.88)
- Expanding Low-Density Latent Regions for Open-Set Object Detection (CVPR 2022 · similarity 0.88)
- The Devil is in the Object Boundary: Towards Annotation-free Instance Segmentation using Foundation Models (ICLR 2024 · similarity 0.87)
- ImOV3D: Learning Open Vocabulary Point Clouds 3D Object Detection from Only 2D Images (NeurIPS 2024 · similarity 0.86)
- Roboflow100-VL: A Multi-Domain Object Detection Benchmark for Vision-Language Models (NeurIPS 2025 · similarity 0.86)
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.