Research map / Detection and segmentation / Open-set and few-shot detection
Few-Shot Object Counting and Detection
Thanh Nguyen, Chau Pham, Khoi Nguyen, Minh Hoai
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 stage-1 pseudo GT boxes instead of ridge regression boxes (ablation) | Counting-DETR · FSCD-147 val | AP: 11.37 -> 17.27 | Improved |
| add stage-1 pseudo GT boxes instead of ridge regression boxes (ablation) | Counting-DETR · FSCD-147 val | MAE: 29.74 -> 20.38 | Improved |
| add Laplace uncertainty loss (ablation) | Counting-DETR · FSCD-147 val | AP: 14.19 -> 17.27 | Improved |
| add Laplace uncertainty loss (ablation) | Counting-DETR · FSCD-147 val | MAE: 23.57 -> 20.38 | Improved |
| add both pseudo boxes and uncertainty loss (ablation) | Counting-DETR · FSCD-147 val | AP: 10.81 -> 17.27 | Improved |
| use fixed-grid instead of learnable anchor points (ablation) | Counting-DETR · FSCD-147 val | AP: 16.46 -> 17.27 | Improved |
| use fixed-grid instead of learnable anchor points (ablation) | Counting-DETR · FSCD-147 val | MAE: 25.20 -> 20.38 | Improved |
| increase anchor points from 100 to 600 (ablation) | Counting-DETR · FSCD-147 val | AP: 11.26 -> 17.27 | Improved |
| increase anchor points from 600 to 700 (ablation) | Counting-DETR · FSCD-147 val | AP: 17.27 -> 15.10 | Hurt |
| use Gaussian instead of Laplace uncertainty loss (ablation) | Counting-DETR · FSCD-147 val | AP: 17.27 -> 14.03 | Hurt |
| use Gaussian uncertainty loss instead of none (ablation) | Counting-DETR · FSCD-147 val | MAE: 23.20 -> 24.46 | Hurt |
| use Counting-DETR over best adapted baseline | FSDetView+pseudo box · FSCD-LVIS test | MAE: 24.89 -> 18.51 | Improved |
Closest papers by meaning
- Point Segment and Count: A Generalized Framework for Object Counting (CVPR 2024 · similarity 0.91)
- A Novel Unified Architecture for Low-Shot Counting by Detection and Segmentation (NeurIPS 2024 · similarity 0.90)
- Few-Shot Pattern Detection via Template Matching and Regression (ICCV 2025 · similarity 0.89)
- A Low-Shot Object Counting Network With Iterative Prototype Adaptation (ICCV 2023 · similarity 0.89)
- DAVE - A Detect-and-Verify Paradigm for Low-Shot Counting (CVPR 2024 · similarity 0.89)
- Single Domain Generalization for Few-Shot Counting via Universal Representation Matching (CVPR 2025 · 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.
