Research map / Detection and segmentation / CNN features
STEERER: Resolving Scale Variations for Counting and Localization via Selective Inheritance Learning
Tao Han, Lei Bai, Lingbo Liu, Wanli Ouyang
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 |
|---|---|---|---|
| replace FPN fusion with FSIA (CFN+UFN+SMG) (ablation) | STEERER · SHHA | MAE: 60.4 -> 54.6 | Improved |
| add soft-mask generator to UFN+CFN (ablation) | FSIA · SHHA | MAE: 55.1 -> 54.6 | Improved |
| add UFN to CFN (ablation) | FSIA · SHHA | MAE: 55.3 -> 55.1 | No clear effect |
| impose selection and inheritance loss on all four resolutions (ablation) | STEERER · SHHA | MAE: 58.4 -> 54.6 | Improved |
| add inheritance mask on top of selection mask (ablation) | MSIL · SHHA | MAE: 675.4 -> 54.6 | Improved |
| use masks at all instead of no mask (ablation) | STEERER · SHHA | MAE: 60.5 -> 54.6 | Improved |
| share the counting head across resolutions (ablation) | STEERER · SHHA | MAE: 329.2 -> 54.6 | Improved |
| aggregate multi-resolution predictions instead of using the highest resolution only (ablation) | STEERER · SHHA | MAE/MSE: 54.5/86.9 -> 63.4/102.5 | Hurt |
| use STEERER instead of FPN fusion for cross-dataset transfer | HRNet · SHHA to QNRF | MAE: 120.3 -> 109.4 | Improved |
Later papers that compared with this method
1 later accepted paper reports STEERER as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Regressor-Segmenter Mutual Prompt Learning for Crowd Counting (CVPR 2024 · better on 6 of 10 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Vision Transformer Off-the-Shelf: A Surprising Baseline for Few-Shot Class-Agnostic Counting (AAAI 2024 · similarity 0.84)
- Learning to Count without Annotations (CVPR 2024 · similarity 0.84)
- T2ICount: Enhancing Cross-modal Understanding for Zero-Shot Counting (CVPR 2025 · similarity 0.83)
- A Fixed-Point Approach to Unified Prompt-Based Counting (AAAI 2024 · similarity 0.83)
- A Novel Unified Architecture for Low-Shot Counting by Detection and Segmentation (NeurIPS 2024 · similarity 0.83)
- OmniCount: Multi-label Object Counting with Semantic-Geometric Priors (AAAI 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.
