Research map / Label-efficient and robust learning / Self-supervised and contrastive
Adaptive Dilated Network With Self-Correction Supervision for Counting
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 self-correction supervision | ADNet / VGG16 bn counting model · ShanghaiTech A | MAE: 61.3 -> 55.4 | Helped |
| add adaptive dilated convolution | VGG baseline · ShanghaiTech A | MAE: 66.5 -> 61.3 | Helped |
| add self-correction supervision (ablation) | VGG baseline · ShanghaiTech A | MAE: 66.5 -> 60.7 | Helped |
| add self-correction supervision (ablation) | CSRNet* · ShanghaiTech A | MAE: 64.2 -> 58.7 | Helped |
| add self-correction supervision (ablation) | MCNN* · ShanghaiTech A | MAE: 108.2 -> 101.5 | Helped |
| increase SC iterations from 2 to 3 or 4 (ablation) | ADSCNet · ShanghaiTech A | MAE: 55.4 -> 57.6; 55.4 -> 58.3 | Hurt |
| use one SC iteration instead of two (ablation) | ADSCNet · ShanghaiTech A | MAE: 55.4 -> 56.2 | Hurt |
| use adaptive rather than static dilation (ablation) | VGG baseline · ShanghaiTech A | MAE: 64.2 -> 61.3 | Helped |
| replace adaptive dilation with multi-branch dilation rates (1,3,5) (ablation) | VGG baseline · ShanghaiTech A | MAE: 61.3 -> 63.6 | Hurt |
| replace adaptive dilation with deformable convolution (ablation) | VGG baseline · ShanghaiTech A | MAE: 61.3 -> 62.6 | Hurt |
Closest papers by meaning
- Learning to Count without Annotations (CVPR 2024 · similarity 0.85)
- A Fixed-Point Approach to Unified Prompt-Based Counting (AAAI 2024 · similarity 0.83)
- Efficient Masked AutoEncoder for Video Object Counting and A Large-Scale Benchmark (ICLR 2025 · similarity 0.83)
- Enhancing Low-Rank Adaptation with Recoverability-Based Reinforcement Pruning for Object Counting (AAAI 2025 · similarity 0.83)
- Frequency-Adaptive Dilated Convolution for Semantic Segmentation (CVPR 2024 · similarity 0.82)
- Single Domain Generalization for Crowd Counting (CVPR 2024 · similarity 0.82)
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.