Research map / Detection and segmentation / Open-set and few-shot detection
Tap and Shoot Segmentation
Ding-Jie Chen, Jui-Ting Chien, Hwann-Tzong Chen, Long-Wen Chang
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 |
|---|---|---|---|
| replace pretrained FCN with simple encoder-decoder (AE) from scratch (ablation) | tap-and-shoot segmentation network · five datasets, 16K updates | mean IoU: GrabCut 0.7228 -> 0.6909; space 652.45MB -> 98.99MB; time 4h -> 1.5h | Hurt |
| extend FCN up-sampling with extra deconvolution layers (FCN-extended) (ablation) | FCN · GrabCut | mean IoU: 0.7228 -> 0.7159 | No clear effect |
| add off-the-shelf defocus estimation channel D to input (ablation) | segmentation network · mean IoU on datasets | mean IoU | Hurt |
| train with blurred-background samples (learn defocus implicitly) (ablation) | RGBS network · datasets | segmentation quality | Improved |
| use tap-and-shoot network instead of Xu et al. 2016 | interactive segmentation · GrabCut and Berkeley | IoU: roughly 15% and 2% IoU increments | Improved |
Closest papers by meaning
- Interactive Image Segmentation With Latent Diversity (CVPR 2018 · similarity 0.85)
- FocusCut: Diving Into a Focus View in Interactive Segmentation (CVPR 2022 · similarity 0.85)
- Prompting Diffusion Models for Zero-Shot Instance Segmentation (NeurIPS 2026 · similarity 0.84)
- Segment This Thing: Foveated Tokenization for Efficient Point-Prompted Segmentation (CVPR 2025 · similarity 0.84)
- NTClick: Achieving Precise Interactive Segmentation With Noise-tolerant Clicks (CVPR 2025 · similarity 0.83)
- Repurposing Stable Diffusion Attention for Training-Free Unsupervised Interactive Segmentation (CVPR 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.
