Research map / Vision-language and multimodal / Human and embodied modeling
Conditional Stroke Recovery for Fine-Grained Sketch-Based Image Retrieval
Zhixin Ling, Zhen Xing, Jian Zhou, Xiangdong Zhou
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 single-anchor InfoNCE with double-anchor InfoNCE (Lda) (ablation) | CSR-GoogLeNet · Sketchy | acc@1: 42.6 -> 44.9 | Improved |
| replace single-anchor InfoNCE with double-anchor InfoNCE (Lda) (ablation) | CSR-GoogLeNet · QMUL-ShoeV2 | acc@1: 42.2 -> 43.7 | Improved |
| add stroke-disorder sketch augmentation (ablation) | Lsa · QMUL-ShoeV2 | acc@1: 42.2 -> 44.2 | Improved |
| add multi-level feature fusion (MF) (ablation) | Lda · Sketchy | acc@1: 44.9 -> 48.2 | Improved |
| add recovery loss without MF and RC (ablation) | Lda · Sketchy | acc@1: 44.9 -> 45.2 | No clear effect |
| add recovery compensation (RC) (ablation) | Lrec + Lda · Sketchy | acc@1: 45.2 -> 47.7 | Improved |
| add MF on top of recovery loss (ablation) | Lrec + Lda · Sketchy | acc@1: 45.2 -> 48.9 | Improved |
| replace stroke disorder with grid disorder (ablation) | CSR · Sketchy | acc@1: 50.8 -> 41.9 | Hurt |
| add channel 4 (moved strokes) of recovery map (ablation) | CSR · QMUL-ShoeV2 | acc@1: 43.9 -> 48.7 | Improved |
| set alpha_p to 2 instead of 0 (use auxiliary anchor) (ablation) | CSR · Sketchy | acc@1: 36.7 -> 50.8 | Improved |
| efficient triplet collection (b(b-1) triplets) (ablation) | Ltrp · Sketchy (b=16) | acc@1: 38.9 -> 45.8 | Improved |
| use sketch rather than image as anchor (ablation) | Ltrp with ETC · Sketchy (b=16) | acc@1: 37.8 -> 45.8 | Improved |
Closest papers by meaning
- Sketch Less for More: On-the-Fly Fine-Grained Sketch-Based Image Retrieval (CVPR 2020 · similarity 0.91)
- Adaptive Fine-Grained Sketch-Based Image Retrieval (ECCV 2022 · similarity 0.89)
- Generalising Fine-Grained Sketch-Based Image Retrieval (CVPR 2019 · similarity 0.89)
- ARNet: Self-Supervised FG-SBIR with Unified Sample Feature Alignment and Multi-Scale Token Recycling (AAAI 2025 · similarity 0.88)
- How to Handle Sketch-Abstraction in Sketch-Based Image Retrieval? (CVPR 2024 · similarity 0.86)
- Sketch Down the FLOPs: Towards Efficient Networks for Human Sketch (CVPR 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.
