Research map / Image generation / Autoregressive and latent generation
Deep Sketch Hashing: Fast Free-Hand Sketch-Based Image Retrieval
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 sketch-token inputs as a bridge between sketches and natural images (ablation) | DSH · TU-Berlin Extension and Sketchy | MAP: 0.497 -> 0.570; 0.682 -> 0.783 | Helped |
| add the natural-image network to the sketch and sketch-token networks (ablation) | DSH · TU-Berlin Extension and Sketchy | MAP: 0.379 -> 0.570; 0.507 -> 0.783 | Helped |
| add semantic factorization loss to cross-view pairwise loss (ablation) | DSH · TU-Berlin Extension and Sketchy | MAP: 0.522 -> 0.570; 0.715 -> 0.783 | Helped |
| add cross-view pairwise loss to semantic factorization loss (ablation) | DSH · TU-Berlin Extension and Sketchy | MAP: 0.485 -> 0.570; 0.667 -> 0.783 | Helped |
| use DSH rather than the strongest listed cross-modality baseline | 128-bit category-level SBIR · TU-Berlin Extension | MAP: 0.425 -> 0.570 | Helped |
| use DSH rather than the strongest listed cross-modality baseline | 128-bit category-level SBIR · Sketchy | MAP: 0.675 (4096-d) -> 0.783 | Helped |
| use compact binary DSH codes rather than continuous-valued SBIR features | SBIR retrieval · TU-Berlin Extension and Sketchy | retrieval time per query (s): DSH: 1.05×10^{-3} s and 3.53×10^{-4} s | Helped |
Closest papers by meaning
- Sketch Me That Shoe (CVPR 2016 · similarity 0.93)
- SketchMate: Deep Hashing for Million-Scale Human Sketch Retrieval (CVPR 2018 · similarity 0.93)
- Zero-Shot Sketch-Image Hashing (CVPR 2018 · similarity 0.92)
- ARNet: Self-Supervised FG-SBIR with Unified Sample Feature Alignment and Multi-Scale Token Recycling (AAAI 2025 · similarity 0.90)
- How to Handle Sketch-Abstraction in Sketch-Based Image Retrieval? (CVPR 2024 · similarity 0.88)
- Sketch Down the FLOPs: Towards Efficient Networks for Human Sketch (CVPR 2025 · similarity 0.88)
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.