Research map / Label-efficient and robust learning / Self-supervised and contrastive
Solving Mixed-Modal Jigsaw Puzzle for Fine-Grained 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 |
|---|---|---|---|
| Replace ImageNet classification pre-training with mixed-modal jigsaw pre-training | GoogleNet FG-SBIR · QMUL Shoe V1 | Acc@1: 43.48%± 1.74 -> 56.52%± 2.75 | Helped |
| Replace ImageNet classification pre-training with mixed-modal jigsaw pre-training | GoogleNet FG-SBIR · QMUL Shoe V2 | Acc@1: 33.99%± 1.09 -> 36.52%± 0.84 | Helped |
| Replace ImageNet classification pre-training with mixed-modal jigsaw pre-training | GoogleNet FG-SBIR · QMUL Chair | Acc@1: 85.16%± 1.56 -> 85.98%± 2.01 | No clear effect |
| Replace ImageNet classification pre-training with mixed-modal jigsaw pre-training | GoogleNet FG-SBIR · QMUL Handbag | Acc@1: 52.62%± 2.04 -> 62.97%± 2.04 | Helped |
| Add Sinkhorn permutation inference instead of classifying among 1000 jigsaw permutations (ablation) | Mixed-modal jigsaw pre-training · QMUL Shoe V1 | Acc@1: 42.78%± 3.75 -> 56.52%± 2.75 | Helped |
| Add Sinkhorn permutation inference instead of classifying among 1000 jigsaw permutations (ablation) | Mixed-modal jigsaw pre-training · QMUL Shoe V2 | Acc@1: 30.24%± 1.74 -> 36.52%± 0.84 | Helped |
| Add Sinkhorn permutation inference instead of classifying among 1000 jigsaw permutations (ablation) | Mixed-modal jigsaw pre-training · QMUL Chair | Acc@1: 79.59%± 1.53 -> 85.98%± 2.01 | Helped |
| Add Sinkhorn permutation inference instead of classifying among 1000 jigsaw permutations (ablation) | Mixed-modal jigsaw pre-training · QMUL Handbag | Acc@1: 49.40%± 3.97 -> 62.97%± 2.04 | Helped |
| Mix photo and edgemap modalities within the puzzle at patch level (ablation) | Jigsaw pre-training · QMUL Shoe V1 | Acc@1: 51.82 -> 54.08 | Helped |
| Mix photo and edgemap modalities within the puzzle at patch level (ablation) | Jigsaw pre-training · QMUL Shoe V2 | Acc@1: 35.32 -> 36.52 | Helped |
| Mix photo and edgemap modalities within the puzzle at patch level (ablation) | Jigsaw pre-training · QMUL Chair | Acc@1: 83.92 -> 85.98 | Helped |
| Mix photo and edgemap modalities within the puzzle at patch level (ablation) | Jigsaw pre-training · QMUL Handbag | Acc@1: 56.31 -> 58.10 | Helped |
Closest papers by meaning
- Generalising Fine-Grained Sketch-Based Image Retrieval (CVPR 2019 · similarity 0.91)
- ARNet: Self-Supervised FG-SBIR with Unified Sample Feature Alignment and Multi-Scale Token Recycling (AAAI 2025 · similarity 0.89)
- You'll Never Walk Alone: A Sketch and Text Duet for Fine-Grained Image Retrieval (CVPR 2024 · similarity 0.88)
- Sketch Down the FLOPs: Towards Efficient Networks for Human Sketch (CVPR 2025 · similarity 0.87)
- How to Handle Sketch-Abstraction in Sketch-Based Image Retrieval? (CVPR 2024 · similarity 0.86)
- Asymmetric Mutual Alignment for Unsupervised Zero-Shot Sketch-Based Image Retrieval (AAAI 2024 · 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.