Research map / Detection and segmentation / CNN features
Correspondence Transformers With Asymmetric Feature Learning and Matching Flow Super-Resolution
Read the paperCheck 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 |
|---|---|---|---|
| Use asymmetric feature learning and flow super-resolution | ACTR with iBOT-B · SPair-71K | PCK@0.1: 57.7 -> 62.1 | Helped |
| Use ACTR instead of TransforMatcher with the same iBOT-B backbone | SPair-71K · SPair-71K | PCK@0.1: 57.9 -> 62.1 | Helped |
| Use ACTR at 512 × 512 resolution instead of VAT‡ | iBOT-B matching · SPair-71K | PCK@0.1: 59.0 -> 65.4 | Helped |
| Use ACTR | PF-PASCAL with iBOT-B · PF-PASCAL | PCK@0.1: 93.3 -> 94.0 | Helped |
| Add source branch positional encoding (ablation) | ACTR · SPair-71K | PCK@0.1: 60.4 (1.7↓) -> 62.1 | Helped |
| Add target branch token reweighting (ablation) | ACTR · SPair-71K | PCK@0.1: 60.7 (1.4↓) -> 62.1 | Helped |
| Add asymmetric cross-attention (ablation) | ACTR · SPair-71K | PCK@0.1: 60.1 (2.0↓) -> 62.1 | Helped |
| Add multi-path super-resolution (ablation) | ACTR · SPair-71K | PCK@0.1: 61.0 (1.1↓) -> 62.1 | Helped |
| Add dual-window flow refinement (ablation) | ACTR · SPair-71K | PCK@0.1: 60.6 (1.5↓) -> 62.1 | Helped |
| Add matching flow super-resolution (ablation) | ACTR · SPair-71K | PCK@0.1: 59.0 (3.1↓) -> 62.1 | Helped |
Closest papers by meaning
- Deep Semantic Feature Matching (CVPR 2017 · similarity 0.89)
- Pixel-level Semantic Correspondence through Layout-aware Representation Learning and Multi-scale Matching Integration (CVPR 2024 · similarity 0.89)
- Unifying Feature and Cost Aggregation with Transformers for Semantic and Visual Correspondence (ICLR 2024 · similarity 0.87)
- SemAlign3D: Semantic Correspondence between RGB-Images through Aligning 3D Object-Class Representations (CVPR 2025 · similarity 0.86)
- Noisy Correspondence Rectification via Asymmetric Similarity Learning (AAAI 2025 · similarity 0.86)
- Win-Win: Training High-Resolution Vision Transformers from Two Windows (ICLR 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.