---
格式版本: 2
标题: "Embedding Rotation Invariance for Provable Multi-Oriented Scene Text Recognition"
原文链接: "https://arxiv.org/abs/2608.10684"
发布日期: "2026-08-11"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Tue, 11 Aug 2026 09:06:46 UTC (405 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-08-12T11:27:52+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-08-12T11:26:08+08:00"
入库时间: "2026-08-12T03:27:52.746Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=inference%20networking&searchtype=all"
匹配关键词:
  - "performance"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 0
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "内容为计算机视觉场景文本识别论文，与超节点/AI Rack/机柜级AI基础设施完全无关，属于明显无关资料。"
AI质检模型: "ali-deepseek-v4-flash"
AI质检时间: "2026-08-12T11:28:04+08:00"
AI主题相关性: 0
AI来源权威性: 0
AI新颖性: 0
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 0
采集批次: "2026年8月12日2点04分20秒"
采集批次ID: "20260812-020420-289"
去重键: "https://arxiv.org/abs/2608.10684"
---

## Computer Science > Computer Vision and Pattern Recognition

## Title:Embedding Rotation Invariance for Provable Multi-Oriented Scene Text Recognition

Authors:[Zhibin Ma](https://arxiv.org/search/cs?searchtype=author&query=Ma,+Z), [Pengwen Dai](https://arxiv.org/search/cs?searchtype=author&query=Dai,+P), [Yi Liu](https://arxiv.org/search/cs?searchtype=author&query=Liu,+Y), [Xugong Qin](https://arxiv.org/search/cs?searchtype=author&query=Qin,+X), [Chenyun Yu](https://arxiv.org/search/cs?searchtype=author&query=Yu,+C), [Xiaochun Cao](https://arxiv.org/search/cs?searchtype=author&query=Cao,+X)

[View PDF](https://arxiv.org/pdf/2608.10684) [HTML (experimental)](https://arxiv.org/html/2608.10684v1)

> Abstract:Multi-oriented text is ubiquitous in real-world scenes and remains a major challenge for scene text recognition (STR). Existing rotation-aware methods explicitly estimate text orientation. However, due to the lack of theoretical guarantees, they are prone to error accumulation, increased computational cost, and strong reliance on data. In this work, we incorporate rotation invariance into the STR framework to address these limitations. Specifically, we adopt an encoder-decoder architecture, embedding rotation equivariance in the encoder and rotation invariance in the decoder to construct a fully rotation-invariant network. On the decoder side, we first identify and prove the rotation-invariant property of the cross-attention mechanism and use it to formulate a rotation-invariant text decoder that maps visual features to output text in a rotation-invariant manner. On the encoder side, we propose a rotation-equivariant local-global extraction network that integrates deep equivariant convolutions with self-attention, enabling rotation-equivariant feature extraction while modeling inter-character dependencies and preserving fine-grained visual details. By integrating the encoder and decoder, we obtain an end-to-end Rotation-Invariant Scene Text Recognition network (RISTER). RISTER provides rotation invariance with theoretical guarantees, enhancing robustness on multi-oriented samples without introducing additional inference computation or relying on data-driven orientation correction. Experiments show that RISTER achieves state-of-the-art performance on both standard and multi-oriented benchmarks, surpassing the second-best model by 4.0 percent in accuracy on the general multi-oriented dataset.

| Subjects: | Computer Vision and Pattern Recognition (cs.CV) |
| --- | --- |
| Cite as: | [arXiv:2608.10684](https://arxiv.org/abs/2608.10684) \[cs.CV\] |
|  | (or [arXiv:2608.10684v1](https://arxiv.org/abs/2608.10684v1) \[cs.CV\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.10684](https://doi.org/10.48550/arXiv.2608.10684) |

## Submission history

From: Zhibin Ma \[[view email](https://arxiv.org/show-email/506e177b/2608.10684)\]  
**\[v1\]** Tue, 11 Aug 2026 09:06:46 UTC (405 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.10684) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))