未验证 提交 18ddb6d5 编写于 作者: E Evezerest 提交者: GitHub

Merge pull request #8097 from Evezerest/dygraph

Update README and add new application document references
English | [简体中文](README_ch.md)
English | [简体中文](README_ch.md) | [हिन्दी](./doc/doc_i18n/README_हिन्द.md) | [日本語](./doc/doc_i18n/README_日本語.md) | [한국인](./doc/doc_i18n/README_한국어.md) | [Pу́сский язы́к](./doc/doc_i18n/README_Ру́сский_язы́к.md)
<p align="center">
<img src="./doc/PaddleOCR_log.png" align="middle" width = "600"/>
......@@ -25,25 +25,33 @@ PaddleOCR aims to create multilingual, awesome, leading, and practical OCR tools
<img src="./doc/imgs_results/ch_ppocr_mobile_v2.0/00006737.jpg" width="800">
</div>
## Recent updates
## 📣 Recent updates
- 💥 **Live Preview: Oct 24 - Oct 26, China Standard Time, 20:30**, Engineers@PaddleOCR will show PP-StructureV2 optimization strategy for 3 days.
- Scan the QR code below using WeChat, follow the PaddlePaddle official account and fill out the questionnaire to join the WeChat group, get the live link and 20G OCR learning materials (including PDF2Word application, 10 models in vertical scenarios, etc.)
<div align="center">
<img src="https://user-images.githubusercontent.com/50011306/196944258-0eb82df1-d730-4b96-a350-c1d370fdc2b1.jpg" width = "150" height = "150" />
</div>
- **🔥2022.8.24 Release PaddleOCR [release/2.6](https://github.com/PaddlePaddle/PaddleOCR/tree/release/2.6)**
- Release [PP-StructureV2](./ppstructure/),with functions and performance fully upgraded, adapted to Chinese scenes, and new support for [Layout Recovery](./ppstructure/recovery) and **one line command to convert PDF to Word**;
- [Layout Analysis](./ppstructure/layout) optimization: model storage reduced by 95%, while speed increased by 11 times, and the average CPU time-cost is only 41ms;
- [Table Recognition](./ppstructure/table) optimization: 3 optimization strategies are designed, and the model accuracy is improved by 6% under comparable time consumption;
- [Key Information Extraction](./ppstructure/kie) optimization:a visual-independent model structure is designed, the accuracy of semantic entity recognition is increased by 2.8%, and the accuracy of relation extraction is increased by 9.1%.
- **🔥2022.7 Release [OCR scene application collection](./applications/README_en.md)**
- **🔥2022.8 Release [OCR scene application collection](./applications/README_en.md)**
- Release **9 vertical models** such as digital tube, LCD screen, license plate, handwriting recognition model, high-precision SVTR model, etc, covering the main OCR vertical applications in general, manufacturing, finance, and transportation industries.
- **🔥2022.5.9 Release PaddleOCR [release/2.5](https://github.com/PaddlePaddle/PaddleOCR/tree/release/2.5)**
- **2022.8 Add implementation of [8 cutting-edge algorithms](doc/doc_en/algorithm_overview_en.md)**
- Text Detection: [FCENet](doc/doc_en/algorithm_det_fcenet_en.md), [DB++](doc/doc_en/algorithm_det_db_en.md)
- Text Recognition: [ViTSTR](doc/doc_en/algorithm_rec_vitstr_en.md), [ABINet](doc/doc_en/algorithm_rec_abinet_en.md), [VisionLAN](doc/doc_en/algorithm_rec_visionlan_en.md), [SPIN](doc/doc_en/algorithm_rec_spin_en.md), [RobustScanner](doc/doc_en/algorithm_rec_robustscanner_en.md)
- Table Recognition: [TableMaster](doc/doc_en/algorithm_table_master_en.md)
- **2022.5.9 Release PaddleOCR [release/2.5](https://github.com/PaddlePaddle/PaddleOCR/tree/release/2.5)**
- Release [PP-OCRv3](./doc/doc_en/ppocr_introduction_en.md#pp-ocrv3): With comparable speed, the effect of Chinese scene is further improved by 5% compared with PP-OCRv2, the effect of English scene is improved by 11%, and the average recognition accuracy of 80 language multilingual models is improved by more than 5%.
- Release [PPOCRLabelv2](./PPOCRLabel): Add the annotation function for table recognition task, key information extraction task and irregular text image.
- Release interactive e-book [*"Dive into OCR"*](./doc/doc_en/ocr_book_en.md), covers the cutting-edge theory and code practice of OCR full stack technology.
- [more](./doc/doc_en/update_en.md)
## Features
## 🌟 Features
PaddleOCR support a variety of cutting-edge algorithms related to OCR, and developed industrial featured models/solution [PP-OCR](./doc/doc_en/ppocr_introduction_en.md) and [PP-Structure](./ppstructure/README.md) on this basis, and get through the whole process of data production, model training, compression, inference and deployment.
......@@ -54,7 +62,7 @@ PaddleOCR support a variety of cutting-edge algorithms related to OCR, and devel
> It is recommended to start with the “quick experience” in the document tutorial
## Quick Experience
## Quick Experience
- Web online experience for the ultra-lightweight OCR: [Online Experience](https://www.paddlepaddle.org.cn/hub/scene/ocr)
- Mobile DEMO experience (based on EasyEdge and Paddle-Lite, supports iOS and Android systems): [Sign in to the website to obtain the QR code for installing the App](https://ai.baidu.com/easyedge/app/openSource?from=paddlelite)
......@@ -62,22 +70,23 @@ PaddleOCR support a variety of cutting-edge algorithms related to OCR, and devel
<a name="book"></a>
## E-book: *Dive Into OCR*
- [Dive Into OCR 📚](./doc/doc_en/ocr_book_en.md)
## 📚 E-book: *Dive Into OCR*
- [Dive Into OCR ](./doc/doc_en/ocr_book_en.md)
<a name="Community"></a>
## Community
## 👫 Community
- **Join us**👬: Scan the QR code below with your Wechat, you can join the official technical discussion group. Looking forward to your participation.
- For international developers, we regard [PaddleOCR Discussions](https://github.com/PaddlePaddle/PaddleOCR/discussions) as our international community platform. All ideas and questions can be discussed here in English.
- For Chinese develops, Scan the QR code below with your Wechat, you can join the official technical discussion group. For richer community content, please refer to [中文README](README_ch.md), looking forward to your participation.
<div align="center">
<img src="https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/dygraph/doc/joinus.PNG" width = "200" height = "200" />
<img src="https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/dygraph/doc/joinus.PNG" width = "150" height = "150" />
</div>
<a name="Supported-Chinese-model-list"></a>
## PP-OCR Series Model List(Update on September 8th)
## 🛠️ PP-OCR Series Model List(Update on September 8th)
| Model introduction | Model name | Recommended scene | Detection model | Direction classifier | Recognition model |
| ------------------------------------------------------------ | ---------------------------- | ----------------- | ------------------------------------------------------------ | ------------------------------------------------------------ | ------------------------------------------------------------ |
......@@ -92,7 +101,7 @@ PaddleOCR support a variety of cutting-edge algorithms related to OCR, and devel
- For a new language request, please refer to [Guideline for new language_requests](#language_requests).
- For structural document analysis models, please refer to [PP-Structure models](./ppstructure/docs/models_list_en.md).
## Tutorials
## 📖 Tutorials
- [Environment Preparation](./doc/doc_en/environment_en.md)
- [PP-OCR 🔥](./doc/doc_en/ppocr_introduction_en.md)
- [Quick Start](./doc/doc_en/quickstart_en.md)
......@@ -153,7 +162,7 @@ PaddleOCR support a variety of cutting-edge algorithms related to OCR, and devel
<a name="Visualization"></a>
## Visualization [more](./doc/doc_en/visualization_en.md)
## 👀 Visualization [more](./doc/doc_en/visualization_en.md)
<details open>
<summary>PP-OCRv3 Chinese model</summary>
......@@ -217,7 +226,7 @@ PaddleOCR support a variety of cutting-edge algorithms related to OCR, and devel
</details>
<a name="language_requests"></a>
## Guideline for New Language Requests
## 🇺🇳 Guideline for New Language Requests
If you want to request a new language support, a PR with 1 following files are needed:
......@@ -230,5 +239,5 @@ More details, please refer to [Multilingual OCR Development Plan](https://github
<a name="LICENSE"></a>
## License
## 📄 License
This project is released under <a href="https://github.com/PaddlePaddle/PaddleOCR/blob/master/LICENSE">Apache 2.0 license</a>
[English](README.md) | 简体中文
[English](README.md) | 简体中文 | [हिन्दी](./doc/doc_i18n/README_हिन्द.md) | [日本語](./doc/doc_i18n/README_日本語.md) | [한국인](./doc/doc_i18n/README_한국어.md) | [Pу́сский язы́к](./doc/doc_i18n/README_Ру́сский_язы́к.md)
<p align="center">
<img src="./doc/PaddleOCR_log.png" align="middle" width = "600"/>
......@@ -25,17 +25,31 @@ PaddleOCR旨在打造一套丰富、领先、且实用的OCR工具库,助力
<img src="./doc/imgs_results/ch_ppocr_mobile_v2.0/00006737.jpg" width="800">
</div>
## 近期更新
## 📣 近期更新
- **💥 直播预告:10.24-10.26日每晚8点半**,PaddleOCR研发团队详解PP-StructureV2优化策略。微信扫描下方二维码,关注公众号并填写问卷后进入官方交流群,获取直播链接与20G重磅OCR学习大礼包(内含PDF转Word应用程序、10种垂类模型、《动手学OCR》电子书等)
<div align="center">
<img src="https://user-images.githubusercontent.com/50011306/196944258-0eb82df1-d730-4b96-a350-c1d370fdc2b1.jpg" width = "150" height = "150" />
</div>
- **🔥2022.8.24 发布 PaddleOCR [release/2.6](https://github.com/PaddlePaddle/PaddleOCR/tree/release/2.6)**
- 发布[PP-StructureV2](./ppstructure/),系统功能性能全面升级,适配中文场景,新增支持[版面复原](./ppstructure/recovery),支持**一行命令完成PDF转Word**
- [版面分析](./ppstructure/layout)模型优化:模型存储减少95%,速度提升11倍,平均CPU耗时仅需41ms;
- [表格识别](./ppstructure/table)模型优化:设计3大优化策略,预测耗时不变情况下,模型精度提升6%;
- [关键信息抽取](./ppstructure/kie)模型优化:设计视觉无关模型结构,语义实体识别精度提升2.8%,关系抽取精度提升9.1%。
- 发布[PP-StructureV2](./ppstructure/README_ch.md),系统功能性能全面升级,适配中文场景,新增支持[版面复原](./ppstructure/recovery/README_ch.md),支持**一行命令完成PDF转Word**
- [版面分析](./ppstructure/layout/README_ch.md)模型优化:模型存储减少95%,速度提升11倍,平均CPU耗时仅需41ms;
- [表格识别](./ppstructure/table/README_ch.md)模型优化:设计3大优化策略,预测耗时不变情况下,模型精度提升6%;
- [关键信息抽取](./ppstructure/kie/README_ch.md)模型优化:设计视觉无关模型结构,语义实体识别精度提升2.8%,关系抽取精度提升9.1%。
- **🔥2022.8 发布 [OCR场景应用集合](./applications)**
- 包含数码管、液晶屏、车牌、高精度SVTR模型、手写体识别等**9个垂类模型**,覆盖通用,制造、金融、交通行业的主要OCR垂类应用。
- **2022.8 新增实现[8种前沿算法](doc/doc_ch/algorithm_overview.md)**
- 文本检测:[FCENet](doc/doc_ch/algorithm_det_fcenet.md), [DB++](doc/doc_ch/algorithm_det_db.md)
- 文本识别:[ViTSTR](doc/doc_ch/algorithm_rec_vitstr.md), [ABINet](doc/doc_ch/algorithm_rec_abinet.md), [VisionLAN](doc/doc_ch/algorithm_rec_visionlan.md), [SPIN](doc/doc_ch/algorithm_rec_spin.md), [RobustScanner](doc/doc_ch/algorithm_rec_robustscanner.md)
- 表格识别:[TableMaster](doc/doc_ch/algorithm_table_master.md)
- **2022.5.9 发布 PaddleOCR [release/2.5](https://github.com/PaddlePaddle/PaddleOCR/tree/release/2.5)**
- 发布[PP-OCRv3](./doc/doc_ch/ppocr_introduction.md#pp-ocrv3),速度可比情况下,中文场景效果相比于PP-OCRv2再提升5%,英文场景提升11%,80语种多语言模型平均识别准确率提升5%以上;
- 发布半自动标注工具[PPOCRLabelv2](./PPOCRLabel):新增表格文字图像、图像关键信息抽取任务和不规则文字图像的标注功能;
......@@ -44,7 +58,7 @@ PaddleOCR旨在打造一套丰富、领先、且实用的OCR工具库,助力
> [更多](./doc/doc_ch/update.md)
## 特性
## 🌟 特性
支持多种OCR相关前沿算法,在此基础上打造产业级特色模型[PP-OCR](./doc/doc_ch/ppocr_introduction.md)[PP-Structure](./ppstructure/README_ch.md),并打通数据生产、模型训练、压缩、预测部署全流程。
......@@ -55,25 +69,26 @@ PaddleOCR旨在打造一套丰富、领先、且实用的OCR工具库,助力
> 上述内容的使用方法建议从文档教程中的快速开始体验
## 快速开始
## 快速开始
- 在线网站体验:超轻量PP-OCR mobile模型体验地址:https://www.paddlepaddle.org.cn/hub/scene/ocr
- 移动端demo体验:[安装包DEMO下载地址](https://ai.baidu.com/easyedge/app/openSource?from=paddlelite)(基于EasyEdge和Paddle-Lite, 支持iOS和Android系统)
- 一行命令快速使用:[快速开始(中英文/多语言/文档分析)](./doc/doc_ch/quickstart.md)
<a name="电子书"></a>
## 《动手学OCR》电子书
- [《动手学OCR》电子书📚](./doc/doc_ch/ocr_book.md)
## 📚《动手学OCR》电子书
- [《动手学OCR》电子书](./doc/doc_ch/ocr_book.md)
<a name="开源社区"></a>
## 开源社区
- **项目合作📑:** 如果您是企业开发者且有明确的OCR垂类应用需求,填写[问卷](https://paddle.wjx.cn/vj/QwF7GKw.aspx)后可免费与官方团队展开不同层次的合作。
- **加入社区👬:** 微信扫描二维码并填写问卷之后,加入交流群领取福利
- **获取PaddleOCR最新发版解说《OCR超强技术详解与产业应用实战》系列直播课回放链接**
- **10G重磅OCR学习大礼包:**《动手学OCR》电子书,配套讲解视频和notebook项目;66篇OCR相关顶会前沿论文打包放送,包括CVPR、AAAI、IJCAI、ICCV等;PaddleOCR历次发版直播课视频;OCR社区优秀开发者项目分享视频。
- **社区项目**🏅️:[社区项目](./doc/doc_ch/thirdparty.md)文档中包含了社区用户**使用PaddleOCR开发的各种工具、应用**以及**为PaddleOCR贡献的功能、优化的文档与代码**等,是官方为社区开发者打造的荣誉墙,也是帮助优质项目宣传的广播站。
- **社区常规赛**🎁:社区常规赛是面向OCR开发者的积分赛事,覆盖文档、代码、模型和应用四大类型,以季度为单位评选并发放奖励,赛题详情与报名方法可参考[链接](https://github.com/PaddlePaddle/PaddleOCR/issues/4982)
## 👫 开源社区
- **📑项目合作:** 如果您是企业开发者且有明确的OCR垂类应用需求,填写[问卷](https://paddle.wjx.cn/vj/QwF7GKw.aspx)后可免费与官方团队展开不同层次的合作。
- **👫加入社区:** 微信扫描二维码并填写问卷之后,加入交流群领取20G重磅OCR学习大礼包
- **包括《动手学OCR》电子书** ,配套讲解视频和notebook项目;PaddleOCR历次发版直播课视频;
- **OCR场景应用模型集合:** 包含数码管、液晶屏、车牌、高精度SVTR模型、手写体识别等垂类模型,覆盖通用,制造、金融、交通行业的主要OCR垂类应用。
- PDF2Word应用程序;OCR社区优秀开发者项目分享视频。
- **🏅️社区项目**[社区项目](./doc/doc_ch/thirdparty.md)文档中包含了社区用户**使用PaddleOCR开发的各种工具、应用**以及**为PaddleOCR贡献的功能、优化的文档与代码**等,是官方为社区开发者打造的荣誉墙,也是帮助优质项目宣传的广播站。
- **🎁社区常规赛**:社区常规赛是面向OCR开发者的积分赛事,覆盖文档、代码、模型和应用四大类型,以季度为单位评选并发放奖励,赛题详情与报名方法可参考[链接](https://github.com/PaddlePaddle/PaddleOCR/issues/4982)
<div align="center">
<img src="https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/dygraph/doc/joinus.PNG" width = "150" height = "150" />
......@@ -82,7 +97,7 @@ PaddleOCR旨在打造一套丰富、领先、且实用的OCR工具库,助力
<a name="模型下载"></a>
## PP-OCR系列模型列表(更新中)
## 🛠️ PP-OCR系列模型列表(更新中)
| 模型简介 | 模型名称 | 推荐场景 | 检测模型 | 方向分类器 | 识别模型 |
| ------------------------------------- | ----------------------- | --------------- | ------------------------------------------------------------ | ------------------------------------------------------------ | ------------------------------------------------------------ |
......@@ -103,7 +118,7 @@ PaddleOCR旨在打造一套丰富、领先、且实用的OCR工具库,助力
<a name="文档教程"></a>
## 文档教程
## 📖 文档教程
- [运行环境准备](./doc/doc_ch/environment.md)
- [PP-OCR文本检测识别🔥](./doc/doc_ch/ppocr_introduction.md)
......@@ -169,7 +184,7 @@ PaddleOCR旨在打造一套丰富、领先、且实用的OCR工具库,助力
<a name="效果展示"></a>
## 效果展示 [more](./doc/doc_ch/visualization.md)
## 👀 效果展示 [more](./doc/doc_ch/visualization.md)
<details open>
<summary>PP-OCRv3 中文模型</summary>
......
......@@ -43,13 +43,13 @@ PaddleOCR场景应用覆盖通用,制造、金融、交通行业的主要OCR
### 金融
| 类别 | 亮点 | 模型下载 | 教程 | 示例图 |
| -------------- | ----------------------------- | -------------- | ------------------------------------- | ------------------------------------------------------------ |
| -------------- | ----------------------------- | -------------- | ----------------------------------------- | ------------------------------------------------------------ |
| 表单VQA | 多模态通用表单结构化提取 | [模型下载](#2) | [中文](./多模态表单识别.md)/English | <img src="https://ai-studio-static-online.cdn.bcebos.com/a3b25766f3074d2facdf88d4a60fc76612f51992fd124cf5bd846b213130665b" width = "200" height = "200" /> |
| 增值税发票 | 关键信息抽取,SER、RE任务训练 | [模型下载](#2) | [中文](./发票关键信息抽取.md)/English | <img src="https://user-images.githubusercontent.com/14270174/185393805-c67ff571-cf7e-4217-a4b0-8b396c4f22bb.jpg" width = "200" /> |
| 印章检测与识别 | 端到端弯曲文本识别 | | | |
| 通用卡证识别 | 通用结构化提取 | | | |
| 印章检测与识别 | 端到端弯曲文本识别 | [模型下载](#2) | [中文](./印章弯曲文字识别.md)/English | <img src="https://ai-studio-static-online.cdn.bcebos.com/498119182f0a414ab86ae2de752fa31c9ddc3a74a76847049cc57884602cb269" width = "150" /> |
| 通用卡证识别 | 通用结构化提取 | [模型下载](#2) | [中文](./快速构建卡证类OCR.md)/English | <img src="https://ai-studio-static-online.cdn.bcebos.com/981640e17d05487e961162f8576c9e11634ca157f79048d4bd9d3bc21722afe8" width = "300" /> |
| 身份证识别 | 结构化提取、图像阴影 | | | |
| 合同比对 | 密集文本检测、NLP串联 | | | |
| 合同比对 | 密集文本检测、NLP关键信息抽取 | [模型下载](#2) | [中文](./扫描合同关键信息提取.md)/English | <img src="https://ai-studio-static-online.cdn.bcebos.com/54f3053e6e1b47a39b26e757006fe2c44910d60a3809422ab76c25396b92e69b" width = "300" /> |
<a name="14"></a>
......
......@@ -43,13 +43,13 @@ PaddleOCR scene application covers general, manufacturing, finance, transportati
### Finance
| Case | Feature | Model Download | Tutorial | Example |
| ----------------------------------- | -------------------------------------------------- | -------------------- | ------------------------------------- | ------------------------------------------------------------ |
| ----------------------------------- | -------------------------------------------------- | -------------------- | ----------------------------------------- | ------------------------------------------------------------ |
| Form visual question and answer | Multimodal general form structured extraction | [Model Download](#2) | [中文](./多模态表单识别.md)/English | <img src="https://ai-studio-static-online.cdn.bcebos.com/a3b25766f3074d2facdf88d4a60fc76612f51992fd124cf5bd846b213130665b" width = "200" height = "200" /> |
| VAT invoice | Key information extraction, SER, RE task fine-tune | [Model Download](#2) | [中文](./发票关键信息抽取.md)/English | <img src="https://user-images.githubusercontent.com/14270174/185393805-c67ff571-cf7e-4217-a4b0-8b396c4f22bb.jpg" width = "200" /> |
| Seal detection and recognition | End-to-end curved text recognition | | | |
| Universal card recognition | Universal structured extraction | | | |
| Seal detection and recognition | End-to-end curved text recognition | [Model Download](#2) | [中文](./印章弯曲文字识别.md)/English | <img src="https://ai-studio-static-online.cdn.bcebos.com/498119182f0a414ab86ae2de752fa31c9ddc3a74a76847049cc57884602cb269" width = "150" /> |
| Universal card recognition | Universal structured extraction | [Model Download](#2) | [中文](./快速构建卡证类OCR.md)/English | <img src="https://ai-studio-static-online.cdn.bcebos.com/981640e17d05487e961162f8576c9e11634ca157f79048d4bd9d3bc21722afe8" width = "300" /> |
| ID card recognition | Structured extraction, image shading | | | |
| Contract key information extraction | Dense text detection, NLP concatenation | | | |
| Contract key information extraction | Dense text detection, NLP concatenation | [Model Download](#2) | [中文](./扫描合同关键信息提取.md)/English | <img src="https://ai-studio-static-online.cdn.bcebos.com/54f3053e6e1b47a39b26e757006fe2c44910d60a3809422ab76c25396b92e69b" width = "300" /> |
<a name="14"></a>
......
......@@ -1029,5 +1029,6 @@ Train:
如需获取已训练模型,请扫下图二维码填写问卷,加入PaddleOCR官方交流群获取全部OCR垂类模型下载链接、《动手学OCR》电子书等全套OCR学习资料🎁
![](https://ai-studio-static-online.cdn.bcebos.com/ea32877b717643289dc2121a2e573526d99d0f9eecc64ad4bd8dcf121cb5abde)
<div align="center">
<img src="https://ai-studio-static-online.cdn.bcebos.com/ea32877b717643289dc2121a2e573526d99d0f9eecc64ad4bd8dcf121cb5abde" width = "150" height = "150" />
</div>
......@@ -84,11 +84,6 @@
* 模型来自模型库PaddleClas,它是一个图像识别和图像分类任务的工具集,助力使用者训练出更好的视觉模型和应用落地。
![](https://ai-studio-static-online.cdn.bcebos.com/606d1afaf0d0484a99b1d39895d394b22f24e74591514796859a9ea3a2799b78)
### 2.2 卡证识别模型
* 检测:DBNet 识别:SVRT
......@@ -179,8 +174,6 @@ C [下载预训练模型 ](https://github.com/PaddlePaddle/PaddleClas/blob/relea
PaddleOCR是一个丰富、领先且实用的OCR工具库,助力开发者训练出更好的模型并应用落地
![](https://ai-studio-static-online.cdn.bcebos.com/16c5e16d53b8428c95129cac4f5520204d869910247943e494d854227632e882)
身份证识别:用现有的方法识别
......
[English](../../README.md) | [简体中文](../../README_ch.md) | [हिन्दी](./README_हिन्द.md) | [日本語](./README_日本語.md) | [한국인](./README_한국어.md) | Pу́сский язы́к
<p align="center">
<img src="../PaddleOCR_log.png" align="middle" width = "600"/>
<p align="center">
<p align="left">
<a href="./LICENSE"><img src="https://img.shields.io/badge/license-Apache%202-dfd.svg"></a>
<a href="https://github.com/PaddlePaddle/PaddleOCR/releases"><img src="https://img.shields.io/github/v/release/PaddlePaddle/PaddleOCR?color=ffa"></a>
<a href=""><img src="https://img.shields.io/badge/python-3.7+-aff.svg"></a>
<a href=""><img src="https://img.shields.io/badge/os-linux%2C%20win%2C%20mac-pink.svg"></a>
<a href=""><img src="https://img.shields.io/pypi/format/PaddleOCR?color=c77"></a>
<a href="https://pypi.org/project/PaddleOCR/"><img src="https://img.shields.io/pypi/dm/PaddleOCR?color=9cf"></a>
<a href="https://github.com/PaddlePaddle/PaddleOCR/stargazers"><img src="https://img.shields.io/github/stars/PaddlePaddle/PaddleOCR?color=ccf"></a>
</p>
## Введение
PaddleOCR стремится создавать многоязычные, потрясающие, передовые и практичные инструменты OCR, которые помогают пользователям обучать лучшие модели и применять их на практике
<div align="center">
<img src="https://user-images.githubusercontent.com/50011306/187821591-6cb09459-fdbf-4ad3-8c5a-26af611c211d.png" width="800">
</div>
<div align="center">
<img src="../imgs_results/PP-OCRv3/en/en_4.png" width="800">
</div>
<div align="center">
<img src="../imgs_results/ch_ppocr_mobile_v2.0/00006737.jpg" width="800">
</div>
## 📣 Последние обновления
- **🔥2022.8.24 Выпуск PaddleOCR [Выпуск /2.6](https://github.com/PaddlePaddle/PaddleOCR/tree/release/2.6)**
- Выпускать [PP-Structurev2](./ppstructure/),с полностью обновленными функциями и производительностью, адаптированными для китайских сцен и новой поддержкой pаспознавание таблиц
[Восстановление макета](./ppstructure/recovery) и **однострочная команда для преобразования PDF в Word**;
- [Анализ макета](./ppstructure/layout) оптимизация: память модели уменьшена на 95%, а скорость увеличена в 11 раз, а среднее время процессорного времени составляет всего 41 мс;
- [Распознавание таблиц](./ppstructure/table) оптимизация: разработано 3 стратегии оптимизации, а точность модели улучшена на 6% при сопоставимых затратах времени;
- [Извлечение ключевой информации](./ppstructure/kie) оптимизация: разработана визуально независимая структура модели, точность распознавания семантической сущности увеличена на 2,8%, а точность извлечения отношения увеличена на 9,1%.
- **🔥2022.7 Выпуск [Коллекция приложений сцены OCR](../../applications/README_en.md)**
- Выпуск **9 вертикальных моделей**, таких как цифровая трубка, ЖК-экран, номерной знак, модель распознавания рукописного ввода, высокоточная модель SVTR и т. д., охватывающих основные вертикальные приложения OCR в целом, производственной, финансовой и транспортной отраслях.
- **🔥2022.5.9 Выпуск PaddleOCR [Выпуск /2.5](https://github.com/PaddlePaddle/PaddleOCR/tree/release/2.5)**
- Выпускать [PP-OCRv3](../doc_en/ppocr_introduction_en.md#pp-ocrv3): При сопоставимой скорости эффект китайской сцены улучшен на 5% по сравнению с ПП-OCRRv2, эффект английской сцены улучшен на 11%, а средняя точность распознавания 80 языковых многоязычных моделей улучшена более чем на 5%.
- Выпускать [PPOCRLabelv2](./PPOCRLabel): Добавьте функцию аннотации для задачи распознавания таблиц, задачи извлечения ключевой информации и изображения неправильного текста.
- Выпустить интерактивную электронную книгу [*"Погружение в OCR"*](../doc_en/ocr_book_en.md), охватывает передовую теорию и практику кодирования технологии полного стека OCR.
- [подробнее](../doc_en/update_en.md)
## 🌟 Функции
PaddleOCR поддерживает множество передовых алгоритмов, связанных с распознаванием текста, и разработала промышленные модели/решения. [PP-OCR](../doc_en/ppocr_introduction_en.md) и [PP-Structure](./ppstructure/README.md) на этой основе и пройти весь процесс производства данных, обучения модели, сжатия, логического вывода и развертывания.
<div align="center">
<img src="https://user-images.githubusercontent.com/50011306/196963669-f53b0ee5-3cb4-481c-b73c-97c4b3e2efb8.png">
</div>
## ⚡ Быстрый опыт
```bash
pip3 install paddlepaddle # for gpu user please install paddlepaddle-gpu
pip3 install paddleocr
paddleocr --image_dir /your/test/image.jpg --lang=ru
```
> Если у вас нет среды Python, выполните [Подготовка среды](../doc_en/environment_en.md). Мы рекомендуем вам начать с [Учебники](#Tutorials).
<a name="книга"></a>
## 📚 Электронная книга: *Погружение в OCR*
- [Погружение в распознавание символов](../doc_en/ocr_book_en.md)
<a name="Сообщество"></a>
## 👫 Сообщество
Что касается международных разработчиков, мы рассматриваем [Обсуждения PaddleOCR](https://github.com/PaddlePaddle/PaddleOCR/discussions) как нашу платформу для международного сообщества. Все идеи и вOCRосы можно обсудить здесь на английском языке.
<a name="Список-поддерживаемых-китайских-моделей"></a>
## 🛠️ Список моделей серии ПП -OCR
| Введение модели | Название модели | Рекомендуемая сцена | Модель обнаружения | Модель распознавания |
| ------------------------------------------------------------ | ---------------------------- | ----------------- | ------------------------------------------------------------ | ------------------------------------------------------------ |
| Ру́сский язы́к:Ру́сский язы́к Сверхлегкая модель ПП-OCRv3 (13.4M) | cyrillic_PP-OCRv3_xx | Мобильный и сервер |[модель вывода](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/Multilingual_PP-OCRv3_det_infer.tar)/[обученный модель](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/Multilingual_PP-OCRv3_det_distill_train.tar) | [модель вывода](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/cyrillic_PP-OCRv3_rec_infer.tar)/[обученный модель](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/cyrillic_PP-OCRv3_rec_train.tar) |
| Английский сверхлегкая модель ПП-OCRv3 (13,4 Мб) | en\_ПП-OCRv3_xx |Мобильный и сервер | [модель вывода](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_det_infer.tar) / [обученный модель](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_det_distill_train.tar) | [вывод модель](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_rec_infer.tar) / [обученный модель](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_rec_train.tar) |
| Сверхлегкая китайская и английская модель ПП-OCRv3 (16,2M) | ch\_ПП-OCRv3_xx | Мобильный и сервер | [вывод модель](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_infer.tar) / [обученный модель](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_distill_train.tar) | [вывод модель](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_rec_infer.tar) / [обученный модель](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_rec_train.tar) |
- Для получения дополнительных загрузок моделей (включая несколько языков) см. [Загрузки моделей серии ПП-OCR](../doc_en/models_list_en.md).
- Для запроса нового языка см [Руководство для новых языковых_запросов](#language_requests).
- Модели структурного анализа документов см [PP-Structure модельs](./ppstructure/docs/модельs_list_en.md).
<a name=" Учебники "></a>
## 📖 Учебники
- [Подготовка окружающей среды](../doc_en/environment_en.md)
- [PP-OCR 🔥](../doc_en/ppocr_introduction_en.md)
- [Быстрый старт](../doc_en/quickstart_en.md)
- [Модель Zoo](../doc_en/модельs_en.md)
- [Модель тренировки](../doc_en/training_en.md)
- [Обнаружение текста](../doc_en/detection_en.md)
- [Распознавание текста](../doc_en/recognition_en.md)
- [Классификация направления текста](../doc_en/angle_class_en.md)
- Модель Сжатие
- [Модель квантования](./deploy/slim/quantization/README_en.md)
- [Модель Обрезка](./deploy/slim/prune/README_en.md)
- [Дистилляция знаний](../doc_en/knowledge_distillation_en.md)
- [Вывод и развертывание](./deploy/README.md)
- [Python Вывод](../doc_en/ inference _ppocr_en.md)
- [Вывод C++](./deploy/cpp_infer/readme.md)
-[Подача](./deploy/pdserving/README.md)
- [Мобильный](./deploy/lite/readme.md)
- [Paddle2ONNX](./deploy/paddle2onnx/readme.md)
-[ВеслоОблако](./deploy/paddlecloud/README.md)
- [Benchmark](../doc_en/benchmark_en.md)
- [PP-Structure 🔥](../../ppstructure/README.md)
- [Быстрый старт](../../ppstructure/docs/quickstart_en.md)
- [Модель Zoo](../../ppstructure/docs/models_list_en.md)
- [Модель тренировки](../doc_en/training_en.md)
- [Анализ макета](../../ppstructure/layout/README.md)
- [Распознавание таблиц](../../ppstructure/table/README.md)
- [Извлечение ключевой информации](../../ppstructure/kie/README.md)
- [Вывод и развертывание](./deploy/README.md)
- [Вывод Python](../../ppstructure/docs/inference_en.md)
- [Вывод С++](../../deploy/cpp_infer/readme.md)
- [Обслуживание](../../deploy/hubserving/readme_en.md)
- [Академические алгоритмы](../doc_en/algorithm_overview_en.md)
- [Обнаружение текста](../doc_en/algorithm_overview_en.md)
- [Распознавание текста](../doc_en/algorithm_overview_en.md)
- [Непрерывной цепью OCR](../doc_en/algorithm_overview_en.md)
- [Распознавание таблиц](../doc_en/algorithm_overview_en.md)
- [Извлечение ключевой информации](../doc_en/algorithm_overview_en.md)
- [Добавьте новые алгоритмы в PaddleOCR](../doc_en/add_new_algorithm_en.md)
- Аннотации и синтез данных
- [Полуавтоматический инструмент аннотации данных: метка ППOCRR](./PPOCRLabel/README.md)
- [Инструмент синтеза данных: Стиль-текст](./StyleText/README.md)
- [Другие инструменты аннотирования данных](../doc_en/data_annotation_en.md)
- [Другие инструменты синтеза данных](../doc_en/data_synthesis_en.md)
- Наборы данных
- [Общие наборы данных OCR (китайский/английский)](../doc_en/dataset/datasets_en.md)
- [Наборы данных Рукописный/_OCR_ наборы данных (китайский)](../doc_en/dataset/handwritten_datasets_en.md)
- [Различные наборы данных OCR (многоязычные)](../doc_en/dataset/vertical_and_multilingual_datasets_en.md)
- [Анализ макета](../doc_en/dataset/layout_datasets_en.md)
- [Распознавание таблиц](../doc_en/dataset/table_datasets_en.md)
- [Извлечение ключевой информации](../doc_en/dataset/kie_datasets_en.md)
- [Структура кода](../doc_en/tree_en.md)
- [Визуализация](#Visualization)
- [Сообщество](#Community)
- [Новые языковые запросы](#language_requests)
- [ЧАСТО ЗАДАВАЕМЫЕ ВOCRОСЫ](../doc_en/FAQ_en.md)
- [Использованная литература](../doc_en/reference_en.md)
- [ЛИЦЕНЗИЯ](#LICENSE)
<a name="language_requests"></a>
## 🇺🇳 Руководство по запросам на новый язык
Если вы хотите **запросить новую языковую модель**, проголосуйте в [Голосуйте за обновление многоязычной модели](https://github.com/PaddlePaddle/PaddleOCR/discussions/7253). Мы будем регулярно обновлять модель по результату. **Пригласите друзей проголосовать вместе!**
Если вам нужно **обучить новую языковую модель** на основе вашего сценария, учебное пособие в [Проекте обучения многоязычной модели](https://github.com/PaddlePaddle/PaddleOCR/discussions/7252) поможет вам подготовить набор данных и показать вам весь процесс шаг за шагом.
Оригинальный [Многоязычный план разработки OCR](https://github.com/PaddlePaddle/PaddleOCR/issues/1048) по-прежнему показывает вам много полезных корпусов и словарей.
<a name=" Визуализация "></a>
## 👀 Визуализация [больше](../doc_en/visualization_en.md)
<details open>
<summary>PP-OCRv3 Многоязычная модель </summary>
<div align="center">
<img src="../imgs_results/PP-OCRv3/multi_lang/japan_2.jpg" width="800">
<img src="../imgs_results/PP-OCRv3/multi_lang/korean_1.jpg" width="800">
</div>
</details>
<details open>
<summary>PP-OCRv3 Aнглийская модель </summary>
<div align="center">
<img src="../imgs_results/PP-OCRv3/en/en_1.png" width="800">
<img src="../imgs_results/PP-OCRv3/en/en_2.png" width="800">
</div>
</details>
<details open>
<summary>PP-OCRv3 Kитайская модель </summary>
<div align="center">
<img src="../imgs_results/PP-OCRv3/ch/PP-OCRv3-pic001.jpg" width="800">
<img src="../imgs_results/PP-OCRv3/ch/PP-OCRv3-pic002.jpg" width="800">
<img src="../imgs_results/PP-OCRv3/ch/PP-OCRv3-pic003.jpg" width="800">
</div>
</details>
<details open>
<summary>PP-Structurev2</summary>
1. анализ макета + распознавание таблиц
<div align="center">
<img src="../../ppstructure/docs/table/ppstructure.GIF" width="800">
</div>
2. SER (Семантическое распознавание объектов)
<div align="center">
<img src="https://user-images.githubusercontent.com/25809855/186094456-01a1dd11-1433-4437-9ab2-6480ac94ec0a.png" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185310636-6ce02f7c-790d-479f-b163-ea97a5a04808.jpg" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185539517-ccf2372a-f026-4a7c-ad28-c741c770f60a.png" width="600">
</div>
3. RE (Извлечение отношений)
<div align="center">
<img src="https://user-images.githubusercontent.com/25809855/186094813-3a8e16cc-42e5-4982-b9f4-0134dfb5688d.png" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185393805-c67ff571-cf7e-4217-a4b0-8b396c4f22bb.jpg" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185540080-0431e006-9235-4b6d-b63d-0b3c6e1de48f.jpg" width="600">
</div>
</details>
<a name="ЛИЦЕНЗИЯ"></a>
## 📄 Лицензия
Этот проект выпущен под <a href="https://github.com/PaddlePaddle/PaddleOCR/blob/master/LICENSE">Apache 2.0 license</a>
[English](../../README.md) | [简体中文](../../README_ch.md) | हिन्दी | [日本語](./README_日本語.md) | [한국인](./README_한국어.md) | [Pу́сский язы́к](./README_Ру́сский_язы́к.md)
<p align="center">
<img src="../PaddleOCR_log.png" align="middle" width = "600"/>
<p align="center">
<p align="left">
<a href="./LICENSE"><img src="https://img.shields.io/badge/license-Apache%202-dfd.svg"></a>
<a href="https://github.com/PaddlePaddle/PaddleOCR/releases"><img src="https://img.shields.io/github/v/release/PaddlePaddle/PaddleOCR?color=ffa"></a>
<a href=""><img src="https://img.shields.io/badge/python-3.7+-aff.svg"></a>
<a href=""><img src="https://img.shields.io/badge/os-linux%2C%20win%2C%20mac-pink.svg"></a>
<a href=""><img src="https://img.shields.io/pypi/format/PaddleOCR?color=c77"></a>
<a href="https://pypi.org/project/PaddleOCR/"><img src="https://img.shields.io/pypi/dm/PaddleOCR?color=9cf"></a>
<a href="https://github.com/PaddlePaddle/PaddleOCR/stargazers"><img src="https://img.shields.io/github/stars/PaddlePaddle/PaddleOCR?color=ccf"></a>
</p>
## प्रस्तावना
पैडलओसीआर का उद्देश्य बहुभाषी,शानदार , ओसीआर और व्यावहारिक ओसीआरउपकरण बनाना है जो यूजर्स को बेहतर मॉडलों के लिए प्रशिक्षित करने और उन्हें व्यवहार में लागू करने में मदद करते हैं।
<div align="center">
<img src="https://user-images.githubusercontent.com/50011306/187821591-6cb09459-fdbf-4ad3-8c5a-26af611c211d.png" width="800">
</div>
<div align="center">
<img src="../imgs_results/PP-OCRv3/en/en_4.png" width="800">
</div>
<div align="center">
<img src="../imgs_results/ch_ppocr_mobile_v2.0/00006737.jpg" width="800">
</div>
## 📣 हाल के अद्यतन
- **🔥2022.8.24 रिलीज Paddleओसीआर [रिलीज/2.6](https://github.com/PaddlePaddle/PaddleOCR/tree/release/2.6)**
- रिलीज [PP-Structurev2](../../ppstructure),फंक्शन और परफॉरमेंस के साथ पूरी तरह से उन्नत, चायनीज शीन्स के अनुकूल, और मदद के लिए [लेआउट रिकवरी](../../ppstructure/recovery) और **पीडीएफ को वर्ड में बदलने के लिए वन लाइन कमांड**;
- [लेआउट एनालाइस](../../ppstructure/layout) ऑप्टिमाइजेशन: मॉडल स्टोरेज में 95% की कमी, जबकि स्पीड में 11 गुना वृद्धि , और एवरेज CPU स टाइम-कॉस्ट केवल 41ms है;
- [टेबल रिकोगनाइजेशन](../../ppstructure/table) ऑप्टिमाइजेशन: 3 ऑप्टिमाइज़ेशन के तरीके डिजाइन किए गए हैं, और तुलनात्मक समय की खपत के तहत मॉडल सटीकता में 6% का सुधार हुआ है;
- [की इंफॉर्मेशन एक्स्ट्रेक्शन](../../ppstructure/kie) ऑप्टिमाइजेशन : एक बिजुवल-स्वतंत्र मॉडल संरचना डिजाइन की गई है, सिमेंटिक एन्टाइटी रिकग्निशन की सटीकता में 2.8% की वृद्धि हुई है, और रिलेशन एक्सट्रैक्शन की सटीकता में 9.1% की वृद्धि हुई है।
- **🔥2022.7 रिलीज [ओसीआर दृश्य आवेदन संग्रह](../../applications/README_en.md)**
- रिलीज **9 वर्टिकल मॉडल** जैसे कि डिजिटल ट्यूब, एलसीडी स्क्रीन, लाइसेंस प्लेट, हस्तलेखन पहचान मॉडल, उच्च-सटीक एसवीटीआर मॉडल, आदि, जो सामान्य रूप से मुख्य ओसीआर वर्टिकल अनुप्रयोगों, विनिर्माण, वित्त और परिवहन उद्योगों को कवर करते हैं।
- **🔥2022.5.9 रिलीज Paddleओसीआर [रिलीज/2.5](https://github.com/PaddlePaddle/Paddleओसीआर/tree/release/2.5)**
- रिलीज [PP-OCRv3](../doc_en/ppocr_introduction_en.md#pp-ocrv3): तुलनात्मक स्पीड के साथ, चाइनीज शीन्स का प्रभाव PP-ओसीआर v2 की तुलना में 5% की और वृद्धि हुयी है इंगलिस शीन्स के प्रभाव में 11% का सुधार हुआ है, और 80 भाषाओं के बहुभाषी मॉडलों की औसत पहचान सटीकता में 5% से अधिक सुधार हुआ है।
- रिलीज़ [PPOCRLabelv2](./PPOCRLabel): टेबल टेबल रिकोगनाइजेशन टास्क की इंफॉर्मेशन एक्स्ट्रेक्शन टास्क और अनियमित टेक्सट इमेज के लिए एनोटेशन फ़ंक्शन एड करे।
- इंटरएक्टिव ई-बुक जारी करें [*"ओसीआर में गोता लगाएँ"*](../doc_en/ocr_book_en.md), ओसीआर पूर्ण स्टैक तकनीक के अत्याधुनिक सिद्धांत और कोड प्रेक्टिस को कवर करता है।
- [और अधिक](../doc_en/update_en.md)
## 🌟 विशेषताएँ
Paddleओसीआर से संबंधित विभिन्न प्रकार के अत्याधुनिक एल्गोरिथ्म को सपोर्ट करता है, और विकसित औद्योगिक विशेष रुप से प्रदर्शित मॉडल/समाधान [PP- OCR](../doc_en/ppocr_introduction_en.md) और [PP-Structure](../../ppstructure/README.md) इस आधार पर और डेटा प्रोडक्शन की पूरी प्रोसेस के माध्यम से प्राप्त करें, मॉडल ट्रेनिंग, दबाव, अनुमान और तैनाती।
<div align="center">
<img src="https://user-images.githubusercontent.com/50011306/196920323-9d386ab0-1233-4415-8508-99d459d256bb.png">
</div>
## ⚡ शीघ्र अनुभव
```bash
pip3 install paddlepaddle # for gpu user please install paddlepaddle-gpu
pip3 install paddleocr
paddleocr --image_dir /your/test/image.jpg --lang=hi
```
> यदि आपके पास पायथन एनवायरनमेंट नहीं है, कृपया फॉलो कीजिए [एनवायरनमेंट प्रिपेरेशन](../doc_en/environment_en.md). हम अनुशंसा करते हैं कि आप इसके साथ शुरुआत करें [ट्यूटोरियल](#Tutorials).
<a name="किताब"></a>
## 📚 ई-बुक: *ओसीआर में गोता लगाएँ*
- [ओसीआर में गोता लगाएँ](../doc_en/ocr_book_en.md)
<a name="कम्युनिटी"></a>
## 👫 समुदाय
अंतरराष्ट्रीय डेवलपर्स के लिए, हम सम्मान करते हैं [पैडलओसीआर चर्चाएँ] (https://github.com/PaddlePaddle/PaddleOCR/discussions) हमारे अंतरराष्ट्रीय कम्युनिटी मंच के रूप में। यहां सभी विचारों और प्रश्नों पर अंग्रेजी में चर्चा की जा सकती है।
<a name="समर्थित-चीनी-मॉडल-लिस्ट"></a>
## 🛠️ PP-ओसीआर श्रृंखला मॉडल सूची
| मॉडल प्रस्तावना | मॉडल नाम | रिकमेंडिड सीन | डिटेक्शन मॉडल | रिकोगनाइजेशन मॉडल |
| ------------------------------------------------------------ | ---------------------------- | ----------------- | ------------------------------------------------------------ | ------------------------------------------------------------ |
| हिन्दी:हिन्दी अल्ट्रा-लाइटवेट PP-ओसीआरv3 सिस्टम (9.9M) | devanagari_PP-ओसीआरv3_xx | मोबाइल और सर्वर |[इन्फरन्स मॉडल](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/Multilingual_PP-OCRv3_det_infer.tar) / [प्रशिक्षितमॉडल](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/Multilingual_PP-OCRv3_det_distill_train.tar) | [इन्फरन्समॉडल](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/devanagari_PP-OCRv3_rec_infer.tar) / [प्रशिक्षित मॉडल](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/devanagari_PP-OCRv3_rec_train.tar) |
| इंग्लिश अल्ट्रा- लाइट वेट PP-ओसीआरv3 मॉडल (13.4M) | en_PP-ओसीआरv3_xx | मोबाइल और सर्वर | [इन्फरन्स मॉडल](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_det_infer.tar) / [प्रशिक्षितमॉडल](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_det_distill_train.tar)| [इन्फरन्समॉडल](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_rec_infer.tar) / [प्रशिक्षित मॉडल](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_rec_train.tar) |
| चाइनीस और इंग्लिश अल्ट्रा- लाइट वेट PP-ओसीआरv3 मॉडल(16.2M) | ch_PP-ओसीआरv3_xx | मोबाइल और सर्वर | [इन्फरन्स मॉडल](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_infer.tar) / [प्रशिक्षित मॉडल](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_distill_train.tar) | [प्रशिक्षित मॉडल](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_rec_infer.tar) / [प्रशिक्षित मॉडल](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_rec_train.tar) |
- अधिक मॉडल डाउनलोड (एकाधिक भाषाओं सहित) के लिए, कृपया [PP-ओसीआर सीरीज मॉडल डाउनलोड](../doc_en/models_list_en.md) देखें।
- एक नए भाषा अनुरोध के लिए, कृपया [नई भाषा अनुरोधों के लिए दिशानिर्देश](#language_requests).
- स्ट्रक्चर मॉडल डोकोमेंट एनालाइज के लिए, कृपया देखें [PP-Structure models](../../ppstructure/docs/models_list_en.md).
<a name="ट्यूटोरियल"></a>
## 📖 ट्यूटोरियल
- [एनवायरनमेंट प्रिपरेशन](../doc_en/environment_en.md)
- [PP-OCR 🔥](../doc_en/ppocr_introduction_en.md)
- [क्विक स्टार्ट](../doc_en/quickstart_en.md)
- [मॉडल जू](../doc_en/models_en.md)
- [मॉडल ट्रेनिंग](../doc_en/training_en.md)
- [टेक्सट डिटेक्शन](../doc_en/detection_en.md)
- [टेक्सट रिकोगनीशन](../doc_en/recognition_en.md)
- [टेक्सट डायरेक्शन क्लासिफिकेशन](../doc_en/angle_class_en.md)
- मॉडल कम्प्रेशन
- [मॉडल परिमाणीकरण](./deploy/slim/quantization/README_en.md)
- [मॉडल प्रूनिंग](./deploy/slim/prune/README_en.md)
- [ज्ञान आसवन](../doc_en/knowledge_distillation_en.md)
- [इन्फरन्स और डिप्लोमेन्ट](./deploy/README.md)
- [Python इन्फरन्स](../doc_en/inference_ppocr_en.md)
- [C++ इन्फरन्स](./deploy/cpp_infer/readme.md)
- [सरविंग](./deploy/pdserving/README.md)
- [मोबाइल](./deploy/lite/readme.md)
- [Paddle2ONNX](./deploy/paddle2onnx/readme.md)
- [पैडल क्लाउड](./deploy/paddlecloud/README.md)
- [Benchmark](../doc_en/benchmark_en.md)
- [PP-Structure 🔥](../../ppstructure/README.md)
- [क्विक स्टार्ट](../../ppstructure/docs/quickstart_en.md)
- [मॉडल जू](../../ppstructure/docs/models_list_en.md)
- [मॉडल ट्रेनिंग](../doc_en/training_en.md)
- [लेआउट एनालाइस](../../ppstructure/layout/README.md)
- [टेबल रिकोगनाइजेशन](../../ppstructure/table/README.md)
- [की इंफॉर्मेशन एक्स्ट्रेक्शन](../../ppstructure/kie/README.md)
- [इन्फरन्स और डिप्लोमेन्ट](./deploy/README.md)
- [Python इन्फरन्स](../../ppstructure/docs/inference_en.md)
- [C++ इन्फरन्स](./deploy/cpp_infer/readme.md)
- [सरविंग](./deploy/hubserving/readme_en.md)
- [एकेडमिक एल्गोरिथम](../doc_en/algorithm_overview_en.md)
- [टेक्स्ट डिनेक्शन](../doc_en/algorithm_overview_en.md)
- [टेक्स्ट रिकोगनाइजेशन](../doc_en/algorithm_overview_en.md)
- [एंड-टू-एंड ओसीआर](../doc_en/algorithm_overview_en.md)
- [टेबल रिकोगनाइजेशन](../doc_en/algorithm_overview_en.md)
- [की इंफॉर्मेशन एक्स्ट्रेक्शन](../doc_en/algorithm_overview_en.md)
- [पैडलओसीआर में नए एल्गोरिदम जोड़ें](../doc_en/add_new_algorithm_en.md)
- डेटा एनोटेशन और सिंथेसिस
- [सेमी-ऑटोमैटिक एनोटेशन टूल: PPओसीआरलेबल](./PPOCRLabel/README.md)
- [डेटा सिंथेसिस टूल: स्टाइल-टेक्सट](./StyleText/README.md)
- [अन्य डेटा एनोटेशन टूल](../doc_en/data_annotation_en.md)
- [अन्य डेटा सिंथेसिस टूल](../doc_en/data_synthesis_en.md)
- डेटा सेट
- [सामान्य ओसीआर डेटासेट (चीनी/अंग्रेज़ी)](../doc_en/dataset/datasets_en.md)
- [हस्तलिखित_ओसीआर_डेटासेट (चीनी)](../doc_en/dataset/handwritten_datasets_en.md)
- [विभिन्न ओसीआर
डेटासेट (बहुभाषी)](../doc_en/dataset/vertical_and_multilingual_datasets_en.md)
- [लेआउट एनालाइस](../doc_en/dataset/layout_datasets_en.md)
- [टेबल रिकोगनाइजेशन](../doc_en/dataset/table_datasets_en.md)
- [की इंफॉर्मेशन एक्स्ट्रेक्शन](../doc_en/dataset/kie_datasets_en.md)
- [कोड संरचना](../doc_en/tree_en.md)
- [विसुमलाइजेशन](#Visualization)
- [कम्युनिटी](#Community)
- [नई भाषा के लिए अनुरोध](#language_requests)
- [सामान्य प्रश्न](../doc_en/FAQ_en.md)
- [रेफरेन्सेस](../doc_en/reference_en.md)
- [लाइसेंस](#LICENSE)
<a name="language_requests"></a>
## 🇺🇳 नई भाषा अनुरोधों के लिए संयुक्त राष्ट्र दिशानिर्देश
अगर आप **एक नए भाषा मॉडल का अनुरोध करना चाहते हैं**, तो कृपया [बहुभाषी मॉडल अपग्रेड के लिए वोट करें](https://github.com/PaddlePaddle/PaddleOCR/discussions/7253) में वोट करें। हम नियमित रूप से परिणाम के अनुसार मॉडल को अपग्रेड करेंगे। **अपने दोस्तों को एक साथ वोट करने के लिए आमंत्रित करें!**
यदि आपको **एक नए भाषा मॉडल को प्रशिक्षित करने** अपने परिदृश्य के आधार पर, तो यह [बहुभाषी मॉडल ट्रेनिंग प्रोजेक्ट ट्रेनिंग](https://github.com/PaddlePaddle/PaddleOCR/discussions/7252) ट्यूटोरियल आपको डेटासेट तैयार करने में मदद करेगा और आपको स्टेप बाए स्टेप पूरा प्रोसेस दिखाएगा
मूल [बहुभाषी ओसीआर विकास योजना](https://github.com/PaddlePaddle/PaddleOCR/issues/1048) अभी भी आपको बहुत सारे उपयोगी संग्रह और शब्दकोश दिखाता है
<a name="विसुमलाइजेशन"></a>
## 👀 विज़ुअलाइज़ेशन [अधिक](../doc_en/visualization_en.md)
<details open>
<summary>PP-OCRv3 बहुभाषी मॉडल</summary>
<div align="center">
<img src="../imgs_results/PP-OCRv3/multi_lang/japan_2.jpg" width="800">
<img src="../imgs_results/PP-OCRv3/multi_lang/korean_1.jpg" width="800">
</div>
</details>
<details open>
<summary>PP-OCRv3 अंग्रेजी मॉडल</summary>
<div align="center">
<img src="../imgs_results/PP-OCRv3/en/en_1.png" width="800">
<img src="../imgs_results/PP-OCRv3/en/en_2.png" width="800">
</div>
</details>
<details open>
<summary>PP-OCRv3 चीनी मॉडल</summary>
<div align="center">
<img src="../imgs_results/PP-OCRv3/ch/PP-OCRv3-pic001.jpg" width="800">
<img src="../imgs_results/PP-OCRv3/ch/PP-OCRv3-pic002.jpg" width="800">
<img src="../imgs_results/PP-OCRv3/ch/PP-OCRv3-pic003.jpg" width="800">
</div>
</details>
<details open>
<summary>PP-Structurev2</summary>
1. लेआउट एनालाइस + टेबल रिकोगनाइजेशन
<div align="center">
<img src="../../ppstructure/docs/table/ppstructure.GIF" width="800">
</div>
2. SER (सिमेंटिक एंटिटी रिकोगनाइजेशन)
<div align="center">
<img src="https://user-images.githubusercontent.com/25809855/186094456-01a1dd11-1433-4437-9ab2-6480ac94ec0a.png" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185310636-6ce02f7c-790d-479f-b163-ea97a5a04808.jpg" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185539517-ccf2372a-f026-4a7c-ad28-c741c770f60a.png" width="600">
</div>
3. RE (रिलेशन एक्सट्रैक्शन)
<div align="center">
<img src="https://user-images.githubusercontent.com/25809855/186094813-3a8e16cc-42e5-4982-b9f4-0134dfb5688d.png" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185393805-c67ff571-cf7e-4217-a4b0-8b396c4f22bb.jpg" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185540080-0431e006-9235-4b6d-b63d-0b3c6e1de48f.jpg" width="600">
</div>
</details>
<a name="लाइसेंस"></a>
## 📄 लाइसेंस
इस प्रोजेक्ट को इन परियोजना के तहत जारी किया गया है <a href="https://github.com/PaddlePaddle/PaddleOCR/blob/master/LICENSE">Apache 2.0 license</a>
[English](../../README.md) | [简体中文](../../README_ch.md) | [हिन्दी](./README_हिन्द.md) | 日本語 | [한국인](./README_한국어.md) | [Pу́сский язы́к](./README_Ру́сский_язы́к.md)
<p align="center">
<img src="../PaddleOCR_log.png" align="middle" width = "600"/>
<p align="center">
<p align="left">
<a href="./LICENSE"><img src="https://img.shields.io/badge/license-Apache%202-dfd.svg"></a>
<a href="https://github.com/PaddlePaddle/PaddleOCR/releases"><img src="https://img.shields.io/github/v/release/PaddlePaddle/PaddleOCR?color=ffa"></a>
<a href=""><img src="https://img.shields.io/badge/python-3.7+-aff.svg"></a>
<a href=""><img src="https://img.shields.io/badge/os-linux%2C%20win%2C%20mac-pink.svg"></a>
<a href=""><img src="https://img.shields.io/pypi/format/PaddleOCR?color=c77"></a>
<a href="https://pypi.org/project/PaddleOCR/"><img src="https://img.shields.io/pypi/dm/PaddleOCR?color=9cf"></a>
<a href="https://github.com/PaddlePaddle/PaddleOCR/stargazers"><img src="https://img.shields.io/github/stars/PaddlePaddle/PaddleOCR?color=ccf"></a>
</p>
## 紹介
PaddleOCR は、さまざまな言語で、優れた最先端かつ実用的な OCR ツールを作成することを目的とし、ユーザーがより優れたモデルをトレーニングし、実践的に対応できるようになるために役立つAIOCRです。
<div align="center">
<img src="https://user-images.githubusercontent.com/50011306/187821591-6cb09459-fdbf-4ad3-8c5a-26af611c211d.png" width="800">
</div>
<div align="center">
<img src="../imgs_results/PP-OCRv3/en/en_4.png" width="800">
</div>
<div align="center">
<img src="../imgs_results/ch_ppocr_mobile_v2.0/00006737.jpg" width="800">
</div>
## 📣 最新アップデート
- **🔥2022.8.24 リリース PaddleOCR [release/2.6 ](https://github.com/PaddlePaddle/PaddleOCR/tree/release/2.6)**
- [PP-Structurev2](../../ppstructure/)がリリース。機能と使いやすさがアップグレード、中国語のさまざまな文字に適応、 [レイアウトの復旧](../../ppstructure/recovery) さらに**1 行のコマンドをPDFへ転換、そして Word**に変換可能。
- [レイアウト分析](../../ppstructure/layout) の最適化:モデルのストレージが 95% 削減、速度が 11 倍向上、平均 CPU 時間コストはわずか 41 ミリ秒です。
- [表認識](../../ppstructure/table) 最適化:3つの最適化戦略設計、モデルの精度が従来より同時間比が 6% 向上。
- [キー情報抽出](../../ppstructure/kie) 最適化:視覚に依存しないモデル構造設計、語彙の実態識別精度が 2.8% 向上、関係抽出の精度が 9.1% 向上。
- **🔥2022.7 リリース [OCR scene application collection](../../applications/README_en.md)**
- デジタルチューブ、液晶画面、ナンバー プレート、手書き認識モデル、高精度 SVTR モデルなど、**9つの垂直モデル**をリリース、一般、製造、金融、運輸業界の主要な OCR 垂直アプリケーションをカバー。
- **🔥2022.5.9 リリース PaddleOCR [release/2.5](https://github.com/PaddlePaddle/PaddleOCR/tree/release/2.5)**
- [PP-OCRv3](../doc_en/ppocr_introduction_en.md#pp-ocrv3)リリース: 同等の速度で、中国語の識別効果は PP-OCRv2 より 5% 向上、英語の識別効果は 11% 向上し、80 言語の多言語モデルの平均認識精度は 5% 以上向上。
- [PPOCRLabelv2](./PPOCRLabel)リリース: 表認識タスク、キー情報抽出タスク、イレギュラーテキスト画像のアノテーション機能を追加。
- インタラクティブな電子書籍 [*"OCR に没入"*](../doc_en/ocr_book_en.md)、 をリリース。 OCRフルスタック技術の最先端の理論とコードの実践をカバー。
- [もっと](../doc_en/update_en.md)
## 🌟 PaddleOCRとは?
PaddleOCRは、OCRに関連するさまざまな最先端のアルゴリズムに対応する、産業用の機能モデル/ソリューション [PP-OCR](../doc_en/ppocr_introduction_en.md)[PP-Structure](../../ppstructure/README.md) を開発。これに基づき、データの生成、モデルのトレーニング、圧縮、推論、展開の全プロセスを実行可能。
<div align="center">
<img src="https://user-images.githubusercontent.com/50011306/195771471-fad5eb1d-190d-4a7b-8b0c-0433fb32445f.png">
</div>
## ⚡ 今すぐトライアル
```bash
pip3 install paddlepaddle # for gpu user please install paddlepaddle-gpu
pip3 install paddleocr
paddleocr --image_dir /your/test/image.jpg --lang=japan # change for i18n abbr
```
>Python環境がない場合は [環境の準備](../doc_en/environment_en.md)に従ってください。[チュートリアル](#Tutorials) から始めることをお勧めします。
<a name="本"></a>
## 📚 電子書籍:*OCRに入る*
- [OCRに没入](../doc_en/ocr_book_en.md)
<a name="コミュニティ"></a>
## 👫コミュニティー
他国の開発者の方は [PaddleOCR Discussions](https://github.com/PaddlePaddle/PaddleOCR/discussions) を国際的なコミュニティ プラットフォームとして使用します。みなさんのアイデアや質問がある場合、ここで英語で話し合うことができます。
<a name="対応中国機種一覧"></a>
## 🛠️ シリーズ モデル式一覧
| モデル紹介 | モデル名 | 推奨のシーン | 検出モデル | 認識モデル |
| ------------------------------------------------------------ | ---------------------------- | ----------------- | ------------------------------------------------------------ | ------------------------------------------------------------ |
| 日本語超軽量 PP-OCRv3 モデル(14.8M) | japan_PP-OCRv3_xx | モバイル & サーバー |[推論モデル](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/Multilingual_PP-OCRv3_det_infer.tar)/[トレーニングモデル](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/Multilingual_PP-OCRv3_det_distill_train.tar) |[推論モデル](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/japan_PP-OCRv3_rec_infer.tar)/[トレーニングモデル](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/japan_PP-OCRv3_rec_train.tar) |
| 英語超軽量PP-OCRv3モデル(13.4M) | en_PP-OCRv3_xx | モバイル & サーバー | [推論モデル](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_det_infer.tar) / [トレーニングモデル](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_det_distill_train.tar) | [推論モデル](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_rec_infer.tar) / [トレーニングモデル](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_rec_train.tar) |
| 中国語と英語の超軽量 PP-OCRv3 モデル(16.2M) | ch_PP-OCRv3_xx | モバイル & サーバー | [推論モデル](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_infer.tar) / [トレーニングモデル](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_distill_train.tar) | [推論モデル](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_rec_infer.tar) / [トレーニングモデル](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_rec_train.tar) |
- その他のモデルのダウンロード (多言語を含む) については、[PP-OCR シリーズ モデルのダウンロード] (../doc_en/models_list_en.md)をご参照ください。
- 新しい言語のリクエストについては、 [新しい言語_リクエストのガイドライン](#language_requests)を参照してください。
- 構造文書分析モデルについては、[PP-Structure models](../../ppstructure/docs/models_list_en.md)をご参照ください。
<a name="チュートリアル"></a>
## 📖 チュートリアル
- [環境の準備](../doc_en/environment_en.md)
- [PP-OCR 🔥](../doc_en/ppocr_introduction_en.md)
- [クイックスタート](../doc_en/quickstart_en.md)
- [Model Zoo](../doc_en/models_en.md)
- [トレーニング モデル](../doc_en/training_en.md)
- [テキスト検出](../doc_en/detection_en.md)
- [テキスト認識](../doc_en/recognition_en.md)
- [テキスト方向の分類](../doc_en/angle_class_en.md)
- モデル圧縮
- [モデルの量子化](./deploy/slim/quantization/README_en.md)
- [モデルの剪裁](./deploy/slim/prune/README_en.md)
- [知識の蒸留](../doc_en/knowledge_distillation_en.md)
- [推論と展開](./deploy/README.md)
- [Python 推論](../doc_en/inference_ppocr_en.md)
- [C++ 推論](./deploy/cpp_infer/readme.md)
- [サービング](./deploy/pdserving/README.md)
- [モバイル](./deploy/lite/readme.md)
- [Paddle2ONNX](./deploy/paddle2onnx/readme.md)
- [PaddleCloud](./deploy/paddlecloud/README.md)
- [Benchmark](../doc_en/benchmark_en.md)
- [PP-Structure 🔥](../../ppstructure/README.md)
- [クイックスタート](../../ppstructure/docs/quickstart_en.md)
- [Model Zoo](../../ppstructure/docs/models_list_en.md)
- [トレーニング モデル](../doc_en/training_en.md)
- [レイアウト分析](../../ppstructure/layout/README.md)
- [表認識](../../ppstructure/table/README.md)
- [キー情報抽出](../../ppstructure/kie/README.md)
- [推論と展開](./deploy/README.md)
- [Python 推論](../../ppstructure/docs/inference_en.md)
- [C++ 推論](./deploy/cpp_infer/readme.md)
- [サービング](./deploy/hubserving/readme_en.md)
- [アカデミックアリゴリズム](../doc_en/algorithm_overview_en.md)
- [テキスト検出](../doc_en/algorithm_overview_en.md)
- [テキスト認識](../doc_en/algorithm_overview_en.md)
- [エンド・ツー・エンド OCR](../doc_en/algorithm_overview_en.md)
- [表認識](../doc_en/algorithm_overview_en.md)
- [キー情報抽出](../doc_en/algorithm_overview_en.md)
- [PaddleOCR に新しいアルゴリズムを追加する](../doc_en/add_new_algorithm_en.md)
- データの注釈と合成
- [半自動注釈ツール: PPOCRLabel](./PPOCRLabel/README.md)
- [データ合成ツール: Style-Text](./StyleText/README.md)
- [その他のデータ注釈ツール](../doc_en/data_annotation_en.md)
- [その他のデータ合成ツール](../doc_en/data_synthesis_en.md)
- データセット
- [一般OCRデータセット(中国語/英語)](../doc_en/dataset/datasets_en.md)
- [HandWritten_OCR_Datasets(中国語)](../doc_en/dataset/handwritten_datasets_en.md)
- [各種OCRデータセット(多言語対応)](../doc_en/dataset/vertical_and_multilingual_datasets_en.md)
- [レイアウト分析](../doc_en/dataset/layout_datasets_en.md)
- [表認識](../doc_en/dataset/table_datasets_en.md)
- [キー情報抽出](../doc_en/dataset/kie_datasets_en.md)
- [コード構造](../doc_en/tree_en.md)
- [視覚化](#Visualization)
- [コミュニティ](#Community)
- [新言語のリクエスト](#language_requests)
- [よくある質問](../doc_en/FAQ_en.md)
- [参考文献](../doc_en/reference_en.md)
- [ライセンス](#LICENSE)
<a name="language_requests"></a>
## 🇺🇳 新しい言語リクエストのガイドライン
**新言語モデルをリクエスト**したい場合、[多言語モデルのアップグレードへの投票](https://github.com/PaddlePaddle/PaddleOCR/discussions/7253)で投票してください。投票結果に応じて定期的にモデルがアップグレードされます。**友達を招待して一緒に投票しましょう!**
シナリオに基づいて**新しい言語モデルをトレーニング** する必要がある場合は、[多言語モデル トレーニング プロジェクト](https://github.com/PaddlePaddle/PaddleOCR/discussions/7252) のチュートリアルがデータセットの準備にご利用でき、 プロセス全体を段階的に表示することができます。
元の[多言語 OCR 開発計画](https://github.com/PaddlePaddle/PaddleOCR/issues/1048) には、まだ多くの有用なコーパスと辞書が表示されています
<a name="ビジュアリゼーション"></a>
## 👀 ビジュアリゼーション [more](../doc_en/visualization_en.md)
<details open>
<summary>PP-OCRv3 多言語モデル</summary>
<div align="center">
<img src="../imgs_results/PP-OCRv3/multi_lang/japan_2.jpg" width="800">
<img src="../imgs_results/PP-OCRv3/multi_lang/korean_1.jpg" width="800">
</div>
</details>
<details open>
<summary>PP-OCRv3 英語 モデル</summary>
<div align="center">
<img src="../imgs_results/PP-OCRv3/en/en_1.png" width="800">
<img src="../imgs_results/PP-OCRv3/en/en_2.png" width="800">
</div>
</details>
<details open>
<summary>PP-OCRv3 中国語 モデル</summary>
<div align="center">
<img src="../imgs_results/PP-OCRv3/ch/PP-OCRv3-pic001.jpg" width="800">
<img src="../imgs_results/PP-OCRv3/ch/PP-OCRv3-pic002.jpg" width="800">
<img src="../imgs_results/PP-OCRv3/ch/PP-OCRv3-pic003.jpg" width="800">
</div>
</details>
<details open>
<summary>PP-Structurev2</summary>
1. レイアウト分析+テーブル認識
<div align="center">
<img src="../../ppstructure/docs/table/ppstructure.GIF" width="800">
</div>
2. SER (セマンティックエンティティ認識)
<div align="center">
<img src="https://user-images.githubusercontent.com/25809855/186094456-01a1dd11-1433-4437-9ab2-6480ac94ec0a.png" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185310636-6ce02f7c-790d-479f-b163-ea97a5a04808.jpg" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185539517-ccf2372a-f026-4a7c-ad28-c741c770f60a.png" width="600">
</div>
3. RE (関係抽出)
<div align="center">
<img src="https://user-images.githubusercontent.com/25809855/186094813-3a8e16cc-42e5-4982-b9f4-0134dfb5688d.png" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185393805-c67ff571-cf7e-4217-a4b0-8b396c4f22bb.jpg" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185540080-0431e006-9235-4b6d-b63d-0b3c6e1de48f.jpg" width="600">
</div>
</details>
<a name="ライセンス"></a>
## 📄 ライセンス
このプロジェクトは以下の場所でリリースされています <a href="https://github.com/PaddlePaddle/PaddleOCR/blob/master/LICENSE">Apache 2.0 license</a>
[English](../../README.md) | [简体中文](../../README_ch.md) | [हिन्दी](./README_हिन्द.md) | [日本語](./README_日本語.md) | 한국인 | [Pу́сский язы́к](./README_Ру́сский_язы́к.md)
<p align="center">
<img src="../PaddleOCR_log.png" align="middle" width = "600"/>
<p align="center">
<p align="left">
<a href="./LICENSE"><img src="https://img.shields.io/badge/license-Apache%202-dfd.svg"></a>
<a href="https://github.com/PaddlePaddle/PaddleOCR/releases"><img src="https://img.shields.io/github/v/release/PaddlePaddle/PaddleOCR?color=ffa"></a>
<a href=""><img src="https://img.shields.io/badge/python-3.7+-aff.svg"></a>
<a href=""><img src="https://img.shields.io/badge/os-linux%2C%20win%2C%20mac-pink.svg"></a>
<a href=""><img src="https://img.shields.io/pypi/format/PaddleOCR?color=c77"></a>
<a href="https://pypi.org/project/PaddleOCR/"><img src="https://img.shields.io/pypi/dm/PaddleOCR?color=9cf"></a>
<a href="https://github.com/PaddlePaddle/PaddleOCR/stargazers"><img src="https://img.shields.io/github/stars/PaddlePaddle/PaddleOCR?color=ccf"></a>
</p>
## 소개
PaddleOCR은 사용자들이 보다 나은 모델을 훈련하여 실전에 투입하는데 도움을 주는 다중 언어로 된 엄청나게 멋지고 주도적이며 실용적인 OCR 툴을 만드는데 목표를 두고 있습니다.
<div align="center">
<img src="https://user-images.githubusercontent.com/50011306/187821591-6cb09459-fdbf-4ad3-8c5a-26af611c211d.png" width="800">
</div>
<div align="center">
<img src="../imgs_results/PP-OCRv3/en/en_4.png" width="800">
</div>
<div align="center">
<img src="../imgs_results/ch_ppocr_mobile_v2.0/00006737.jpg" width="800">
</div>
## 📣최근 업데이트
- **🔥2022년 8월 24일에 패들 OCR 출시 [출시/2.6](https://github.com/PaddlePaddle/PaddleOCR/tree/release/2.6)**
차이니즈 씬에 맞춘 완전 업그레이드 된 기능과 성능을 갖춘 ; [PP-Structurev2](../../ppstructure/) 출시, 그리고 레이아웃 리커버리 ](../../ppstructure/recovery) 신규 지원 및 PDF 를 워드로 전환하는 원 라인 명령
- [레이아웃 분석](../../ppstructure/layout) 최적화: 95% 감소된 모델 저장, while 반면 속도는 11배 증가하고, 평균 CPU 시간 비용은 41ms에 불과함;
- [표 인식](../../ppstructure/table) 최적화: 3 최적화 전략이 디자인되고 모델 정확도는 비교 가능한 시간 소비 하에 6% 개선됨;
- [핵심 정보 추출](../../ppstructure/kie) 최적화: 시각에 의존하지 않는 모델 구조가 디자인되고, 의미체 인식 정확도가 2.8% 증가되며 관계 추출 정확도는 9.1% 증가됨.
- **🔥2022년 7월 출시[OCR 씬 애플리케이션 컬렉션](../../applications/README_en.md)**
디지털 튜브, LCD 스크린, 라이선스 플레이트, 수기 인식 모델, 고정밀 SVTR 모델 등등과 같은 “9수직 모델” 출시로, 일반적으로 주된 OCR 수직 애플리케이션, 제조, 금융 및 수송 산업 커버
- **🔥2022년 5월 9일에 패들 OCR 출시 [출시/2.5](https://github.com/PaddlePaddle/PaddleOCR/tree/release/2.5)**
- [PP-OCRv3](../doc_en/ppocr_introduction_en.md#pp-ocrv3)출시: 5%.비교 가능한 속도로, 차이니즈 씬의 효과는 PP-OCRv2와 비교해 볼 때 추가로 5% 정도 더 개선되고, 잉글리쉬 씬 효과는 11% 개선되었으며, 80개 언어 다중 언어 모델 평균 인식 정확도는 5% 이상 개선됨.
- [PPOCRLabelv2](./PPOCRLabel)출시: 표 인식 업무, 핵심 정보 추출 업무 및 불규칙한 텍스트 이미지주석 기능 추가.
- 쌍방향e-북 출시 [*"OCR 뛰어들기"*](../doc_en/ocr_book_en.md), 첨단 이론 및 OCR 정식 스택 기술 코드 연습 포함.
- [추가](../doc_en/update_en.md)
## 🌟특징
패들OCR은 OCR 관련 다양한 첨단 알고리즘 지원 [PP-OCR](../doc_en/ppocr_introduction_en.md)[PP-Structure](../../ppstructure/README.md) 이를 기반으로, 그리고 전체 데이터 생산 처리, 모델 훈련, 압축, 추론 및 배치를 통해 획득.
<div align="center">
<img src="https://user-images.githubusercontent.com/50011306/196963392-6cd1b251-109b-49c3-9b3d-ccf203dcec49.png">
</div>
## ⚡ 신속한 경험
```bash
pip3 install paddlepaddle # for gpu user please install paddlepaddle-gpu
pip3 install paddleocr
paddleocr --image_dir /your/test/image.jpg --lang=korean
```
>만일 당신이 파이톤 환경이 없다면 [환경 준비]를 따르기 바람(../doc_en/environment_en.md). 우리는 당신이[사용지침 프로그램]으로 시작할 것을 권장합니다.(#Tutorials).
<a name="북"></a>
## 📚 E-북: *OCR로 뛰어들기*
- [OCR로 뛰어들기](../doc_en/ocr_book_en.md)
<a name="커뮤니티"></a>
## 👫 커뮤니티로
국제 개발자들을 위해 우리는 [PaddleOCR 논의하기](https://github.com/PaddlePaddle/PaddleOCR/discussions) 를 우리의 국제 커뮤니티로 간주. 모든 아이디어와 질문은 여기서 영어로 논의 가능.
<a name="지원됨 – 차이니즈-모델-목록- "></a>
## PP-OCR 시리즈 모델 목록
| 모델 소개 |모델 명 | 권장 씬| 감지 모델 |인식 모델 |
| ------------------------------------------------------------ | ---------------------------- | ----------------- | ------------------------------------------------------------ | ------------------------------------------------------------ |
| 한국어 초경량 PP-OCRv3 모델(14.8M) | korean_PP-OCRv3_xx | 모바일 & 서버 | [추론 모델](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/Multilingual_PP-OCRv3_det_infer.tar) / [훈련 모델](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/Multilingual_PP-OCRv3_det_distill_train.tar) | [추론 모델](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/korean_PP-OCRv3_rec_infer.tar) / [훈련 모델](https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/korean_PP-OCRv3_rec_train.tar) |
| 영어 초경량 PP-OCRv3 모델(13.4M) | en_PP-OCRv3_xx | 모바일 & 서버 | [추론 모델](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_det_infer.tar) / [훈련 모델](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_det_distill_train.tar) | [추론 모델](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_rec_infer.tar) / [훈련 모델](https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_rec_train.tar) |
| 중국어 및 영어 초경량 PP-OCRv3 model(16.2M) | ch_PP-OCRv3_xx | 모바일 & 서버 | [추론 모델](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_infer.tar) / [훈련 모델](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_distill_train.tar) | [추론 모델](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_rec_infer.tar) / [훈련 모델](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_rec_train.tar) |
- (다중 언어를 포함하여)더 많은 모델을 다운로드 하려면, [PP-OCR 시리즈 모델 다운로드](../doc_en/models_list_en.md)를 참조할 것.
- 신규 언어 요청에 대해서는, [신규 언어 요청 지침](#language_requests)을 참조할 것.
- 구조적 문서 분석 모델에 대해서는, [PP-Structure models](../../ppstructure/docs/models_list_en.md).을 참조할 것.
<a name="사용 지침 프로그램"></a>
## 📖 사용 지침 프로그램
- [환경 준비](../doc_en/environment_en.md)
- [PP-OCR 🔥](../doc_en/ppocr_introduction_en.md)
- [신속한 시작](../doc_en/quickstart_en.md)
- [동물원 모델](../doc_en/models_en.md)
- [모델 훈련](../doc_en/training_en.md)
- [텍스트 감지](../doc_en/detection_en.md)
- [텍스트 인식](../doc_en/recognition_en.md)
- [텍스트 방향 분류](../doc_en/angle_class_en.md)
- 모델 압축
- [모델 계량화](./deploy/slim/quantization/README_en.md)
- [모델 전지작업](./deploy/slim/prune/README_en.md)
- [지식 정제](../doc_en/knowledge_distillation_en.md)
- [추론 및 배치](./deploy/README.md)
- [파이톤 추론](../doc_en/inference_ppocr_en.md)
- [C++ 추론](./deploy/cpp_infer/readme.md)
- [서빙](./deploy/pdserving/README.md)
- [모바일](./deploy/lite/readme.md)
- [Paddle2ONNX](./deploy/paddle2onnx/readme.md)
- [패들 클라우드](./deploy/paddlecloud/README.md)
- [Benchmark](../doc_en/benchmark_en.md)
- [PP-Structure 🔥](../../ppstructure/README.md)
- [신속한 시작](../../ppstructure/docs/quickstart_en.md)
- [동물원 모델](../../ppstructure/docs/models_list_en.md)
- [모델 훈련](../doc_en/training_en.md)
- [레이아웃 분석](../../ppstructure/layout/README.md)
- [표 인식](../../ppstructure/table/README.md)
- [핵심 정보 추출](../../ppstructure/kie/README.md)
- [추론 및 배치](./deploy/README.md)
- [파이톤 추론](../../ppstructure/docs/inference_en.md)
- [C++ 추론](./deploy/cpp_infer/readme.md)
- [서빙](./deploy/hubserving/readme_en.md)
- [학문적 알고리즘](../doc_en/algorithm_overview_en.md)
- [텍스트 감지](../doc_en/algorithm_overview_en.md)
- [텍스트 인식](../doc_en/algorithm_overview_en.md)
- [종단종OCR](../doc_en/algorithm_overview_en.md)
- [표 인식](../doc_en/algorithm_overview_en.md)
- [핵심 정보 추출](../doc_en/algorithm_overview_en.md)
- [PaddleOCR에 신규 알고리즘 추가](../doc_en/add_new_algorithm_en.md)
- 데이터 주석 및 합성
- [반-자동 주석 툴: PPOCRLabel](./PPOCRLabel/README.md)
- [데이터 합성 툴: 스타일-텍스트](./StyleText/README.md)
- [기타 데이터 주석 툴](../doc_en/data_annotation_en.md)
- [기타 데이터 합성 툴](../doc_en/data_synthesis_en.md)
- 데이터세트
- [일반 OCR 데이터세트(중국어/영어)](../doc_en/dataset/datasets_en.md)
- [수기_OCR_데이터세트(중국어)](../doc_en/dataset/handwritten_datasets_en.md)
- [다양한 OCR 데이터세트(다중언어)](../doc_en/dataset/vertical_and_multilingual_datasets_en.md)
- [레이아웃 분석](../doc_en/dataset/layout_datasets_en.md)
- [표 인식](../doc_en/dataset/table_datasets_en.md)
- [핵심 정보 추출](../doc_en/dataset/kie_datasets_en.md)
- [코드 구조](../doc_en/tree_en.md)
- [시각화](#Visualization)
- [커뮤니티](#Community)
- [신규 언어 요청](#language_requests)
- [자주 묻는 질문](../doc_en/FAQ_en.md)
- [추론](../doc_en/reference_en.md)
- [라이선스](#LICENSE)
<a name="language_requests"></a>
## 신규 언어 요청에 대한 유엔 가이드라인
만일 신규 언어 모델을 요청하고자 한다면**, [다중 언어 모델 업그레이드 투표하기](https://github.com/PaddlePaddle/PaddleOCR/discussions/7253)에서 투표하기 바람. 우리는 결과에 따라 규칙적으로 모델을 업그레이드 시킬 것임**함께 투표하고자 당신의 친구들을 초대할 것!**
만일 당신이 시나리오 기반 “신규 언어 모델”을 훈련하고자 한다면, [다중 언어 모델 훈련 프로젝트](https://github.com/PaddlePaddle/PaddleOCR/discussions/7252) 를 통해 당신의 데이터세트를 작성하는데 도움이 되고 단계별로 전체 절차를 보여줄 것입니다.
원본[다중 언어 OCR 개발 계획](https://github.com/PaddlePaddle/PaddleOCR/issues/1048)은 여전히 수많은 유용한 말뭉치와 사전을 보여줍니다.
<a name="시각화"></a>
## 👀 시각화[추가](../doc_en/visualization_en.md)
<details open>
<summary>PP-OCRv3 다중 언어 모델</summary>
<div align="center">
<img src="../imgs_results/PP-OCRv3/multi_lang/japan_2.jpg" width="800">
<img src="../imgs_results/PP-OCRv3/multi_lang/korean_1.jpg" width="800">
</div>
</details>
<details open>
<summary>PP-OCRv3 영어 모델</summary>
<div align="center">
<img src="../imgs_results/PP-OCRv3/en/en_1.png" width="800">
<img src="../imgs_results/PP-OCRv3/en/en_2.png" width="800">
</div>
</details>
<details open>
<summary>PP-OCRv3 중국어 모델</summary>
<div align="center">
<img src="../imgs_results/PP-OCRv3/ch/PP-OCRv3-pic001.jpg" width="800">
<img src="../imgs_results/PP-OCRv3/ch/PP-OCRv3-pic002.jpg" width="800">
<img src="../imgs_results/PP-OCRv3/ch/PP-OCRv3-pic003.jpg" width="800">
</div>
</details>
<details open>
<summary>PP-Structurev2</summary>
1. 레이아웃 분석 + 표 인식
<div align="center">
<img src="./ppstructure/docs/table/ppstructure.GIF" width="800">
</div>
2. SER (의미체 인식)
<div align="center">
<img src="https://user-images.githubusercontent.com/25809855/186094456-01a1dd11-1433-4437-9ab2-6480ac94ec0a.png" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185310636-6ce02f7c-790d-479f-b163-ea97a5a04808.jpg" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185539517-ccf2372a-f026-4a7c-ad28-c741c770f60a.png" width="600">
</div>
3. RE (관계 추출)
<div align="center">
<img src="https://user-images.githubusercontent.com/25809855/186094813-3a8e16cc-42e5-4982-b9f4-0134dfb5688d.png" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185393805-c67ff571-cf7e-4217-a4b0-8b396c4f22bb.jpg" width="600">
</div>
<div align="center">
<img src="https://user-images.githubusercontent.com/14270174/185540080-0431e006-9235-4b6d-b63d-0b3c6e1de48f.jpg" width="600">
</div>
</details>
<a name="라이선스"></a>
## 📄 라이선스
본 프로젝트는 <a href="https://github.com/PaddlePaddle/PaddleOCR/blob/master/LICENSE">Apache 2.0 license</a> 하에 출시됨.
......@@ -104,6 +104,19 @@ paddleocr --image_dir=ppstructure/recovery/UnrealText.pdf --type=structure --rec
通过OCR技术:
版面恢复分为2种方法,详细介绍请参考:[版面恢复教程](https://github.com/PaddlePaddle/PaddleOCR/blob/dygraph/ppstructure/recovery/README_ch.md)
- PDF解析
- OCR技术
通过PDF解析(只支持pdf格式的输入):
```bash
paddleocr --image_dir=ppstructure/recovery/UnrealText.pdf --type=structure --recovery=true --use_pdf2docx_api=true
```
通过OCR技术:
```bash
# 中文测试图
paddleocr --image_dir=ppstructure/docs/table/1.png --type=structure --recovery=true
......
......@@ -91,9 +91,9 @@ paddleocr --image_dir=ppstructure/docs/table/table.jpg --type=structure --layout
Key information extraction does not currently support use by the whl package. For detailed usage tutorials, please refer to: [inference document](./inference_en.md).
<a name="216"></a>
#### 2.1.6 layout recovery
#### 2.1.6 layout recovery(PDF to Word)
Two layout recovery methods are provided, For detailed usage tutorials, please refer to: [Layout Recovery](../recovery/README.md).
Two layout recovery methods are provided, For detailed usage tutorials, please refer to: [Layout Recovery](https://github.com/PaddlePaddle/PaddleOCR/blob/dygraph/ppstructure/recovery/README.md).
- PDF parse
- OCR
......
......@@ -19,7 +19,7 @@ PDF2Word是PaddleOCR社区开发者 [whjdark](https://github.com/whjdark) 基于
> - 初次安装程序根据不同设备需要等待1-2分钟不等
> - 使用Office与WPS打开的Word结果会出现不同,推荐以Office为准
> - 本程序使用 [QPT](https://github.com/QPT-Family/QPT) 进行应用程序打包,感谢 [GT-ZhangAcer](https://github.com/GT-ZhangAcer) 对打包过程的支持
> - 应用程序不支持盗版Windows系统,若在安装过程中出现报错或缺少依赖,推荐直接使用 `paddleocr` whl包应用PDF2Word功能,详情可查看[链接](https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.6/ppstructure/docs/quickstart.md)
> - 应用程序仅支持正版win10,11系统,不支持盗版Windows系统,若在安装过程中出现报错或缺少依赖,推荐直接使用 `paddleocr` whl包应用PDF2Word功能,详情可查看[链接](https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.6/ppstructure/docs/quickstart.md)
### 脚本启动界面
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册