OCR: Difference between revisions

From LemonWiki共筆
Jump to navigation Jump to search
mNo edit summary
m (Text replacement - ": Image:Owl icon.jpg " to "{{Tips}} ")
Tags: Mobile edit Mobile web edit
 
(53 intermediate revisions by the same user not shown)
Line 1: Line 1:
OCR (optical character recognition), 光學字元辨識
OCR (optical character recognition), [https://zh.wikipedia.org/wiki/%E5%85%89%E5%AD%A6%E5%AD%97%E7%AC%A6%E8%AF%86%E5%88%AB 光學字元辨識]、圖片轉文字


== scan ==
 
== OCR tools ==
=== 圖片轉換成文字 ===
* {{Gd}} [https://docs.google.com/ Google DOCs]: 上傳文件後，檔案名稱點選右鍵，「選擇開啟工具」 --> 「Google 文件」<ref>[http://docs.google.com/support/bin/answer.py?answer=176692&hl=en Uploading and exporting: Uploading image files with text to Google Docs]、[https://support.google.com/drive/answer/176692?hl=zh-Hant&visit_id=1-636534874969716350-2978233269&rd=1 將 PDF 和相片檔案轉換為文字 - 電腦 - Google 雲端硬碟說明]</ref> 英文可以順利辨識、簡體中文遇到問題。
** 教學: [https://buzzorange.com/techorange/2019/12/09/convert-picture-into-word/ 不要浪費時間 key 資料啦！拍照上傳 Google 雲端，按個右鍵就自動幫你轉文字 | TechOrange]
 
* [https://www.google.com/photos/about/ Google Photos] 將圖片上傳到 Google Photos，再點選「複製圖像中的文字」{{access | date=2022-09-30}}
 
* {{Gd}} [https://line.me/zh-hant/ 免費通話、免費傳訊的應用程式「LINE」]
** [https://mrmad.com.tw/line-ocr 【教學】LINE 透過 OCR 文字辨識功能，直接讓圖片轉成文字技巧 - 瘋先生]
 
* [https://www.onlineocr.net/dashboard Free Online OCR - convert scanned PDF and images to Word, JPEG to Word] 不註冊有前 5 頁的額度、註冊會員有總共 50 頁的免費額度 {{access | date=2018-02-06}}
 
* [https://udn.com/news/story/7088/3326897 Google Keep內建辨識功能 將圖片內容轉文字輸出 | 社群網路 | 數位 | 聯合新聞網] {{access | date=2018-08-27}}
 
* [https://zhtw.109876543210.com/ 免費在線OCR - 在線圖片識別 - 免費OCR軟件 - 免費OCR轉換成Word - 在線文字識別轉換 - 圖片文字識別軟件 - 圖片轉文字] 免費版限制每次可以批次上傳 3 頁、每天轉換 10 頁的 PDF 檔。 {{access | date=2018-10-24}}
 
* MS Office 2003 需額外安裝的Office 工具: Microsoft Office Document Imaging ([http://fun.idv.tw/fun/2008/12/ms_office_ocr.html 你也可以輕鬆做文字辨識(OCR)])
*# (.pdf檔案轉為.mdi) PDF列印到 MS Office 2003 Document Imaging
*# (.mdi檔案轉為word檔) MS Office 2003 Document Imaging(.mdi) -> 使用OCR辨識/傳送文字到Word 
 
* [http://www.microsoft.com/downloads/details.aspx?FamilyID=dd172063-9517-41d8-82af-29c38f7437b6&DisplayLang=zh-tw Microsoft Office Document Imaging 中文簡體OCR辨識引擎]
 
* [http://imageconvert.nsspot.net/?m=img2ocr Optical Character Recognition tool that extracts text from major image format - Online Image, PDF, Latex, OCR Converter] 繁體中文辨識結果不佳。{{access | date=2020-10-29}}
 
* ''$'' [http://www.newocr.com/ Free Online OCR - Convert JPEG, PNG, GIF, BMP, TIFF, PDF, DjVu to Text] 可指定語言。線上網頁轉換，需要逐頁下載轉換後的檔案。使用 API 免費額度 20 頁<ref>[http://www.newocr.com/api/ Free Online OCR - OCR API]</ref>
 
* [https://ocr.space/ Best Free OCR API, Online OCR, Searchable PDF - Fresh 2022 On-Premise OCR Software] 可指定語言
 
* [https://chatgpt.com/ ChatGPT] 顯示錯誤訊息「Tesseract 不支援繁體中文」 (Traditional Chinese language data for Tesseract is not available in this environment) {{access | date=2024-08-23}}
 
* [https://claude.ai/new Claude] 可處理中文圖轉字，但字出錯仍需要人工校稿。
 
* ''$'' [https://cloud.google.com/vision/?hl=zh-tw Vision AI | 透過機器學習技術取得圖片的深入分析結果  |  Cloud Vision API  |  Google Cloud]
 
=== PDF轉換成文字 ===
{{Tips}} 講個秘訣：因為線上服務免費版會限制 PDF 檔案頁數，可使用切割軟體 [[PDF split and merge tools]]
 
* [[Document_converter#PDF.E8.BD.89.E6.8F.9B.E6.88.90.E7.B4.94.E6.96.87.E5.AD.97 | PDF轉換成文字]]
 
== OCR scripts & API ==
 
[https://github.com/ocropus/ocropy ocropus/ocropy: Python-based tools for document analysis and OCR]
* Script Language: Python
* Support Language: < 10. {{exclaim}} 沒有提供中文 model 檔案 {{access | date=2022-04-20}} More on [https://github.com/ocropus-archive/DUP-ocropy/wiki/Models Models · ocropus-archive/DUP-ocropy Wiki]
* License: [https://github.com/ocropus/ocropy/blob/master/LICENSE Apache License 2.0]
 
[https://github.com/tesseract-ocr/tesseract tesseract-ocr/tesseract: Tesseract Open Source OCR Engine (main repository)] {{access | date=2022-06-19}}
* Script Language: C++; Fork on PHP [https://github.com/thiagoalessio/tesseract-ocr-for-php thiagoalessio/tesseract-ocr-for-php: A wrapper to work with Tesseract OCR inside PHP.] <ref>[https://github.com/tesseract-ocr/tesseract/blob/main/doc/tesseract.1.asc#languages-and-scripts LANGUAGES AND SCRIPTS]</ref>，
* Support Language: 100+ contains Traditional Chinese 但是繁體中文辨識結果不佳。 {{access | date=2022-04-20}}. More on [https://tesseract-ocr.github.io/tessdoc/Data-Files-in-different-versions.html Languages/Scripts supported in different versions of Tesseract | tessdoc]
* License: [https://github.com/tesseract-ocr/tesseract/blob/main/LICENSE Apache License 2.0]. PHP Fork: [https://github.com/thiagoalessio/tesseract-ocr-for-php/blob/main/MIT-LICENSE MIT License]
 
{{Acronym| acronym=API| def=應用程式介面（英語：application programming interface）}} of OCR services
 
Azure AI Vision/[https://azure.microsoft.com/zh-tw/services/cognitive-services/computer-vision/ 電腦視覺 | Microsoft Azure]: [https://docs.microsoft.com/zh-tw/azure/cognitive-services/Computer-vision/quickstarts-sdk/client-library?pivots=programming-language-rest-api&tabs=visual-studio 快速入門：光學字元辨識 (OCR) 用戶端程式庫或 REST API - Azure Cognitive Services | Microsoft Docs]
* Support Language: 支援中文<ref>[https://docs.microsoft.com/zh-tw/azure/cognitive-services/computer-vision/language-support#optical-character-recognition-ocr 語言支援 - 電腦視覺 - Azure Cognitive Services | Microsoft Docs]</ref>
 
[https://cloud.google.com/vision Vision AI | 透過機器學習技術取得圖片的深入分析結果  |  Cloud Vision API  |  Google Cloud]
* Support Language: 支援繁體中文 ({{kbd | key=zh-Hant}})<ref>[https://cloud.google.com/vision/docs/languages OCR Language Support  |  Cloud Vision API  |  Google Cloud]</ref>
 
[https://ocr.space/OCRAPI Free OCR API]
* Support Language: 支援繁體中文 ({{kbd | key=cht}})
 
[https://aws.amazon.com/tw/rekognition/?blog-cards.sort-by=item.additionalFields.createdDate&blog-cards.sort-order=desc Amazon Rekognition – 影片與影像 – AWS]: [https://docs.aws.amazon.com/rekognition/latest/dg/text-detection.html?pg=ln&sec=ft Detecting text - Amazon Rekognition]:
* Support Language: {{exclaim}} 不支援中文<ref>[https://aws.amazon.com/tw/about-aws/whats-new/2021/11/amazon-rekognition-text-detection-7-new-languages-accuracy/ Amazon Rekognition text detection supports 7 new languages and improves accuracy] " Amazon Rekognition is designed to detect words in English, Arabic, Russian, German, French, Italian, Portuguese and Spanish."</ref><ref>[https://docs.aws.amazon.com/rekognition/latest/dg/text-detection.html Detecting text - Amazon Rekognition]</ref> {{access | date=2022-04-20}}
 
 
相關頁面
* [[Troubleshooting of Amazon Rekognition]]
* [[Troubleshooting of Azure Cognitive Services API]]
 
== 常用文件的解析度設定 ==
常用用途的解析度設定
常用用途的解析度設定
* 文字辨識 75~150 dpi
* 文字辨識 75~150 dpi
* 圖文交雜 100~150 dpi
* 圖文交雜 100~150 dpi
* 圖檔(螢幕上觀看) 150~250 dpi
* 圖檔(螢幕上觀看) 150~250 dpi {{exclaim}} 個人經驗: 簡報掃描的圖檔，如果是小字 300 dpi 可以辨識，但建議調整到 600 dpi。
* 圖檔(有列印需求) 300 dpi以上
* 圖檔(有列印需求) 300 dpi以上
* 名片 150~200 dpi
* 名片 150~200 dpi
Line 11: Line 82:
出處：PCHome 2005/8
出處：PCHome 2005/8


== OCR Software ==
== Related Pages ==
* MS Office 2003 需額外安裝的Office 工具: Microsoft Office Document Imaging
* [[Document_converter#PDF_%E8%BD%89%E6%8F%9B%E6%88%90%E7%B4%94%E6%96%87%E5%AD%97_(TXT)|Convert PDF to TXT]]
* [http://www.microsoft.com/downloads/details.aspx?FamilyID=dd172063-9517-41d8-82af-29c38f7437b6&DisplayLang=zh-tw Microsoft Office Document Imaging 中文簡體OCR辨識引擎]
 
* [https://docs.google.com/ Google DOCs]: [http://docs.google.com/support/bin/answer.py?answer=176692&hl=en Uploading and exporting: Uploading image files with text to Google Docs] {{access | date=2018-02-06}}
== References ==
* ''$'' [https://www.onlineocr.net/dashboard Free Online OCR - convert scanned PDF and images to Word, JPEG to Word] 不註冊有前 5 頁的額度、註冊會員有總共 50 頁的免費額度 {{access | date=2018-02-06}}
<references />
 
相關文章
* [https://www.ptt.cc/bbs/EZsoft/M.1516336833.A.EAF.html (請問) 有沒有可以大量OCR(圖文轉換)的軟體 - 看板 EZsoft - 批踢踢實業坊] {{access | date=2018-02-15}}


[[Category:Tool]]
[[Category:Tool]]

Latest revision as of 20:15, 8 January 2025

OCR (optical character recognition), 光學字元辨識、圖片轉文字


OCR tools[edit]

圖片轉換成文字[edit]

  • Google Photos 將圖片上傳到 Google Photos，再點選「複製圖像中的文字」[Last visited: 2022-09-30]
  • MS Office 2003 需額外安裝的Office 工具: Microsoft Office Document Imaging (你也可以輕鬆做文字辨識(OCR))
    1. (.pdf檔案轉為.mdi) PDF列印到 MS Office 2003 Document Imaging
    2. (.mdi檔案轉為word檔) MS Office 2003 Document Imaging(.mdi) -> 使用OCR辨識/傳送文字到Word
  • ChatGPT 顯示錯誤訊息「Tesseract 不支援繁體中文」 (Traditional Chinese language data for Tesseract is not available in this environment) [Last visited: 2024-08-23]
  • Claude 可處理中文圖轉字，但字出錯仍需要人工校稿。

PDF轉換成文字[edit]

Owl icon.jpg 講個秘訣：因為線上服務免費版會限制 PDF 檔案頁數，可使用切割軟體 PDF split and merge tools

OCR scripts & API[edit]

ocropus/ocropy: Python-based tools for document analysis and OCR

tesseract-ocr/tesseract: Tesseract Open Source OCR Engine (main repository) [Last visited: 2022-06-19]

API of OCR services

Azure AI Vision/電腦視覺 | Microsoft Azure: 快速入門：光學字元辨識 (OCR) 用戶端程式庫或 REST API - Azure Cognitive Services | Microsoft Docs

  • Support Language: 支援中文[4]

Vision AI | 透過機器學習技術取得圖片的深入分析結果  |  Cloud Vision API  |  Google Cloud

  • Support Language: 支援繁體中文 (zh-Hant)[5]

Free OCR API

  • Support Language: 支援繁體中文 (cht)

Amazon Rekognition – 影片與影像 – AWS: Detecting text - Amazon Rekognition:

  • Support Language: Icon_exclaim.gif 不支援中文[6][7] [Last visited: 2022-04-20]


相關頁面

常用文件的解析度設定[edit]

常用用途的解析度設定

  • 文字辨識 75~150 dpi
  • 圖文交雜 100~150 dpi
  • 圖檔(螢幕上觀看) 150~250 dpi Icon_exclaim.gif 個人經驗: 簡報掃描的圖檔，如果是小字 300 dpi 可以辨識，但建議調整到 600 dpi。
  • 圖檔(有列印需求) 300 dpi以上
  • 名片 150~200 dpi

出處：PCHome 2005/8

Related Pages[edit]

References[edit]

  1. ↑ Uploading and exporting: Uploading image files with text to Google Docs、將 PDF 和相片檔案轉換為文字 - 電腦 - Google 雲端硬碟說明
  2. ↑ Free Online OCR - OCR API
  3. ↑ LANGUAGES AND SCRIPTS
  4. ↑ 語言支援 - 電腦視覺 - Azure Cognitive Services | Microsoft Docs
  5. ↑ OCR Language Support  |  Cloud Vision API  |  Google Cloud
  6. ↑ Amazon Rekognition text detection supports 7 new languages and improves accuracy " Amazon Rekognition is designed to detect words in English, Arabic, Russian, German, French, Italian, Portuguese and Spanish."
  7. ↑ Detecting text - Amazon Rekognition

相關文章

Retrieved from "https://wiki.planetoid.info/index.php?title=OCR&oldid=25324"