# 클라우드 OCR 비교 - Amazon Textract/Rekognition vs Google Vision vs Microsoft Cognitive

> Clean Markdown view of GeekNews topic #106. Use the original source for factual precision when an external source URL is present.

## Metadata

- GeekNews HTML: [https://news.hada.io/topic?id=106](https://news.hada.io/topic?id=106)
- GeekNews Markdown: [https://news.hada.io/topic/106.md](https://news.hada.io/topic/106.md)
- Type: news
- Author: [xguru](https://news.hada.io/@xguru)
- Published: 2019-07-19T07:20:07+09:00
- Updated: 2019-07-19T07:20:07+09:00
- Original source: [amplenote.com](https://www.amplenote.com/blog/2019_examples_amazon_textract_rekognition_microsoft_cognitive_services_google_vision)
- Points: 4
- Comments: 1

## Topic Body

손글씨/웹페이지/칠판 3가지로 테스트했는데, 텍스트 추출하는 것만 비교하면 MS의 압승. 가격도 싼편.

## Comments



### Comment 117

- Author: xguru
- Created: 2019-07-19T07:26:36+09:00
- Points: 2

한국의 경우는 특히나 Form 형태(테이블)의 문서가 많아서, 텍스트추출만 비교하기 보다는 테이블 인식쪽 비교 테스트가 필요하긴 할듯.

테이블이 포함된 복합문서 인식에 있어서는 모두들 추가로 서비스를 제공중

Azure Form Recognizer - https://azure.microsoft.com/en-us/services/cognitive-services/form-recognizer/

Google Document Understanding AI - https://cloud.google.com/solutions/document-understanding/

Amazon Textract Form Extraction - https://docs.aws.amazon.com/textract/latest/dg/how-it-works-analyzing.html