Cloud vs. On Device OCR Processing for Mobile Applications

Mobile OCR Engine, Cloud OCR SDK
4.0, 11
Technology & Features, Comparisons

This is a feature and usage scenario comparison between ABBYY's Cloud OCR Service and the Mobile OCR Engine.

Usage Scenarios

Cloud OCR SDK On Device OCR with Mobile OCR Engine
Target Audience (Mobile) Developers with need for OCR as a service;
ISVs developing in the Cloud,
(Mobile) Developers, ISVs,
Hardware Manufacturers
Integration “High level” integration via web service;
REST web-service calls
“Low level” integration via native API or
via provided wrappers
Network Connection Required Not required, on-device “off-line” processing
File Management Upload document images to the “cloud” On device image processing
ProcessingAsynchronous” out of the system
receive the results,
re-OCR if needed
Synchronous” in the system:
getting image,
providing the results
Scalability Cloud based, ABBYY's responsibility On mobile devices
sequential processing within an application instance
Security HTTPs,
Microsoft Azure infrastructure
Custom implementation
Maintenance none
just use the service
Simple customization on embedding into the application.
No maintenance after release


Cloud OCR SDK On Device OCR with Mobile OCR Engine
OCR - Text Recognition yes yes
Barcode Recognition (OBR) yes, with auto-location yes, without auto-location
ICR yes no
Zonal/Field Level OCR yes, API allows to define recognition zones yes, custom fields can be defined and OCRed
OCR Languages Over 200
all available ABBYY recognition languages
without additional costs
(Asian Languages, Arabic, Hebrew, etc)
62 languages including CJK, Cyrillic, Greek
Business Card Reading 27 languages including CJK, Cyrillic, Greek 21 languages including CJK, Cyrillic, Greek
Historic Font OCR yes no
Forms Matching no no
Document Separation no no
GUI Elements no, processing only no, processing only
Export Formats TXT, ABBYY XML, ALTO XML,
XML of Machine Readable Zones (MRZ) of passports and ID cards,
Business Card recognition XML, vCard
Office Formats: Doc(X), ODT, XML(X), PPT(X)
Recognition results are provided as a structure via API:
- plain text
- business cards data sorted by fields

Development & Deployment

Cloud OCR SDK On Device OCR with Mobile OCR Engine
Trial simple online subscription Trial Software License Agreement to be signed before the SDK is provided
Development Tools all development tools that can interact with a web-service native OS tools that can integrate programming libraries
API compact API,
designed for web-service implementation
compact API,
wrappers for iOS and Android
RAM Consumption
Memory Footprint
needed for the code to send the image
and retrieve the results
depending on the operating system and specific recognition tasks,
OCR of Latin and Cyrillic 10 MB RAM, CJK up to 28 MB
Operating Systems no limitation,
network access and REST Calls have to be available
Android, iOS, Symbian, others
Application Size minimal size increase, OCR Engine and dictionaries add up to 20 MB to your application
Code Samples Sample Code online
Included in the SDK

Business Models

Cloud OCR SDK On Device OCR with Mobile OCR Engine
Payments as you go,
Developer & Runtime Licensing
Maintenance Costs not needed, service always up to date yes, to get major new technology versions
Project Licensing low volume pricing only online sales,
high volume production licenses on request
yes, on request

See also:

This website uses cookies which enable you to see pages or use other functions of our websites. You can turn off such cookies in your browser’s settings. If you continue to use these pages, you consent to the use of cookies.