TestPlant and ABBYY FineReader Engine at #ABBYYSummit16

80 views

Published on

// www.ABBYY.com/resp/promo/test-automation //
Jonathan Gillaspie, CDO at TestPlant, discusses the process of integrating FineReader Engine into eggPlant Functional, and the impact this has had on TestPlant’s success.

Published in: Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
80
On SlideShare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
5
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide
  • Welcome
  • Automated testing tool eggPlant.
    Visual tool for automating/testing systems through the user interfaces
  • Two System model
    From any desktop platform to anything with a user-interface
    Desktops, Mobile, Automotive (IVI), Vending Machine, IoT
    Non-Invasive
  • -- This slide introduces the key differentiator of ePF, image-based testing. The intent is that we engage with the software exactly as a user does which is why the title is “User-centric test automation” rather than “image-based testing”. So we should introduce it that way…
    -- The key difference with ePF is that it interacts with the SUT exactly as a user would, via the screen, swipes, text, etc. Not by looking at code-level properties and injecting changes. We do this using intelligent image recognition and OCR…
  • * 10 Billion Mobile Devices
    200 Languages
    -- ePF’s UI approach means that it doesn’t depend on the underlying technology.
    -- This means we can test any device, operating system, browser or technology.
  • In the past…
    Clipboard
    Text Image Generator
  • Problems
    Sub-pixel aliasing
    Low Resolutions
  • Unprecedented accuracy and Reliability
    98%+

    Learning was powerful but disruptive to testing

  • eggPlant for Localization testing…
    Value of -- CJK

  • Use’Mame
    Value of Dictionaries
    Pattern Matching (Regular Expressions)

  • TestPlant and ABBYY FineReader Engine at #ABBYYSummit16

    1. 1. ABBYY Technology Summit 2016 Jonathan Gillaspie CDO TestPlant © ABBYY Confidential
    2. 2. © ABBYY Confidential 2
    3. 3. Two-System Model © ABBYY Confidential 3 From any desktop platform, to anything with a user interface
    4. 4. Full User-Interface Control © ABBYY Confidential 4
    5. 5. Universal © ABBYY Confidential 5
    6. 6. Demo of eggPlant © ABBYY Confidential 6
    7. 7. How do we deal with text? © ABBYY Confidential 7
    8. 8. OCR © ABBYY Confidential 8
    9. 9. Accuracy • Previous Accuracy: 80-90% • With FineReader Engine: 98+% • Can achieve 99%+ accuracy with machine learning technologies but… – Machine learning was too aggressive – TestPlant needed to deactivate the learning – Now consistently makes the same errors © ABBYY Confidential 9
    10. 10. Universal Platform Support © ABBYY Confidential 10 • SDK on Windows, Mac, Linux • English, European Languages • Chinese - Japanese – Korean (CJK)
    11. 11. Characters vs. Words © ABBYY Confidential 11
    12. 12. OCR in eggPlant © ABBYY Confidential 12
    13. 13. Questions? © ABBYY Confidential 13 Jonathan Gillaspie CDO TestPlant
    14. 14. © ABBYY Confidential 14 Thank You!

    ×