Skip to content

Latest commit

 

History

History
42 lines (31 loc) · 1.4 KB

2019-10-24-Henri_Helvetica-fAuLT_Text.md

File metadata and controls

42 lines (31 loc) · 1.4 KB

fAuLT Text

Speaker: Henri Helvetica
Twitter: @HenriHelvetica Deck Slide: https://www.dropbox.com/s/9i5rz4vdb3l6aco/fAuLT_TEXT_v1_PDF.pdf?dl=0

Web Performance

When speaking Compression of images think about: loading time, formats & sizes

You can't talk about accessibility without talking about user experience.

Missing ALT Text on images leads to poor user experience

Some data, WEB AIM

  • 68% of sites, missing alt TEXT
  • Tenon.io: issues by affected users, nearly half where vision related
  • Top test failures fit in 5 categories, ALT text is among them

Alarming Data

Everyday 2 billion photos are uploaded to web.
1.36 Billion photos are missing ALT text

AUTOMATE ALT TEXT (AAT)

2.2 billion have vision impairment across the world. (roughly 1/4 of the worldwide population)

  • We're applying Computer Vision (CV) to add automated ALT TEXT
  • Machine learning & artificial intelligence CV is hard!
  • The data set provided to the algorithm is really important.

IMAGENET (website with 15M images In 22k categories)

CV FAILURES

Gender was misidentified in

  • Up to 1% of lighter skinned males
  • Up to 7% on white females
  • Up to 12% on dark skin males
  • Up to 35% on dark skin females

References

The WebAIM Million
http://webaim.org/projects/million