Lately I’ve been quite immersed in the world of Microsoft to find my way around the new job, and whilst doing so I discovered a few things you might have missed. Especially in the field of accessibility there is some splendid stuff happening.
During the key note of //build last week in San Francisco there were quite a few mind-blowing demos. The video of they keynote is available here (even for download).
- A “What’s in this image” demo using the Computer Vision API
- The Caption Bot, which automatically detects content in images and proposes alternative text
- The Custom Recognition Intelligent Service (CRIS), which allows you to train a system to get better results when converting audio to text.
The results of CRIS are impressive. For example, Cornelia showed off how a set of audio files with interviews with children became much more usable. On the right in this screenshot you see the traditional results of speech to text APIs and on the right what CRIS was able to extract after being trained up:
The Microsoft Cognitive Services: Give Your Apps a Human Side breakout session at build had some more interesting demos:
- Tele2 using the translation API to do live translation of phone calls into other languages. You hear the message in the original language, a beep and then the translation. (34:50 onwards)
- ProDeaf doing the same to translate live audio into sign language across languages. (40:00 onwards)
- F12 tools for Edge now have not only an accessibility tab in the DOM viewer, but also a live updating Accessibility Tree viewer
- Narrator for Windows will have a developer mode. You can turn this on using “Narrator + Caps Lock + Shift + F12” and it will only read out the app that you chose rather than the whole operating system, including your editor and the things you type in. To avoid the mistake of looking at the screen whilst using a screen reader, it also automatically hides the screen except for the currently read out part.
All in all, there is a lot of great stuff happening in the world of accessibility at Microsoft. I had a very easy time researching my talk next week at Funka.