Improving the internet for the visually impaired

Making every image on the internet accessible by using machine learning to generate the captions.

Register Your Interest

The Internet Is Broken

There are 250+ million people with visual impairments 1 .

People with visual impairments frequently rely on a Screen Reader to navigate the internet. The Screen Reader reads out what’s on each page. When the Screen Reader reaches an image it has to rely on the “alt tag” to provide a caption for the image.

For the majority of the internet these “alt tags” are either missing (think of any user-uploaded content such as Twitter or LinkedIn), superficial or inaccurate.

Imagine seeing a job post where the information to apply is in a screenshot but all you hear is “Image”.

This happens to millions of people everyday.

an unaccessible image on twitter. The photo shows Satya Nadella retweeting a photo by Microsoft Research.

We rely on images for context.

But it's missing for the 250+ million people with visual impairments 1 .

The Solution

We have created a plugin which works with your existing screen reader to generate the image descriptions for every image on the internet.

It works by using by machine learning to generate the descriptions when you open the page.

The descriptions include:

  • What is in the photo
  • What the text says
  • Who is in the photo
  • What their facial expressions are

If you are interested in trying our product then please register your interest here.

a visually impaired child using a computer

Try it for yourself

an example photo of a previously inaccessible image. the photos shows A cat wearing a suit and tie. The text says What am I doing here
Without Visual Cognition:

Description: "Image"



With Visual Cognition:

Description: "A cat wearing a suit and tie. The text says What am I doing here"

News

Survey icon

We came 3rd in the McKinsey Venture Academy!


After a fantastic day and very tough competition, we came 3rd! Thank you to the judges for their excellent feedback and to Luke O'Neil for his mentorship throughout.

Survey icon

We came 3rd in the Microsoft Imagine Cup UK!


We are really pleased to have come 3rd in the Microsoft UK Image Cup. It is fantastic to see how both Microsoft, and the teams competing, are focused on #TechForGood.

Survey icon

We're in the UK Final of the Microsoft Imagine Cup!


We are honoured to be in the UK Final of the Microsoft Imagine Cup! especially given our inspiration for Visual Cognition was Microsoft Research's SeeingAI. Check out our submission where Luke Harries talks about how our infrastructure and AI is powered by Azure.

Survey icon

We're in the Final of McKinsey Ventures Academy!


We're really excited to be in the Final of the McKinsey Ventures Academy. We have received fantastic mentorship from the team at Quantum Black and we look forward to meeting the judges and other teams on the 27th of April.

Survey icon

Survey Released


We want your feedback of how we can best develop our tool! Please fill in the following survey to help us tailor it to your needs. Click here to fill in the Survery.

Visual Cognition logo

Pilot Study Announced


We proud to announce that we will be shortly launching our pilot study. It will be jointly run with UCL and the Global Disability Innovation Hub. If you would like to be notified when open applications please register your interest here.

The Visual Cognition team collecting the Microsoft Prize at HackCambridge

Winner of Microsoft Prize!


VisualCognition are honoured to win the Microsoft Prize at the University of Cambridge's Hackathon. We look forward to continuing building on the Microsoft Cogntive Services Platform to deliver VisualCognition.

Our Team

With experience in Medicine, Engineering, and Machine Learning we are committed to making the internet more accessible.

headshot of Luke looking happy

Luke Harries LinkedIn Profile

Medicine and Cognitive Neuroscience
University of Cambridge

MSc Computer Science
UCL

headshot of Filip wearing a suit

Filip Kozera LinkedIn Profile

Information Engineering
University of Cambridge

headshot of Mateusz in black and white

Mateusz Staniszewski LinkedIn Profile

Mathematics
Imperial College London

We want to make the internet work for everyone.

Register Your Interest