Sign language is a essential communication system for deaf people, which plays a crucial role in promoting inclusion in modern society and which, like other tools for people with other types of disability, deserves to be highlighted so that it can be developed and made widely available.
By recognising and supporting sign language, a path towards equality is opening up and the full participation of all people in various spheres, and promoting its use in everyday life, as well as within companies and organisations, not only guarantees access to information and communication, but also enriches organisations, fosters teamwork and strengthens the corporation’s most human values.
In this post, we’ll look at how the technology and, specifically, Sign4all, the first sign language translation app, are setting real milestones for inclusion.
Discover Sign4all: the first sign language translation app
It is incredible how technology is driving change in every aspect of our lives and bringing about improvements in social areas that were perhaps not so visible to the wider public, such as the importance of sign language.
Currently, in Spain, 98% of deaf-mute people use standardised sign language, which means that, in order to communicate with them, we should all learn it in the same way that we learn English and other languages. In fact, why learn a foreign language and not another one that gives us the chance to communicate with people who are much closer to us?
Sign4all, the first sign language translation app, which is capable of recognising and interpreting the Spanish sign language alphabet in real time, was developed by the Robotics and 3D Vision Group at the University of Alicante with the aim of breaking down barriers between deaf and hearing people.
How does Sign4all work?
Thanks to the deep learning technology and various computer vision techniques, this app captures and interprets sign language, translating it into the mainstream language.
To help us understand how the sign language translation app works, its mechanism involves capturing the person and analysing every detail of the movements they make with their arms and hands. It then encodes the left side of the body in blue and the right side in red, whilst ensuring the user’s anonymity at all times.
Of course, we mustn’t forget that effective communication is always two-way, so the app also works the other way round, using a virtual avatar that signs the words as the listener types them in Spanish.
And the best thing about it is twofold: firstly, sig4all can translate conversations in real time; and secondly, thanks to the tool’s low cost, it is accessible to anyone who needs it when they are unable to be accompanied by an interpreter.
How does deep learning work?
As we’ve mentioned, the first sign language translation app is based on deep learning technology, but would you know what that actually involves? Pay attention!
Deep learning is a technique for artificial intelligenceinspired by the way the human brain works to learn and recognise patterns.
For example: imagine you’re teaching a computer to recognise cats in photos.
Instead of telling it directly how to recognise a cat, you show it lots of photos of cats over and over again so that it can learn what they look like. In this way, the system begins to look for common patterns, such as shapes, colours and textures, so that, as it sees more and more photos of cats, it adjusts its internal connections (known as «artificial neurons» or «nodes») and improves its ability to identify cats on its own.
Just like a child’s learning process, in which increasingly complex concepts are integrated and more neural connections are formed; in deep learning, layers of these «artificial neurons» are used to process information at ever-deeper levels. Each layer detects specific features and passes this information on to the next layer to carry out more comprehensive and detailed analyses.
Once the system has seen enough cats, it is able to generalise and recognise them in images it has never seen before, because it has learnt to detect the essential characteristics that define the animal.
Returning to Sign4all, which is the subject at hand, the mechanism would work in the same way: the more signs it sees and words it hears, the better equipped it will be to act as a translator and interpreter between deaf and hearing people.
Find out about the latest developments to date from Sign4all
The developers of this sign language translation app are currently working on a collaborative project with the Spanish Language and Sign Languages Research Group (Griles) at the University of Vigo, who are expanding the database that Sig4all will use in the near future.
We hope you found this post on inclusion, deep learning and the digital revolution interesting. Subscribe to Educa.pro for more professional development!
