Challenge Description
Autistic people tend to be visual learners. Several of non-verbal autistic use electronical devices to communicate using pictures. This challenge consists of implementing the other way around, recognizing voices and translating it to pictures to increase the understanding of autistic people with limited verbal language.
Challenge Resources
Autistic people are normally visual learners. Around 95% of what they learn comes from visual information. Although they can hear and several times understand what you say, the pictograms would increase their understanding of what it has been said.
Voice Recognition
Natural Language Processing
Connect Text to Pictograms
There are 2 data sources that needs to be considered to be able to implement this challenge:
1.- Models for voice recognition to text
2.- Data sources for the pictograms. In this case our recommendation is to use ARAASAC pictograms
Interesting related papers:
- Céline Vaschalde, Pauline Trial, Emmanuelle Esperança-Rodier, Didier Schwab, Benjamin Lecouteux.
Automatic pictogram generation from speech to help the implementation of a mediated communication. Conference on Barrier-free Communication, Nov 2018, Geneva, Switzerland. ffhal-01880744f - https://cran.r-project.org/web/packages/googleLanguageR/index.html
- https://cloud.google.com/speech-to-text
- https://realpython.com/python-speech-recognition/
Your voice is one of the ways to recognize you. Recording and analyzing voices from someone it might be subject to strict regulations through GDPR and other ethical regulations that you need to investigate in detail.
If you plan to convert this as a certified medicine product in Autism, there are plenty of regulatory steps that needs to be full-filled in order to be commercialized as a certified product. Please inform yourself of the relevant steps to achieve this.
Graphical Explanation
Data Product – Potential Market
This data challenge could be transformed in a formal data product or solutions that could have its own market. In this case it could sell in at least 2 ways:
Software as a Services
Suscription
There are already in the market SEVERAL applications that works to create pictograms to speech. So far I hadn’t identify any app implementing the other way around.
The use of pictograms also have an interesting market between the following people:
– People with limited hearing, which may include senior people.
– Foreigners trying to learn a new language
– Tourists using it during their travel to other countries where they don’t speak the language