Real-time speech recognition ?
Hi,
I am a 53 year old man who has progressive MS that has led to my normal speech being virtually unintelligible. I am practically bedbound (I live in a care home) and I cannot control my left arm or leg so sign language is not an option for me. My typing with my right hand is done with only one finger and so standard text to speech software applications are far too slow for me to have a normal-speed verbal conversation.
I have participated in research into EEG and ,although specialist equipment was used, I have seen that it is possible to use an EEG headset to interpret brain activity and generate text on a screen.
I am very keen to use a bluetooth eeg to record brain activity whilst I am listening to an audiobook. Both files can then be compared using ai.(possibly ChatGPT)
I would then record my brain activity whilst reading text from an old school exercise book that has never been published on the internet or even typed before. If the ai analysis of the first files can be used to interpret the second brain activity then this should prove the principle and it should be possible to generate real-time speech from measured brain activity.
I that realise real-time speech generation is considered to be the holy grail of speech generation but I genuinely believe that the technology exists to make it possible.
If I could have the use of a bluetooth eeg headset I will use the methods described above to produce a real-time speech generation solution.
All of the methods described above can be seen on many published research documents however there is one aspect of my procedure that ensures that the solution development is entirely unique.
I would be very interested in all comments on this matter.
Comments
Hello Greg,
Am I understanding that you participated in EEG research that enabled it to recognize your silent speech and display that as text on a screen? Can you mention the organization / research project that did this work? I assume there in the UK.
There are a few different research projects that use 'machine learning' (AI) techniques to recognize patterns.
Here is a 'review paper' that contrasts various imagined speech projects:
https://pmc.ncbi.nlm.nih.gov/articles/PMC9086783/
This Google Search query uncovers more research papers:
https://www.google.com/search?q=cortical+eeg+speech+recognition
Let me know the project you were associated with.
Best regards, William
The research was conducted by a team from The University of Essex. I do not know the specification of the eeg cap or the signal amplifier that was used but I do know that they were using a “speller” software system that narrowed down the chosen letter by recording my brain activity as the software highlighted a column and or row that contained the letter. (this solution was far too slow for a conversation to happen but it did give me hope that a non-invasive solution is possible.)
Yes, here is a commercial speller based on OpenBCI hardware. I'm unclear the current status or pricing, you can contact them.
https://conscious-labs.com/
https://conscious-labs.com/eeg-devices/11-8-channels-EEG-ConsciousMinds.html
Conscious Labs is located in Paris and also supplies these electrodes in our shop:
https://shop.openbci.com/products/thinkpulse-active-electrode-8-channel-starter-kit
It seems that most of the research is hampered by a poor s/n ratio.(which is probably adversely affected by the skull ). From looking at many of my personal MRI scans it would appear that the majority of scarring (sclerosis) is situated just below my brain stem hence normal, speech related, brain activity is generated by my brain and sent to my nervous system but cannot progress beyond the scarring . However, the associated brain activity could be detected at a point just above my brain stem, at a point that is NOT affected by my skull.
Does anyone know of an EEG device that can detect brain activity in the area just above the brain stem?
I believe if you run this Google Search query, you will get the same output from Gemini AI, that I did just now.
https://www.google.com/search?q=detecting+EEG+from+brain+stem?
I've pasted a partial clip of that below, minus the links:
@gregward44, did you try contacting Conscious Labs at the links I gave on September 18? Their speller uses a computer monitor screen with flickering labels for the alphabet letters / numbers. The EEG headset then picks up these visual evoked potentials and allows the subject to type text. It can also be customized to allow other menu-type selections, such as home automation equipment.