PLEASE READ THE DESCRIPTION CAREFULLY AND APPLY ONLY IF YOU HAVE RELEVANT EXPERIENCE AND KNOW HOW TO FIX THIS ISSUE!
It is an educational mobile app for Android for kids 2-4 years old - Learn the English alphabet. The idea is when kid presses a button and speaks a letter from the alphabet - image of the letter is displayed on the screen.
The main point is the speech recognition everything else has been made. Developed with Unity.
The current speech recognition performance is poor. Streaming speech recognition is used for recognizing letter sounds BUT most of the times it doesn’t recognize the letter spoken, sometimes it shows different image, sometimes it recognizes the letter but shows image with delay.
With google cloud there was a big delay in response so it is also not acceptable and it requires user to press stop service button after voice recognition finishes. It is not easy to be used by 2-3 year old children.
The job is to improve the speech recognition performance to be excellent!
If you have any questions feel free to ask.
9 freelancers are bidding on average €11032 for this job
I take pride in my attention to detail, and technical expertise required to deliver high quality, bug free code for your job. Thank you for your time and consideration Best regards