Voice Recognition
Last updated
Was this helpful?
Last updated
Was this helpful?
I built this Voice Recognition mobile App for the Creationwide hackathon. The motivation behind the idea is that I don't like the way the password is designed and I believe there are better ways of doing it. I wonder how many of you don't like to come up with and remember those complicated passwords and how many times you forget about them and have to reset them. I just tried to solve one issue that bothers so many people for so long with a new solution. The essence of a password is that it's unique to everyone and it's only associated with the individual who has it. I started brainstorming. I asked myself: in addition to those bunch of random combinations of numbers created by us, what else are unique to us and easier to create and maintain? The answer is Biometrics! As one of biometrics, our voice is unique. Why not use it as a password?
I did a short customer market research online and offline and confirmed this is a problem that worths solving. Given that it's a one-day hackathon, I didn't spend too much time and resources to dig into the competitive analysis. During my short research, I did find one competitor- weChat who offers voice authentication service for login. But it's not the primary authentication process. It has a totally different voice authentication process from the product I was about to design. It asks you to say a couple of numbers a few times. Then next time if you wish to use voice to authenticate, you read the numbers again. The product design idea that I had is that App should be able to authenticate people by listening to the phrase they speak or the song they sing.
Since time is a constraint, I limited the minimum viable product (MVP) features to 3 features.
The app is able to record users' voices.
The app is able to recognize users' voices next time when users log in.
If the app recognizes users' voice, it let them in, otherwise, it blocks them and throws errors.
I used a simple Hybrid Mobile Technology stack including angular js, ionic, Cordova to shape the foundation of the App. Since it's hybrid, the App is going to be able to deploy across platforms such as Android, IOS. The main tech issue that needs to be solved in this product and bring the product to success is voice recognition. It's not going to be feasible to write a voice recognition engine in a day hackathon. I did some research and tracked down a startup that provides a voice recognition engine through API. I called them and used their products for free for my demo purpose.
After all the tech issues were tackled, the final product was born. There are two buttons on the login screen. One saying record your voice and another saying login. The first time when you login, you can click the record button to record your words or just hum couple words from one song. The second time when you login, you can say the same words or sing that same song again. The app is able to recognize if that's your voice or not and decide if it allows you to login or rejects you. It's a fun project. It has a lot of places to improve and a lot more features to add on. I am still looking for better solutions to tackle passwords issues. I am glad that we already have more and more good solutions coming out these days. You can check out my project on my GitHub repo .
If you want to learn more about it, let's catch up! Send me a message or shoot me an email at