As people age, dementia and other factors affect how they remember things, making it especially difficult to remember credentials such as passwords and pins. This presents a barrier of entry for older generation of people to access services like online banking. To solve this problem, instead of using credentials that a person has to remember as identification, they can instead log in with identification that is native to every person, their face and voice.
What it does
The goal of Alice is to reduce the barrier of entry for older generation to access online banking. Instead of verifying a person’s identification using passwords and pins, they simply have to use their webcam to take a picture of their face and speak into their microphone to identify themselves.
How we built it
We built the project with the Nuance Speech Recognition and Speech to Text API, the IBM Cloud Watson Conversation API, the Clarifai APi, IBM Watson Recognition System on BlueMix through the Cloud Foundry command utility line and Google Cloud Storage. We also designed our front-end through Bootstrap and hard-coded HTML templates and design for SASS/CSS markup.
Challenges we ran into
We ran into challenges when implementing the sox command utility line and the Cloud Foundry Utility line and it running properly in terminal.
Accomplishments that we're proud of
We're super proud of the fact that we utilized the Clarifai APi and IBM’s Watson API recognition system on Bluemix in order to design a smart lock capable of unlocking based on the facial features of an individual. We're also proud that we had the opportunity to implement two different machine learning mechanisms through the Nuance API, the Clarifai APi, IBM Cloud and Google Cloud to enable authentification through facial recognition and speech recognition.
What we learned
We learned that the IBM Cloud requires you to download the CF Command Line utility system to implement different parts of the Watson APi and how to implement machine learning API's that ranged from Nuance to IBM Cloud through React components and integrated them into a localhost server using postscript and PythonAnywhere. We also learned that integrating ML components in a web application require
What's next for Alice
We'd like to see Alice pursue finger recognition as well as development on IOS and Android platforms. As well, we would like to explore the how to store images efficiently on databases, such potentially exploring is Apple's face id algorithm to store only the key features.