Inspiration
One of the biggest problems of online education is to prevent students from just joining online classes and not attending them. I myself have experienced many times in hybrid classes that students join the class but when called by teachers, who is busy in teaching the students in the offline mode, do not respond and later when asked to switch on the video often give the excuse of poor network connectivity. It is a major problem because it is deeply harming the future workforce of our nation and we do not have a check for it. This problem is not a small problem and is affecting the future generation adversely.
What it does
The video calling application will have an in-built bot that will use users’ registered images for face recognition and mark attendance. The bot, once enabled, will start taking snapshots after random intervals of time (0-40 seconds) every 5 minutes without requiring the participant to turn on the video. These snapshots will then be processed by face-api.js, a JavaScript API for face detection and recognition in the browser implemented on top of the tensorflow.js core API, to recognize the face registered with the account. If it detects and recognizes the face, it will mark the participant present for that particular timestamp. Later, the results of all timestamps will be used to evaluate if the participant was present for the whole duration of time which will be indicated in the participant’s panel.
How we built it
I first designed the app first on Figma, I started with coding it out using ejs and nodejs. We used face-api.js for implementing the face-recognition feature.
Challenges we ran into
It was quite challenging to work with face-api.js for the first time but tutorials from Web Dev Simplified helped me create that functionality.
Accomplishments that we're proud of
This was my first ML project and I am proud to step into the world of AI-ML :)
What we learned
I learned using ejs (it's pretty much the same as HTML) and using the face-api.js library.
What's next for Attentive
I will try to scale it up and make turn it into a production-level app.
Log in or sign up for Devpost to join the conversation.