BookFace

• 

  BookFace

• 

  Login

• 

  Initialise Account

• 

  Verify Initialisation

• 

  Scan others

• 

  See chats with others

Inspiration

Social media and most networking services have built a world where individuals are disconnected and separated from one another. We are often following and talking to others we have never even met in person. Therefore, we decided to develop BookFace, an intuitive social app that makes physical interaction a necessity in networking. We aim to inspire the humanity in individuals and improve the human connection aspect of social networking.

What it does

BookFace enforces physical, face-to-face interaction before connecting online through facial recognition. The process is as follows:

1. The user sets up a profile like usual but they must upload a verified picture of their face to their account
2. Users are then only allowed to chat with others after scanning their face in person

How we built it

BookFace is a mobile application built in SwiftUI. It is also connected to a Python based server that runs a Convolution Neural Network to detect faces and recognise users. This is all connected using Supabase as a backend.

import SwiftUI
import Supabase
import MTCNN

Challenges we ran into

1. Hyper-tuning and optimising a complex neural network
2. Creating a functional and high speed cloud server
3. Doing it all in 24h

Accomplishments that we're proud of

1. Creating a proprietary image verification system
2. Implementing a state of the art machine learning pipeline
3. Creating a functioning cloud server to send and receive data instantly
4. Choosing softwares and tech stacks to work with

What we learned

1. Creating data pipelines
2. Pivoting quickly between different ideas
3. How to develop in a team by playing to our strengths

What's next for BookFace

1. A fully functioning messaging chat
2. A social media page
3. Enhanced verification

Built With

• machine-learning
• opencv
• python
• supabase
• swiftui