TVS Credit EPIC IT Challenge - Case Study 2

Overview

The application contains two modules, one is Automated KYC system and the other is Video based Customer Assessment for lending purposes.

Module - 1

• Prompting user to upload documents for KYC.
• Matching faces in Photo ID and Live Image using SSDMobileNet and FaceRecognition model from Faceapi.js
• Storing KYC document and current location in database for future reference

Demo

module1_automated_kyc_tvs_EDIT.mp4

Module - 2

• Connecting Agent and Customer via video call.
• Allow Agent to Verify Customer through Face and Geolocation verification against the documents provided during KYC
• Upload / Download Documents between the users

Demo

Untitled.mp4

How to Run

• Clone the frontend and backend repo on your system
• Run npm install in both frontend and backend folder
• Start frontend by using command npm start
• Start backend using command node server
• Start peerjs using command peerjs --port 4201
• Access the client on http://localhost:3000

Technologies Used

Frontend

• Reactjs
• React-bootstrap
• Peerjs for peer to peer video call
• Socket.io-client for Realtime updates
• Navigator API for Webcam and GPS access
• LocationIQ API for Reverse Geolocation
• Faceapi.js for Face detection and Recognition

Backend

• Nodejs
• Expressjs
• MongoDB Atlas
• Socket.io
• Multer