Skip to content

An AI-powered insurance platform that detects fraud and verifies documents in real time using XGBoost, CNN-based forgery detection, OCR, and LLMs. Built for Hack-AI-Thon by SBI Life, it enhances efficiency using Neo4j.

License

Notifications You must be signed in to change notification settings

parth1899/Digi-Sure

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

AI-Powered Fraud Detection and Document Verification

This innovative insurance platform transforms fraud detection for verified document retrieval using advanced AI-powered mechanisms. By reducing the reliance on manual verification, the system enhances efficiency while delivering real-time fraud analysis through multiple layers of security checks.

This project was created for the Hackathon Hack-AI-Thon conducted by SBI.

Demo Video

Demo.mp4

Architecture

image

Technology Stack

This project leverages the following technologies:

  • TypeScript
  • Flask
  • Neo4J
  • XGBoost
  • CNN
  • Mistral OCR
  • Isolation Forest

Database Setup

To set up the Neo4j database, run the following Cypher queries:

// 1. Create a unique constraint on the email field
CREATE CONSTRAINT user_email IF NOT EXISTS
ON (u:User) ASSERT u.email IS UNIQUE;

// 2. Create indexes for better query performance
CREATE INDEX user_name IF NOT EXISTS
FOR (u:User) ON (u.name);

CREATE INDEX user_surname IF NOT EXISTS
FOR (u:User) ON (u.surname);

Contributors

About

An AI-powered insurance platform that detects fraud and verifies documents in real time using XGBoost, CNN-based forgery detection, OCR, and LLMs. Built for Hack-AI-Thon by SBI Life, it enhances efficiency using Neo4j.

Resources

License

Stars

Watchers

Forks

Languages

  • TypeScript 62.2%
  • Python 36.7%
  • Other 1.1%