Skip to content

README.md

Latest commit

 

History

History
84 lines (68 loc) · 2.92 KB

README.md

File metadata and controls

84 lines (68 loc) · 2.92 KB
Raw

Banking Management Desktop Application

A desktop banking system with AI facial recognition for secure transactions.
This project was developed as an academic assignment to practice Fullstack Development with Python and Tkinter.

📌 Project Information

  • Team size: 4
  • Role: Fullstack Developer
  • Duration: Feb 2025 – May 2025
  • Languages/Frameworks: Python, Tkinter

📌 Overview

The system is designed to manage banking operations including customer accounts, transactions, and services.
It integrates AI facial recognition to enhance security for high-value transactions.

📌 Features

  • Authentication (login/register)
  • Main Bank Interface for account management
  • Transaction History tracking
  • Money Transfer functionality
  • Settings and Other Services
  • AI Facial Recognition for transactions ≥ 10M VND

📌 System Design

  • Database schema in Microsoft SQL Server
  • Integrated cv2 Camera and Haarcascade for face detection
  • Built Training Pipeline with real face dataset, monitoring overfitting
  • Implemented secure transaction flow with facial recognition

📌 AI Training Pipeline

The project integrates an AI facial recognition system to secure high-value transactions. The training pipeline was designed to process real facial datasets and includes the following steps:

1. Data Collection:

  • Gathered a dataset of real face images.
  • Organized into training and validation sets.

2. Preprocessing:

  • Applied image resizing and normalization.
  • Used Haarcascade for face detection.
  • Augmented data to increase variability.

3. Model Training:

  • Built a recognition model using OpenCV (cv2).
  • Trained with supervised learning on facial features.
  • Monitored training and validation accuracy.

4. Overfitting Monitoring:

  • Compared training vs validation performance.
  • Identified signs of overfitting (high training accuracy but lower validation accuracy).
  • Adjusted dataset size and preprocessing to mitigate.

5. Integration:

  • Connected the trained model with the Tkinter banking interface.
  • Facial recognition triggered for transactions ≥ 10M VND.
  • Ensured secure flow before transaction approval.

📌 How to Run

  1. Clone the repository: 
    git clone https://github.com/DuyKhoiCoder30062004/Python-PyGroup3.git
  2. Install required dependencies: pip install opencv-python pip install pyodbc
  3. Configure the SQL Server connection string.
  4. Run the application: python main.py

📌 Screenshots

📌 Contribution

Developed by a team of 4 students as part of academic coursework.

📌 License

This project is for educational purposes only.