face verification

Introduction

• This is the final project of python course at NCTU, a prototype of course roll call system.
• Compare faces to the trained model of dataset through haarcascade which is provided by opencv.
• Roll call results would be maintained in the database.

---

User Guide

• Install required pacakges.
• Run main.py to go through the entire process
  • Register student id to the database
  • Verify ID by face
  • Do the course roll call.
• Run database.py to obtain contents of roll-call table.

---

Installation

• packages: sqlite3 / numpy / pandas / opencv-contrib-python
• Install packages via pip

pip install numpy
pip install pandas
pip install opencv-contrib-python

---

Database interface

• Each course has its own database file. e.g. Calculus.db
• Inside the database, create a table named member_list for roll calling by default.
• We implement this database interface via sqlite3, which has been pre-installed already.

---

Result

• Multi-face detection
• Before roll call
• new student registeration

---

Issue

• Poor Performance of going through the whole process of main.py
• Lack of verifying student id when registering course.