Student Tracker

Automate your Attendance Using Facial Recognition

Introduction • Requirements • Installation • Usage • How it Works • Thanks ❤

---

Introduction

Detect & Recognize Faces from Live Feed, Static Image or Video. Attendace is marked & saved in Csv format. Graphical User Interface is designed & build using Tkinter. Built using face recognition library that is using state of art dlib's facial recogniton having 99.38% accuracy.

Requirements

• Python 3.3+
• macOs or Linux or Windows
• Cmake

Installation

Get it up & running

• Download our program from here
• Unzip the downloaded zip file
• Install all the dependencies from the requirement.txt

Building the source code

1. Clone the repository

git clone https://github.com/arhamshah/Attendance-Face-Detection.git
cd Attendance-Face-Detection

2. Install Cmake

• MacOs or Linux

brew install cmake

• Windows

Install Cmake & build or download a pre-configured enviorment of windows-vm here

3. Download & Install all the Dependencies

pip install -r requirements.txt

Usage

Checkout Video Tutorial

Adding Image to the database

• In order to add a person to the database, Enter name in the text box & choose "Add Image to Database" option.
• Capture user's image & it would be saved in database/.

Marking Attendance Using Webcam

• Choose "Start Program with Live Camera" option & Attendance would be updated in attendanceWebcam.py.

Marking Attendance by Importing Image

• Choose "Import Image/Video" option & select folder where image is present. Attendance would be updated in attendanceImage.py.

Marking Attendance by Importing Video

• Choose "Import Image/Video" option & select video. Attendance would be updated in attendanceVideo.py.

Accessing Attendance

• Choose "Open Attendance Sheet" option & select mode by which attendance is marked (i.e. live video, image, video).

How it Works

Checkout article by Adam Geitgey on Face Rencogniton

• Face is detected by Hog algorithm
• Face detected is encoded by 128 measurements & saved for recognition
• When program is initiated User's face is similar detected & encoded by 128 measurements
• Later these encoded measurements are compared for recognizing face from Database
• If encodings are matched, Attendance is written in Csv File with Name & Time

Thanks

• Adam Geitgey for creating face-recognition library to provide an easy way of using of dlib's state of art recognition model.
• Davis King for creating dlib, which provides facial features, face encoding models & face detection algorithms.
• Shoutout to developers & contributors of OpenCv, Pillow, Pip, Numpy, Scikit-Image, Tkinter, Scipy.