Skip to content

Latest commit

 

History

History
58 lines (35 loc) · 2.96 KB

README.md

File metadata and controls

58 lines (35 loc) · 2.96 KB

Introduction

Variable Value
Name Abhishek Ramesh Gadekar
Institute Indian Institute of Technology, Jodhpur
Department Elecrtical Engineering
Graduation Year 2024
Email [email protected]

Are You Working? 😉

Facial Recognition for verifying if the correct individual is taking test, verifying if he logged in and logged out during the correct peroid. This can be used for monitoring purposes and for identification purposes.

abhi

How It Works?

  1. Run the app by clicking start

Screenshot 2022-10-31 at 10 33 43 PM

  1. Press clock in and the application identifies you and starts counting your logged in time. (Great for remote working companies)

Screenshot 2022-10-31 at 10 34 06 PM

  1. Log out by pressing clock out, it shows the the time peroid for which you worked and makes sure the user was infront of the screen for the time peroid.

Screenshot 2022-10-31 at 10 35 11 PM

System Design :

       A cross-platform app was developed using PyQt5, Python, OpenCV , Face_Recognition & Qt-Designer

The AI functionality of face_recognition was implemented in the backend. The functionality was implemented in the backend because using it one can make necessary logic implementations to do the correct prediction .

1. Frontend : PyQt5 & Qt-Designer ( to develop cross platform functional GUI)

2. Backend : Python, OpenCV , Face_Recognition , other libraries

Install Dependencies:

  1. pip install dlib==19.18.0 (the newwer version may create errors)
  2. pip install face-recognition
  3. pip install opencv-python
  4. pip install numpy
  5. pip install PyQt5 (check below if working on M1 mac for installation)

Steps to Run the Application:

  1. Clone the repository and open the project in pycharm and run the mainwindow.py and/before install all the necessary packages stated above. (for windows or intel based macs, m1 mac users please download intel based pycharm and run the mainwindow.py file)

Future scope of the project:

  1. The current model doesnt ditinguish between a photo and real person and thus can be easily fooled. It can be resolved mostly by implementing a full body detection algorithm.
  2. Apart from that there are some bugs(logical erros) which I couldnt resolve due to less time available while making this project.
  3. It is not developed forr multiple users, therfore it can be scaled up to handle simulataneous monitoring of multiple users parallely.