The problem Auto-Call solves

Our project will allow students to mark their attendance in a hassle free way. This will be a paperless method to mark attendance as opposed to the traditional way of taking attendance using pen and paper. This will also increase the efficiency and productivity as it will increase the teaching learning time.

Attendance Tracking: Facial recognition systems can help educational institutions track attendance accurately and efficiently, which can be useful for a variety of purposes, such as grading or tracking student progress.

It will also allow flexible and personalised time for evryone to mark attendance.

This project can be expanded for office and school uses.

Challenges we ran into

Challenges we ran into:

1. Initially, we wished to use the Viola Jones Haar Cascade Classifier available as part of the OpenCV library for face detection, however we came across the Histogram of Oriented Gradients (HOG) method that is available with dlib and used that instead.
2. While using pip to install dlib, we realised that it is available only as source code, hence needs a C++ compiler. We were building the project in Python and so needed to install CMake as well as Visual C++. However, even after installing the requirements, we were unable to install dlib as an error stating that Visual C++ is not installed kept popping up.
   We resolved the issue by going directly to a GitHub repository for dlib and downloaded the latest version compatible with our machines.
3. The face recognition code written using the library face_recognition was extremely slow with many lags in the live webcam feed. To resolve this, we only processed every other frame and rescaled the current frame to a small size before processing. This was a method detailed by the author of the face_recognition library in the documentation.
4. However, the project still was quite slow since the images of people already identified were being stored in the database and encoded repeatedly each time the face recognition sub-program was run. To optimise the process, we directly saved the image encodings into the database and skipped the step of saving images.
5. All three members of the team were from different colleges over India and also resided in different states. It was a challenge coordinating the work through different college schedules and communicating only virtually