In traditional classrooms, it is practically
impossible for the teacher to pay attention
to every student present in the class.
A system that could maintain the
record of each student’s performance on
the basis of their activeness in the class is
required.
Mark student presence automatically
Record each student’s performance on the basis of their activeness in the class.
Calculate attention level of each student with the teacher.
Generate an automated feedback for the teacher so they know which topics are needed to be taught in “a nicer way”.
Express.js used for creating APIs
Python 3.7 for computer vision
MongoDB for database
Flutter for UI
Application Programming Interfaces (APIs)
The software uses Representational State Transfer (REST) architecture.
APIs interact directly with the database to perform Create, Read, Update, Delete (CRUD) operations.
HTTP calls are made to make GET/POST requests to the API URL.
API Hosted on Heroku Platform
To integrate the modules we developed an algorithm, Aditya-Chirag-Rahul-Ankit (ACRA) Algorithm
The algorithm requires two input parameters: n and z
where, n is the total number of frames in the video and z
is the interval number initialized by the user as desired.
Every time the algorithm runs, it captures each face present
in the frame and crops them for better detailing.
The head angle function calculates the angle of the cropped
frame containing the face.
The function returns a binary value depending upon the angles
calculated.
The data is saved with the timestamp.
After the final iteration of the algorithm the data fetched is
stored on MongoDB which is ready to be used by the app.
Face found more than 75% of iteration will be marked present
for the class (83% for the case where z=6 and number of times face found=5)
The accuracy of the algorithm depends directly on z, i.e., the number of intervals.
Teachers could easily access the system and it will be feasible for them to use it anytime and anywhere.
Flutter SDK is used to make an application for accessibility of the data in a proper and systematic manner.
Head Angle Detection of student
Students in a classroom
Application UI
Automated the process of taking attendance.
Calculated each student’s activeness in the class with the help of head pose estimation module.
Successfully implemented the integration algorithm which helped in storing the useful data on the MongoDB remote database.
Generated an automated feedback mobile application for the teacher to observe the alertness of every student.
