The problem teaching-made-easy solves

Pandemic has affected almost every person irrespective of their occupation. Even though the medium of education has shifted from Offline to Online,and work from home is encouraged for everyone,yet during these tough times,teachers’ lives have become tougher. As many of our teachers have children,who too have online classes from hom,it becomes really hard for our teachers to manage their time in evaluating assignments and tests. In addition, keeping an eye on every student is also a challenge during online classes. Students of the primary school also struggle as many of them don’t have knowledge of operating computers, and parents have to sit with them throughout the class to assist them in doing so. Keeping in mind that many parents are working, online classes fail to be interactive, comfortable, and productive for many students.We have created an Assignment Checker using NLP.The checker will take the questions and their model answers/ ideal answers as input from the teacher. The students will submit their answers through a form,from which data will be fetched and fed into our Deep Learning Model of NLP following which the grade will be displayed.The web page interface allows options to create assignments, as well as show students’ performance and class performance.For keeping an eye on every student during the online sessions hosted on any platform, we have created a python program using OpenCV library, which will take the video feed of live class as input and give tags to everything it identifies as human.The tags will be - attentive, not attentive, confused, and speaking. Also, many students get unnoticed during classes as they don’t know about the feature of raising their hands and many don’t know how to raise their hands virtually.We have created another deep learning modelwhich will also take the live feed of online class as input. It will try to identify hand gestures of students, and will classify them as “raising hand”, “mute” and “unmute”

Challenges we ran into

Creating a perfect model for the task was a big hurdle, as its about assignments and if we create a incomplete model it will indeed increase troubles for teachers. Also we are frshers and just learnt openCV, so creating its model was also very troublesome. The biggest trouble we faced was of hosting our model as it was very heavy model, but we found out ng rok server as a solution. But again we need to start server again and again as it has time limit. Also it has concurrent watching limit of 20. But to host it we got a domain and we redirct it every time a new link has to be generated due to time limit of ng rok