MishMash hackathon is India’s largest online diversity hackathon. The focus will be to give you, regardless of your background, gender, sexual orientation, ethnicity, age, skill sets and viewpoints, an opportunity to showcase your talent. The Hackathon is Live from 6:00 PM, 23rd March to 11:55 PM, 1st April, 2020
Azure Services Used: Azure App Engine(for deployment), Azure Speech to text(to identify voice of students).
🔥 Your Pitch
Problem Statement: As more and more people move towards online education, the value of these certifications becomes important as anyone can go online and play the videos without ever watching them just to get the certification. So companies and institutes of higher educations find It difficult to assess their true credibility. With answers to assignments for popular courses available easily online, one does not need to learn anything to get certified.
Proposed Solution: We want to change this by using Deep Learning techniques to assess the Attention of students in online classes by using different features: Eye tracking, Changes in Facial Expressions and voice inputs(for live classes). We are proposing a proprietary algorithm that takes these features and calculates a score that represents the attention of a student during a lecture.
This score can then be used as a metric to identify whether the student deserves to get certified or do they need to start over and pay more attention in classes. The Market size of ed-tech in India alone is worth billions, and through this system, we want to be the ones who can bring credibility to online certifications.
Detailed Methodology: The webcam of a student is used to take their video while they are attending an online lecture and this video is fed to the model, which then uses eye-tracking to ensure the student is watching the screen, and at the same time, a Facial Expression Recognizer is used to detect if the Expressions are changing with time, as research suggests when humans are paying attention their expressions are changing constantly. Similarly, a study conducted recently by Microsoft suggested that the average attention span is only 8 seconds, to account for this our solution rates the attention of students every 10 seconds. So their changing attention throughout the lecture can be observed and quantified. The Azure voice to text is also used in the system which can be used for live classes held online as if any students are interacting with their teacher it indicates that they are in fact paying attention.
The application is packed as a web app, where you can upload the video feed of any student and then watch the system grade their attention in real-time, updating every 10 seconds.
🔦 Any other specific thing you want to highlight?
The ongoing pandemic is going to change the way people perceive the Ed-Tech sector, right now many view it as a supplement to regular classes, but this situation has shown us it might very well become the normal way of teaching, but one of the fundamental issues that online learning faces is to assess whether a student is actually listening, the current infrastructure does not account for this and through this service, we want to fill this gap, to make online education credible.
✅ Checklist
Before you post the issue:
[X] You have followed the issue title format.
[X] You have mentioned the correct labels.
[X] You have provided all the information correctly.
ℹ️ Project information
Project Name: Old Skool Analytics
Short Project Description: Using Deep Learning to measure the attention of students in online classes.
Team Name: Brown_Boys
Team Members: Manmeet Singh Sethi @ManmeetSethi, Rajveer Beerda @rajveerbeerda
Demo Link: https://youtu.be/8Ew_A-mB6b4
Repository Link(s): https://github.com/rajveerbeerda/attention-score
Presentation Link: https://drive.google.com/file/d/1zVDN6i3jQcvSt0Rz9-KSuqphL_7REqI0/view?usp=sharing
Azure Services Used: Azure App Engine(for deployment), Azure Speech to text(to identify voice of students).
🔥 Your Pitch
Problem Statement: As more and more people move towards online education, the value of these certifications becomes important as anyone can go online and play the videos without ever watching them just to get the certification. So companies and institutes of higher educations find It difficult to assess their true credibility. With answers to assignments for popular courses available easily online, one does not need to learn anything to get certified.
Proposed Solution: We want to change this by using Deep Learning techniques to assess the Attention of students in online classes by using different features: Eye tracking, Changes in Facial Expressions and voice inputs(for live classes). We are proposing a proprietary algorithm that takes these features and calculates a score that represents the attention of a student during a lecture. This score can then be used as a metric to identify whether the student deserves to get certified or do they need to start over and pay more attention in classes. The Market size of ed-tech in India alone is worth billions, and through this system, we want to be the ones who can bring credibility to online certifications.
Detailed Methodology: The webcam of a student is used to take their video while they are attending an online lecture and this video is fed to the model, which then uses eye-tracking to ensure the student is watching the screen, and at the same time, a Facial Expression Recognizer is used to detect if the Expressions are changing with time, as research suggests when humans are paying attention their expressions are changing constantly. Similarly, a study conducted recently by Microsoft suggested that the average attention span is only 8 seconds, to account for this our solution rates the attention of students every 10 seconds. So their changing attention throughout the lecture can be observed and quantified. The Azure voice to text is also used in the system which can be used for live classes held online as if any students are interacting with their teacher it indicates that they are in fact paying attention.
The application is packed as a web app, where you can upload the video feed of any student and then watch the system grade their attention in real-time, updating every 10 seconds.
🔦 Any other specific thing you want to highlight?
The ongoing pandemic is going to change the way people perceive the Ed-Tech sector, right now many view it as a supplement to regular classes, but this situation has shown us it might very well become the normal way of teaching, but one of the fundamental issues that online learning faces is to assess whether a student is actually listening, the current infrastructure does not account for this and through this service, we want to fill this gap, to make online education credible.
✅ Checklist
Before you post the issue: