SpeakerFind

• 

  Display of when speaking is detected

• 

  Program sends a desktop notification of speaker when found

Inspiration

As many students are forced to learn remotely, teachers are also trying their best to adapt. One thing they are very limited on, however, is the administration of exams. Each day, students devise new ways to cheat, and my intent is to help prevent that. One method that teachers use for exams is having everyone turn on their camera in a Zoom call. However, in my 500+ student classes, it's very hard to track everyone, along with the fact that everyone is on mute. If there is some way to still detect cheating through a muted video, the amount of cheating will significantly decrease.

What it does

• Uses machine learning and artificial intelligence to analyze patterns in lip movement to detect speaking.

How I built it

• Utilizes Python along with the face_recognition, OpenCV, os, and math packages

Challenges I ran into

• Overestimating the amount of time allotted
• Figuring out which packages to use
• Thinking of a project

Accomplishments that I'm proud of

• Completing an almost complete product
• Learning more about artificial intelligence
• Completing my first hackathon

What I learned

Although the amount of time allowed seems like a lot, it goes by really fast. I didn't spend my time wisely and ended up having to rush in the end.

What's next for SpeakerFind

I want to brush up on the speaker detection algorithm and produce more precise results. I also want to try and integrate the program into a website or app. Also, I think it would be really cool to try out this product in an actual video call platform.

Built With

• face-recognition
• math
• numpy
• opencv
• os
• python