Skip to content
/ Meet.OZ Public

Meet.OZ is a AI driven Platform specially built for Organizations and Colleges to Analyze Emotions of Participants during an Online Meeting or an Online Class.

Notifications You must be signed in to change notification settings

prkhrv/Meet.OZ

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Meet.OZ

Meet.OZ is a AI driven Platform specially built for Organizations and Colleges to Analyze Emotions of Participants during an Online Meeting or an Online Class.

How to Run

1. Clone the Repo

git clone git@github.com:prkhrv/Meet.OZ.git

2. Install Dependencies

npm install

3. Run on localhost

npm start

Tracks

Revamping the OldInnovating with AI

Inspiration

COVID transformed the way we interact and operate. It also caused a major emotional drought since humans were not designed to relate with one another through screens. As video conferencing becomes the new normal, marketers and managers continue to struggle with the lack of nuance and engagement in remote communication, and people with sensory impairments are likely to have it much harder.

What it does

Meet.OZ is an approach to revamping the old by utilizing the Power of AI. It simply takes a live video feed and analyzes the user behavior through it. Following the conversation, a dashboard presents emotional data and data-derived insights such as the most prevalent emotions throughout the call, variations in feelings over time, and noticeable contrasts or rapid shifts in mood.

How we built it

We built it using Tensorflow, OpenCV, and face-api.js to recognize faces and further assess their emotions, The Client Platform is built with React JS and Agora-SDK and of course tonnes of RED BULL !!!!.

Challenges we ran into

We started working with Azure Face API but soon realized that emotion analyses were discontinued earlier this year.

The Pretrained AI models we were trying to use were not accurate so we had to deep dive into some recent research. After scratching our heads for a very long we realized that we can ensemble multiple models trained on a variety of datasets. Also, on the client side, It was a tedious task to understand Agora SDK.

Accomplishments that we're proud of

We are proud that we were able to create a working prototype. We learned to ensemble multiple models (which I feel is pretty cool).

What we learned

We learned to ensemble multiple models and also the working of webRTC for video streaming platforms.

What's next for Meet. OZ

We want to convert it into a browser extension or a desktop app so that it can run in the background and can be used with any video conferencing software. We truly believe that creating a whole new Video conferencing platform will cause more hassle than integrating Meet.OZ as a plugin on existing platforms.

About

Meet.OZ is a AI driven Platform specially built for Organizations and Colleges to Analyze Emotions of Participants during an Online Meeting or an Online Class.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published