︎

Initially, I came up with the idea of developing a chat room to display the person's name when their face appears on the video chat.

However, upon further contemplation, I realized this approach would only be feasible if I had prior data on the person's picture. Realizing this limitation, I revised my plan to create a chatroom that can detect audio and video inputs, with an additional feature that can recognize facial expressions. With this feature, users can easily read other users' emotions, which may lead to better conversation.

I used the library face-api-js .

Reference: https://github.com/justadudewhohacks/face-api.js

• face-api.js : a wrapper around tensorflow which is a popular machine learning library out there and allows to do real-time face detection in the browser.

  
  
• 
  

I wanted to make a chat room that reads facial expressions. I’ve been always thinking about how amazing artificial intelligence + machine learning can detect and read facial expressions and wanted to use that as a new interaction method for this chat.