A web app for discerning the emotions behind text messages and images, intended to help people with ASD who are struggling with the remote nature of the COVID-19 Pandemic.
This app was made with React and Flask, and makes use of Google Cloud Computing NLP API & Tensorflow to analyze the seniments of text and images.
The app allows users to analyze any text they wish, i.e a text they received from a friend.
Upon submitting, the user can see whether there may be strong emotion behind it or whether it is likely positive or negative.
A larger absolute number means it is likely a stronger emotion behind the text.
Users may also upload images, or take a live picture from their own webcam.
Upon submitting their image they can see a bar chart breakdown displaying the likelihood that their picture shows each emotion. There are 7 possible emotions that the model checks for.
Link to trained model: https://drive.google.com/file/d/1KzIxlzgZN-eQk_f45SDd-VY_F7bGEi4x/view
MIT