Emotion Recognizer

Real time Facial Emotion classification using Tensorflow js. The model is trained on fer2013 dataset using keras. For each picture a chart is ploted showing scores of the different emotions and the dorminant emotion is the class emotion.

PICTURE :- A pretrained keras model is loaded into the browser and used to detect faces and classify emotions. WEBCAM :- Transfer learning is used to rapidly train and customize the model to this task for webcam pictures.

Instructions

To get a copy of the project running on your local machine.

Prerequisites

• Python 3.6+
• Node.js version 8.9 or above
• Vue cli
• tensorflow.js

Setting up the project

First, clone this repository to your local machine:

$ git clone https://github.com/Onroe/emotion-classifier.git

Running the App

####Extract the files

####Run the front end -Open a new nodejs terminal window, then navigate to the client folder:

Install dependencies:

$npm install

Then run the app:

$ npm run serve

####NOTE

• Make sure saved model is accessible over http as a json file:

-Check the URL where app is running - http://localhost:8080/picture.

-Check the URL where app is running - http://localhost:8080/index.