https://github.com/prudhvignv/facial-emotion-recognition-using-cnn

A deep learning and image processing project used to predict the emotions of a person in image.
https://github.com/prudhvignv/facial-emotion-recognition-using-cnn

cnn dataset deep deeplearning facial-emotions google-colabs keras neural-network notebook

Last synced: about 19 hours ago
JSON representation

A deep learning and image processing project used to predict the emotions of a person in image.

• Host: GitHub
• URL: https://github.com/prudhvignv/facial-emotion-recognition-using-cnn
• Owner: PrudhviGNV
• License: mit
• Created: 2020-07-05T13:56:07.000Z (almost 5 years ago)
• Default Branch: master
• Last Pushed: 2020-09-11T12:49:21.000Z (almost 5 years ago)
• Last Synced: 2025-06-24T22:50:35.260Z (2 days ago)
• Topics: cnn, dataset, deep, deeplearning, facial-emotions, google-colabs, keras, neural-network, notebook
• Language: Jupyter Notebook
• Homepage:
• Size: 9.34 MB
• Stars: 38
• Watchers: 1
• Forks: 15
• Open Issues: 3
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Facial Emotion Recognition using CNN :

[![LICENCE.md](https://img.shields.io/github/license/PrudhviGNV/py-automl)](https://github.com/PrudhviGNV/py-automl/blob/master/LICENCE.md)[![Ask Me Anything !](https://img.shields.io/badge/Ask%20me-anything-1abc9c.svg)](https://GitHub.com/PrudhviGNV)[![Open Source Love svg2](https://badges.frapsoft.com/os/v2/open-source.svg?v=103)](https://github.com/PrudhviGNV/open-source-badges/)

[![Awesome Badges](https://img.shields.io/badge/badges-awesome-green.svg)](https://github.com/PrudhviGNV/badges)


This is completely a deep learning project entirely based on neural networks and I think Facial emotion recognition(FER) project is one of the classical projects in deep learning.



We all know facial emotions play a vital role in our day-to-day life. so, we need a system which is capable of recognising our  facial emotions and able to act accordingly.





 

This project is the first step in that system.







![me emotion](https://user-images.githubusercontent.com/39909903/86574626-f8af6f80-bf2a-11ea-93f2-c58bd08bb871.PNG) --

![ntr emotion](https://user-images.githubusercontent.com/39909903/86574777-2f858580-bf2b-11ea-8aee-597f89f2b6b9.PNG)


 


Here I trained the convolution neural network with kaggle facial emotion dataset. so that it learns patterns for each facial expression and able to detect facial emotions







-------

### I wrote a medium blog on this project. you can find [here](https://medium.com/@prudhvi.gnv/ultimate-guide-for-facial-emotion-recognition-using-a-cnn-f9239fdc63ad?source=---------2------------------) for quick understanding and insights.

-----

_Feel free to download, clone , fork the project._  

## Installation of python libraries:

  * keras with tensorflow as backend

  * OpenCV

  * numpy

  * pandas

  * matplotlib

## DataSet:

In this project I used kaggle dataset i.e  


  https://www.kaggle.com/c/challenges-in-representation-learning-facial-expression-recognition-challenge/overview

  


    we can use a kaggle API to access and download that dataset.

## Usage:

This project is in notebook format. I wrote the program in google colabs.





**Google colabs** is a cloud based notebook IDE which offers a free GPU and allows us to write python code snippets with ease.

It is almost similar to jupyter notebook but slight variations.

Some of the code snippets in the program is specific to google colabs. Dont get confuse.







In the notebook, I  explain all the concepts, pipeline , terminology , functions which is involved in the program in a clear and concised manner. 

Even I also mention some of my insights , intuitions and observations in the project.


I document almost everything  that is used in the  program.


This notebook file is a complete wrapout  of practical exposure, understanding and theory.







  * facial_Emotion_Recognition_using_cnn.ipnb -->  main file

  * images folder --> consists of images to be tested and also contains screenshots of results of that images.

  * model.h5 --> saved weights of the model which is trained with 200 epochs

  * model.json --> saved architecture of the neural network.

  * haarcascade classifiers --> higly pretrained classifiers to detect faces.







The following is the results of the images. I also add the number of persons in the image and the relevant emojis to their emotions of the persons to create sense of feel.





![multistar_emotion](https://user-images.githubusercontent.com/39909903/86573388-33180d00-bf29-11ea-844a-09c8cea77467.PNG)

![mb_emotion](https://user-images.githubusercontent.com/39909903/86573436-41febf80-bf29-11ea-9f47-76a62f6a6a8e.PNG)

![khans_emotions](https://user-images.githubusercontent.com/39909903/86573368-2c899580-bf29-11ea-8588-3600004d7b77.PNG)

  


  


  Thanks, Hope you enjoy the project! :)

  


  


  


  

  ## Contributing:

  all pull requests are welcome. I felt glad if you give me any suggestions. (^_^)

 [![GitHub issues](https://img.shields.io/github/issues/PrudhviGNV/Facial-emotion-recognition-using-CNN)](https://github.com/PrudhviGNV/Facial-emotion-recognition-using-CNN/issues)

  


  


  ## Licence & copyright: 

  © Prudhvi GNV, _Computer Science


  Licensed under the [MIT License](LICENSE)_

  [![GitHub license](https://img.shields.io/github/license/PrudhviGNV/Facial-emotion-recognition-using-CNN)](https://github.com/PrudhviGNV/Facial-emotion-recognition-using-CNN/blob/master/LICENSE)

  

  --------

## Contact: