Skip to content

A deep convolutional neural network system for live emotion detection

License

Notifications You must be signed in to change notification settings

Zebreu/ConvolutionalEmotion

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

12 Commits
 
 
 
 
 
 
 
 

Repository files navigation

UPDATE

Note to anyone who'd like to try this project: DeCAF is not an active project anymore, but [Caffe] (http://caffe.berkeleyvision.org/) is and it's quick and easy to replace the DeCAF lines in my code for Caffe. I really suggest you to try it. Another project of mine uses it and you can take a look there: https://github.com/Zebreu/SketchingAI

ConvolutionalEmotion

A deep convolutional neural network system for live emotion detection

How to run:

Python 2.7 64 bit and DeCAF (https://github.com/UCB-ICSI-Vision-Group/decaf-release/) needs to be installed, as well as numpy, opencv, pygame (only if you want to play the simple game that changes based on your facial expression) and sklearn. Works on Linux (tested on Ubuntu, with Anaconda).

Download Cohn-Kanade+ dataset. Not every sequence is annotated, so you can call remove() from emotionclassification.py to delete sequences without emotion labels.

Execute emotionclassification.py to produce features files from the dataset and to save a classifier. You might want to change the path to the Haar cascade files from OpenCV (the scripts assume they're in the same folder, so you can just copy them in there too).

Execute zengame.py.

Paper about the system available on arXiv.org: http://arxiv.org/pdf/1408.3750v1.pdf

About

A deep convolutional neural network system for live emotion detection

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages