johnpaulada/emotion-detection-docker

Emotion detection from (almost) scratch in Python. This repo contains both the Dockerfile and the app code.

Emotion Detection in Python

Emotion detection from (almost) scratch in Python.

Install

With Docker

1. Install Docker
2. Run git clone https://github.com/johnpaulada/emotion-detection-docker.git in your console
3. Run cd emotion-detection-docker in your console
4. Run docker pull johnpaulada/emotion-detection or docker load -i emotion-detection.tar in your console
5. Run docker run -v 'pwd'/app:/home/app -it johnpaulada/emotion-detection in your console
6. Run cd app
7. You're all set!

Usage

For help, you can run python3 app.py -h.

Extracting Data

To extract data, run python3 app.py --data path-to-images-folder.

For example:

python3 app.py --data ./images

Training

To train the program, run python3 app.py --train.

Prediction

To predict the emotions of input images, run python3 app.py --predict path-to-input-image. The command supports the prediction of multiple images.

For example:

python3 app.py --predict ./happy-person.jpg ./angry_face.jpg

Stack

1. Python
2. Numpy
3. OpenCV
4. dlib
5. imutils
6. Scikit-Learn
7. Docker

License

MIT