Emotion recognition on collected Google images

This project uses images collected from image.google.com to train an emotion classifier.

Current state:

It's still under development, and the test accuracy is 52%.

Data collection: search & batch download on image.google.com.
Data processing(process_original_images.py): convert profile images to face images using OpenCV.
Network building: currently I use an extended LeNet-5.
Training: AdamOptimization or AMSGrad.
Predict: classify the input image that contains at least one human face.

# Will train the network first if a pretrained one doesn't exist.
$ python emotion_recognition.py katy_perry.jpg

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
original_data		original_data
processed_data		processed_data
saves_and_logs		saves_and_logs
.gitignore		.gitignore
AMSGrad.py		AMSGrad.py
README.md		README.md
data_loader.py		data_loader.py
emotion_recognition.py		emotion_recognition.py
haarcascade_eye.xml		haarcascade_eye.xml
haarcascade_frontalface_default.xml		haarcascade_frontalface_default.xml
katy_perry.jpg		katy_perry.jpg
log_utils.py		log_utils.py
network_builder.py		network_builder.py
process_original_images.py		process_original_images.py
quick_summary.txt		quick_summary.txt
real_world_test.png		real_world_test.png
th (1).jpg		th (1).jpg
th.jpg		th.jpg