emotion2filter

Program detects emotion and applies simple filters to face by emotion.

• Place datasets to required directories (which explained text file in dataset/ directory)

1. Run landmark-model.py to train model to detect facial landmarks
2. Run emotion-model.py to train model to detect facial emotions.
3. Run run_camera.py to see program outputs. FPS is low due to CNN performances. I neither have a GPU nor implementation of GPU

• Program detects faces with openCV's haar cascade.(Frontal faces only)

#Datasets references

DATASET	train data	test data
FACIAL LANDMARK	train: http://users.sussex.ac.uk/~is321/training_images.npz	test: http://users.sussex.ac.uk/~is321/test_images.npz
EMOTION	https://www.kaggle.com/datasets/jonathanoheix/face-expression-recognition-dataset	test also included in link

Sample outputs

video ->

##filter result of neutral face

filter result of sad face

landmark train results.