what-the-face-classification 0.0.27

CNN image classification trained on FER2013 for the 7 emotion categories.

This package was created as a part of Talentsprint and Techwise program by James Rodgers of team "What The Face". It uses a pytorch CNN network trained on FER2013 and CK+ images to classify an image passed into it.

Initialize class and pass a jpg PIL image to the predict() method.

classes: Image_classifier() returns self

functions: predict(image) returns [classified_name: str, probability: double]

    Exceptions for predict:
        If the face detection fails and the probability of a category is low it will return:
            returns ['no face detected', '0'] 

Example:

from PIL import Image from what_the_face_classification.inference import Image_classifier

classifier = Image_classifier() image = Image.open(test.jpg)

prediction = classifier(image)

print(prediction)