process_model.py

import cv2
import glob
import random
import numpy as np
import sys
import logging

logging.basicConfig(level=logging.DEBUG, format='%(asctime)s %(levelname)-8s %(message)s',
                    datefmt='%m-%d %H:%M',
                    filename='logs/process_model.log')
args = sys.argv
logging.debug('Fisherface training initialized')

def get_files_from_emotion(emotion):  # Define function to get file list, randomly shuffle it and split 80/20
    files = glob.glob('basis_data/dataset/%s/*' % emotion)
    random.shuffle(files)
    training = files[:int(len(files) * 0.8)]  # get first 80% of file list
    prediction = files[-int(len(files) * 0.2):]  # get last 20% of file list
    return training, prediction

def image_preprocessing(item):
    image = cv2.imread(item)  # open image
    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)  # convert to grayscale
    return gray

def make_sets():
    training_data = []
    training_labels = []
    prediction_data = []
    prediction_labels = []
    for emotion in emotions:
        training, prediction = get_files_from_emotion(emotion)
        # Append data to training and prediction list, and generate labels 0-7
        for item in training:
            img = image_preprocessing(item)
            training_data.append(img)  # append image array to training data list
            training_labels.append(emotions.index(emotion))

        for item in prediction:  # repeat above process for prediction set
            img = image_preprocessing(item)
            prediction_data.append(img)
            prediction_labels.append(emotions.index(emotion))


    return training_data, training_labels, prediction_data, prediction_labels

def run_recognizer():
    training_data, training_labels, prediction_data, prediction_labels = make_sets()

    fishface.train(training_data, np.asarray(training_labels))

    cnt = 0
    correct = 0
    incorrect = 0
    for image in prediction_data:
        pred, conf = fishface.predict(image)
        if pred == prediction_labels[cnt]:
            correct += 1
            cnt += 1
        else:
            incorrect += 1
            cnt += 1
    return ((100 * correct) / (correct + incorrect))

if len(args) > 1:
    tags = ', '.join(args[1:])
    logging.debug(tags.upper())

emotions = ["anger", "disgust", "happy", "neutral", "surprise"]  # Emotion list
fishface = cv2.face.FisherFaceRecognizer_create()

# Now run it
metascore = []

for i in range(0, 10):
    correct = run_recognizer()
    logging.debug("{} : {}%".format(i, int(correct)))
    metascore.append(correct)

logging.debug("{} iterations - {}% average\n".format(len(metascore), np.mean(metascore)))
fishface.write('basis_data/models/detection_model.xml')