creating_img_from_csv.py

"""
This code is simply to create jpeg images from the pixel values in each row from the FER2013 dataset and storing
them in appropriate class folders making it easier for us to train the model later on.
"""



import pandas as pd
import numpy as np
import cv2
import os
l=[1,1,1,1,1,1,1]
data=pd.read_csv('fer2013.csv')
pathname=os.path.abspath('.')+'/exp_images/'
for row in data.itertuples():
    img=np.asarray(row[2].split(' '), dtype=np.uint8)
    img=np.resize(img, (48, 48))
    
    if row[1]==0:
        cv2.imwrite(pathname+'Angry/'+str(l[0])+'.jpg', img[:, :, 1])

        l[0]=l[0]+1
    elif row[1]==1:
        cv2.imwrite(pathname+'Disgust/'+str(l[1])+'.jpg', img[:, :, 1])
        l[1]=l[1]+1
    elif row[1]==2:
        cv2.imwrite(pathname+'Fear/'+str(l[2])+'.jpg', img[:, :, 1)]
        l[2]=l[2]+1
    elif row[1]==3:
        cv2.imwrite(pathname+'Happy/'+str(l[3])+'.jpg', img[:, :, 1])
        l[3]=l[3]+1
    elif row[1]==4:
        cv2.imwrite(pathname+'Sad/'+str(l[4])+'.jpg', img[:, :, 1])
        l[4]=l[4]+1
    elif row[1]==5:
        cv2.imwrite(pathname+'Surprise/'+str(l[5])+'.jpg', img[:, :, 1])
        l[5]=l[5]+1
    elif row[1]==6:
        cv2.imwrite(pathname+'Neutral/'+str(l[6])+'.jpg', img[:, :, 1])
        l[6]=l[6]+1
    else:
        pass
        
print('images have been created and saved')