Making Training Dataset from Images

Help
#1

Hey
I have a folder of unlabeled colored images and i am building a CNN.
Now the problem is that how do I make a Training Dataset from that folder with images converted to numpy arrays and I have labels of those pictures in a Pandas Series.

#2

Hi @talal
try this if this helps or you can try modifying things here. Also please feel free to use google, stackoverflow, etc. You will learn a lot from these things or doing research by yourself.

labels_csv = pd.read_csv('lablels csv path')
filenames = ['file path where images are located' + fname + '.jpg' for fname in labels_csv['id']]

You can display an image using the given code:
from IPython.display import Image
Image(‘image location’)

Thanks!

1 Like
#3

Thankyou so much @manish_kc_06!!