CelebA datasets Identities or classes

Question

I want to use CelebA dataset on face recognition project, but when I downloaded the dataset I found that all images are in one folder without any identities or sub folder assign to classes. Also, the downloaded information files have no any idea about the classes or identities! Any information about the 10,000 identities that mentioned on its paper and websites.

Answer 1

import os
import numpy as np
import cv2
import shutil

def read_identity(identities_filename):
    identities = []
    with open(identities_filename, 'r') as f:
        for line in f.readlines()[1:]:
            pair = line.strip().split()
            identities.append(pair)
            #print(identities)
    return np.array(identities)

def create_identity(celebA_dir,identity_path,output_dir):

    img_identity = read_identity(os.path.expanduser(identity_path))
    #print(img_identity[0][1])
    i=0
    for filename in os.listdir(celebA_dir):

        if filename == img_identity[i][0]:
            # print("Good ", filename ,img_identity)
            # # ----create the sub folder in the output folder
            # save_dir = os.path.join(output_dir,img_identity[i][1])
            # i += 1
            # if not os.path.exists(save_dir):
            #     os.makedirs(save_dir)

            # # ----copy image source to its identity destination
            destination = os.path.join(output_dir, img_identity[i][1])
            source = os.path.join(celebA_dir, img_identity[i][0])
            i += 1
            shutil.copy2(source, destination)







if __name__ == '__main__':

    celebA_dir = r"...." #Add your directory where dataset exist
    output_dir = r"...." #Add directory where you want to save datasets with identities
    identity_path = r"identity_CelebA.txt"

create_identity(celebA_dir,identity_path,output_dir)