import os
import numpy as np
from keras.preprocessing import image
from keras.applications.resnet50 import preprocess_input
from tqdm import tqdm
class DataSet:
def __init__(self):
self.features = []
self.model = []
def extract_feature(self, img_path):
img = image.load_img(img_path, target_size=(331, 331))
x = image.img_to_array(img)
x = np.expand_dims(x, axis=0)
x = preprocess_input(x)
self.features = self.model.predict(x)
def save(self, save_path='test.txt'):
np.savetxt(save_path, self.features, delimiter=',')
def load(self, load_path='test.txt'):
self.features = np.loadtxt(load_path, delimiter=',')
def iterate_folder(self, main_foldername='test_images', save_main_foldername='extracted_features'):
foldernames = [f for f in os.listdir(main_foldername) if
f.isnumeric() and not f.startswith('.')]
for foldername in tqdm(foldernames):
filenames = sorted(os.listdir(main_foldername + '/' +foldername))
for filename in filenames:
img_path = main_foldername + '/' + foldername + '/' + filename
self.extract_feature(img_path)
if not os.path.exists(save_main_foldername + '/' + foldername):
os.makedirs(save_main_foldername + '/' + foldername)
self.save(save_path=save_main_foldername + '/' + foldername + '/' + filename[:-4]+'.txt')
