Image Augmentation for Increasing sample of dataset

Augmentation.py

import tensorflow 
from PIL import Image
import glob
from tensorflow.keras.preprocessing.image import ImageDataGenerator
from tensorflow.keras.preprocessing.image import img_to_array
from tensorflow.keras.preprocessing.image import load_img
import numpy as np

aug = ImageDataGenerator(
	rotation_range=30,
	zoom_range=0.15,
	width_shift_range=0.2,
	height_shift_range=0.2,
	shear_range=0.15,
	horizontal_flip=True,
	fill_mode="nearest")

train_aug = os.path.join('/content/drive/My Drive/Colab Notebooks/Dataset/', 'training')
valid_aug = os.path.join('/content/drive/My Drive/Colab Notebooks/Dataset/', 'validation')

train_cln_aug = [x[1] for x in os.walk(train_aug)]
train_cln_aug = train_cln_aug[0]

valid_cln_aug = [x[1] for x in os.walk(valid_dir)]
valid_cln_aug = valid_cln_aug[0]

 
for tcg in train_cln_aug:
  
  save_to_dir = '/content/drive/My Drive/Colab Notebooks/Dataset/training/' + tcg
  filename =  save_to_dir +'/*.jpg'
  im=Image.open(filename)
  image_list.append(im)

  image = image_list[0]
  image = img_to_array(image)
  image = np.expand_dims(image, axis=0)
  imageGen = aug.flow(image, batch_size=1, save_to_dir=save_to_dir ,save_prefix="image", save_format="jpg")
  
  # generating 10 sample for each training image
  total = 0
  final = 10
  for image in imageGen:
    
    total += 1
    if total ==final :
      break

for vcg in train_cln_aug:
  
  save_to_dir = '/content/drive/My Drive/Colab Notebooks/Dataset/validation/' + vcg
  filename =  save_to_dir +'/*.jpg'
  im=Image.open(filename)
  image_list.append(im)

  image = image_list[0]
  image = img_to_array(image)
  image = np.expand_dims(image, axis=0)
  imageGen = aug.flow(image, batch_size=1, save_to_dir=save_to_dir ,save_prefix="image", save_format="jpg")
  
  # generating  2 sample for each validation image
  total = 0
  final = 2
  for image in imageGen:
    total += 1
    if total ==final :
      break

inplaceAugmentation.py

import tensorflow 
from tensorflow.keras.preprocessing.image import ImageDataGenerator

# Create training ImageDataGenerator object
train_data_gen = ImageDataGenerator(rotation_range=50,
                                    width_shift_range=0.2,
                                    height_shift_range=0.2,
                                    zoom_range=0.3,
                                    horizontal_flip=True,
                                    vertical_flip=True,
                                    fill_mode='constant',
                                    cval=0,
                                    rescale=1./255)
# Create validation ImageDataGenerator objects				    
valid_data_gen = ImageDataGenerator(rotation_range=45,
                                    width_shift_range=0.2,
                                    height_shift_range=0.2,
                                    zoom_range=0.3,
                                    horizontal_flip=True,
                                    vertical_flip=True,
                                    fill_mode='constant',
                                    cval=0,
                                    rescale=1./255)
test_data_gen = ImageDataGenerator(rescale=1./255)

SEED = 1234
tensorflow.random.set_seed(SEED)
# Training
training_dir = os.path.join(dataset_dir, 'training')
train_gen = train_data_gen.flow_from_directory(training_dir,
                                               target_size=(256, 256),
                                               batch_size=Batch_size,
                                               classes=classes,
                                               class_mode='categorical',
                                               shuffle=True,
                                               seed=SEED)  # targets are directly converted into one-hot vectors

# Validation
valid_dir = os.path.join(dataset_dir, 'valid')
valid_gen = valid_data_gen.flow_from_directory(valid_dir,
                                           target_size=(256, 256),
                                           batch_size=Batch_size, 
                                           classes=classes,
                                           class_mode='categorical',
                                           shuffle=False,
                                           seed=SEED)
# Test
test_dir = os.path.join(dataset_dir, 'testing')
test_gen = test_data_gen.flow_from_directory(test_dir,
                                             target_size=(256, 256),
                                             batch_size=10, 
                                             shuffle=False,
                                             seed=SEED,
                                             class_mode=None,
                                             )