Step I: Data Augmentation

from tensorflow.keras.preprocessing.image import ImageDataGenerator
import matplotlib.pyplot as plt
import numpy as np
import os
from tensorflow.keras.preprocessing.image import load_img, img_to_array, array_to_img

# Load the single image
img_path = 'path_to_your_image.jpg'
img = load_img(img_path)
x = img_to_array(img)
x = x.reshape((1,) + x.shape)

# Define the augmentation configuration
datagen = ImageDataGenerator(
    rotation_range=40,
    width_shift_range=0.2,
    height_shift_range=0.2,
    shear_range=0.2,
    zoom_range=0.2,
    horizontal_flip=True,
    fill_mode='nearest')

# Generate batches of augmented images and save them
save_to_dir = 'path_to_save_augmented_images'
if not os.path.exists(save_to_dir):
    os.makedirs(save_to_dir)

i = 0
for batch in datagen.flow(x, batch_size=1, save_to_dir=save_to_dir, save_prefix='aug', save_format='jpeg'):
    i += 1
    if i > 20:
        break  # Generate 20 augmented images

# Visualize the augmented images
for file in os.listdir(save_to_dir):
    img = load_img(os.path.join(save_to_dir, file))
    plt.imshow(img)
    plt.show()