tutorial53_adversarialExampleUsingFSGM.py

# https://www.tensorflow.org/tutorials/generative/adversarial_fgsm

import tensorflow as tf
import matplotlib as mpl
import matplotlib.pyplot as plt

mpl.rcParams['figure.figsize'] = (8, 8)
mpl.rcParams['axes.grid'] = False

# 사전훈련된 MobileNetV2 모델과 ImageNet 의 클래스 이름들을 불러옴

pretrained_model = tf.keras.applications.MobileNetV2(include_top = True, weights = 'imagenet')
pretrained_model.trainable = False

# ImageNet 클래스 레이블
decode_predictions = tf.keras.applications.mobilenet_v2.decode_predictions

# 이미지가 MobileNetV2 에 전달될 수 있도록 전처리해주는 헬퍼 메서드
def preprocess(image):
    image = tf.cast(image, tf.float32)
    image = image / 255
    image = tf.image.resize(image, (224, 224))
    image = image[None, ...]
    return image

# 확률 벡터에서 레이블을 추출해주는 헬퍼 메서드
def get_imagenet_label(probs):
    return decode_predictions(probs, top = 1)[0][0]

# 원본 이미지
image_path = tf.keras.utils.get_file('YellowLabradorLooking_new.jpg', 
        'https://storage.googleapis.com/download.tensorflow.org/example_images/YellowLabradorLooking_new.jpg')
image_raw = tf.io.read_file(image_path)
image = tf.image.decode_image(image_raw)

image = preprocess(image)
image_probs = pretrained_model.predict(image)

plt.figure()
plt.imshow(image[0])
_, image_class, class_confidence = get_imagenet_label(image_probs)
plt.title('{} : {:.2f}% Confidence'.format(image_class, class_confidence * 100))
plt.show()

# 적대적 이미지 생성하기
# FGSM 실행
loss_object = tf.keras.losses.CategoricalCrossentropy()

def create_adversarial_pattern(input_image, input_label):
    with tf.GradientTape() as tape:
        tape.watch(input_image)
        prediction = pretrained_model(input_image)
        loss = loss_object(input_label, prediction)
        
    # 입력 이미지에 대한 손실 함수의 기울기를 구함
    gradient = tape.gradient(loss, input_image)
    # 왜곡을 생서하기 위해 그래디언트의 부호를 구함
    signed_grad = tf.sign(gradient)
    return signed_grad

# 생성한 왜곡을 시각화
# 이미지의 레이블을 원-핫 인코딩 처리
labrador_retriever_index = 208
label = tf.one_hot(labrador_retriever_index, image_probs.shape[-1])
label = tf.reshape(label, (1, image_probs.shape[-1]))

perturbations = create_adversarial_pattern(image, label)
plt.imshow(perturbations[0])
plt.show() # added

def display_images(image, description):
    _, label, confidence = get_imagenet_label(pretrained_model.predict(image))
    plt.figure()
    plt.imshow(image[0])
    plt.title('{} \n {} : {:.2f}% Confidence'.format(description, label, confidence * 100))
    plt.show()
    
epsilons = [0, 0.01, 0.1, 0.15]
descriptions = [('Epsilon = {:0.3f}'.format(eps) if eps else 'Input') for eps in epsilons]

for i, eps in enumerate(epsilons):
    adv_x = image + eps * perturbations
    adv_x = tf.clip_by_value(adv_x, 0, 1)
    display_images(adv_x, descriptions[i])