Face recognition physical adversarial attack

examples/face_adversarial_attack/example/AFR.py

@@ -0,0 +1,289 @@
+# Copyright 2022 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+import os
+import re
+import numpy as np
+import matplotlib.image as mp
+import dlib
+import face_recognition as fr
+import face_recognition_models as frm
+from PIL import Image, ImageDraw
+import mindspore
+from mindspore.dataset.vision.py_transforms import ToPIL as ToPILImage
+from mindspore import Parameter,ops, nn,Tensor
+from mindspore.dataset.vision.py_transforms import ToTensor
+import mindspore.dataset.vision.py_transforms as P
+from loss_design import TrainOneStepCell,MyWithLossCell,FaceLoss_no_target_attack,FaceLoss_target_attack
+from FaceRecognition.eval import get_net
+
+class Attack(object):
+    """
+    Class used to create adversarial facial recognition attacks
+    """
+
+    def __init__(self,input_img,target_img,seed=None):
+        """
+        Initialization for Attack class.
+
+        Args:
+            input_img : Image to train on.
+            target_img : Image to target the adversarial attack against.
+            seed : optional Sets custom seed for reproducability. Default is generated randomly.
+
+        """
+
+        if (seed != None): np.random.seed(seed)
+        self.MEAN = Tensor([0.485, 0.456, 0.406])
+        self.STD = Tensor([0.229, 0.224, 0.225])
+        self.LOSS = Tensor(0)
+        self.expand_dims = mindspore.ops.ExpandDims()
+        self.imageize = ToPILImage()
+        self.tensorize = ToTensor()
+        self.normalize = P.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+        self.resnet = get_net()
+        self.input_tensor = Tensor(self.normalize(self.tensorize(input_img)))
+        self.target_tensor = Tensor(self.normalize(self.tensorize(target_img)))
+        mp.imsave('./outputs/input图像.jpg', np.transpose(self._reverse_norm(self.input_tensor).asnumpy(), (1, 2, 0)))
+        mp.imsave('./outputs/target图像.jpg', np.transpose(self._reverse_norm(self.target_tensor).asnumpy(), (1, 2, 0)))
+
+
+        self.input_emb = self.resnet(self.expand_dims(self.input_tensor,0))
+        self.target_emb = self.resnet(self.expand_dims(self.target_tensor,0))
+        self.adversarial_emb = None
+        self.mask_tensor = self._create_mask(input_img)
+        self.ref = self.mask_tensor
+        self.pm = Parameter(self.mask_tensor)
+        self.opt = nn.Adam([self.pm], learning_rate=0.01, weight_decay=0.0001)
+
+
+
+    def train(self,attack_method):
+        """
+        Optimized adversarial image.
+        """
+
+        if attack_method == "non-target attack":
+            LOSS = FaceLoss_no_target_attack(self.target_emb)
+        if attack_method == "target_attack":
+            LOSS = FaceLoss_target_attack(self.target_emb)
+
+        net_with_criterion = MyWithLossCell(self.resnet, LOSS,self.input_tensor)
+        train_net = TrainOneStepCell(net_with_criterion, self.opt)
+
+        for i in range(2000):
+
+            self.mask_tensor = Tensor(self.pm)
+
+            grads,loss = train_net(self.mask_tensor)
+
+            print("epoch %d ,loss: %f \n " % (i, loss.asnumpy().item()))
+
+            self.mask_tensor = ops.clip_by_value(self.mask_tensor, Tensor(0, mindspore.float32), Tensor(1, mindspore.float32))
+
+        adversarial_tensor = self._apply(
+            self.input_tensor,
+            (self.mask_tensor - self.MEAN[:, None, None]) / self.STD[:, None, None],
+            self.ref)
+
+        adversarial_tensor = self._reverse_norm(adversarial_tensor)
+
+        return adversarial_tensor, self.mask_tensor
+
+    def test(self):
+        """
+        Test the recognition of adversarial images by the model.
+        """
+
+        adversarial_tensor = self._apply(
+            self.input_tensor,
+            (self.mask_tensor - self.MEAN[:, None, None] )/ self.STD[:, None, None],
+            self.ref)
+
+        self.adversarial_emb = self.resnet(self.expand_dims(adversarial_tensor, 0))
+        self.input_emb = self.resnet(self.expand_dims(self.input_tensor, 0))
+        self.target_emb = self.resnet(self.expand_dims(self.target_tensor, 0))
+
+        adversarial = np.argmax(self.adversarial_emb.asnumpy())
+        target = np.argmax(self.target_emb.asnumpy())
+        input = np.argmax(self.input_emb.asnumpy())
+
+        print("input:", input)
+        print("input_confidence:", self.input_emb.asnumpy()[0][input])
+        print("================================")
+        print("adversarial:", adversarial)
+        print("adversarial_confidence:", self.adversarial_emb.asnumpy()[0][adversarial])
+        print("Confidence changes for target::", self.adversarial_emb.asnumpy()[0][target])
+        print("Confidence changes for input:", self.adversarial_emb.asnumpy()[0][input])
+        print("================================")
+        print("target:", target)
+        print("target_confidence:", self.target_emb.asnumpy()[0][target])
+        print("input:%d, target:%d, adversarial:%d" % (input,target, adversarial))
+
+    def test1(self,adversarial_tensor):
+        self.adversarial_emb = self.resnet(self.expand_dims((adversarial_tensor - self.MEAN[:, None, None]) / self.STD[:, None, None], 0))
+        self.input_emb = self.resnet(self.expand_dims(self.input_tensor, 0))
+        self.target_emb = self.resnet(self.expand_dims(self.target_tensor, 0))
+
+        adversarial = np.argmax(self.adversarial_emb.asnumpy())
+        target = np.argmax(self.target_emb.asnumpy())
+        input = np.argmax(self.input_emb.asnumpy())
+
+        print("input:", input)
+        print("input_confidence:", self.input_emb.asnumpy()[0][input])
+        print("================================")
+        print("adversarial:", adversarial)
+        print("adversarial_confidence:", self.adversarial_emb.asnumpy()[0][adversarial])
+        print("Confidence changes for input:", self.adversarial_emb.asnumpy()[0][input])
+        print("================================")
+        print("target:", target)
+        print("target_confidence:", self.target_emb.asnumpy()[0][target])
+        print("input:%d, target:%d, adversarial:%d" % (input, target, adversarial))
+
+
+
+    def _reverse_norm(self, image_tensor):
+        """
+        Reverses normalization for a given image_tensor
+
+        Args:
+            image_tensor : Tensor
+
+        Returns:
+            Tensor
+        """
+        tensor = image_tensor * self.STD[:, None, None] + self.MEAN[:, None, None]
+
+        return tensor
+
+    def _apply(self,
+               image_tensor,
+               mask_tensor,
+               reference_tensor
+               ):
+        """
+        Apply a mask over an image.
+
+        Args:
+            image_tensor : Canvas to be used to apply mask on.
+            mask_tensor : Mask to apply over the image.
+            reference_tensor : Used to reference mask boundaries
+
+        Returns:
+            Tensor
+        """
+        tensor = mindspore.numpy.where((reference_tensor == 0), image_tensor, mask_tensor)
+
+        return tensor
+
+    def _create_mask(self, face_image):
+        """
+        Create mask image.
+
+        Args:
+            face_image : image of a detected face.
+
+        Returns:
+            mask_tensor : A mask image.
+        """
+
+        mask = Image.new('RGB', face_image.size, color=(0, 0, 0))
+        d = ImageDraw.Draw(mask)
+        landmarks = fr.face_landmarks(np.array(face_image))
+        area = [landmark
+                for landmark in landmarks[0]['chin']
+                if landmark[1] > max(landmarks[0]['nose_tip'])[1]]
+        area.append(landmarks[0]['nose_bridge'][1])
+        d.polygon(area, fill=(255, 255, 255))
+        mask_array = np.array(mask)
+        mask_array = mask_array.astype(np.float32)
+
+        for i in range(mask_array.shape[0]):
+
+            for j in range(mask_array.shape[1]):
+
+                for k in range(mask_array.shape[2]):
+
+                    if mask_array[i][j][k] == 255.:
+                        mask_array[i][j][k] = 0.5
+                    else:
+                        mask_array[i][j][k] = 0
+
+        mask_tensor = Tensor(mask_array)
+        mask_tensor = mask_tensor.swapaxes(0, 2).swapaxes(1, 2)
+        mask_tensor.requires_grad = True
+        return mask_tensor
+
+    def _reverse_norm(self, image_tensor):
+        """
+        Reverses normalization for a given image_tensor.
+
+        Args:
+            image_tensor : Tensor.
+
+        Returns:
+            Tensor.
+        """
+        tensor = image_tensor * self.STD[:, None, None] + self.MEAN[:, None, None]
+        return tensor
+
+
+def detect_face(image_loc):
+    """
+    Helper function to run the facial detection and alignment process using
+    dlib.  Detects a given face and aligns it using dlib's 5 point landmark
+    detector.
+
+     Args:
+        image_loc : image file location.
+
+     Returns:
+        face_image : Resized face image.
+
+     """
+
+    detector = dlib.get_frontal_face_detector()
+    shape_predictor = dlib.shape_predictor(frm.pose_predictor_model_location())
+    image = dlib.load_rgb_image(image_loc)
+    dets = detector(image, 1)
+
+    faces = dlib.full_object_detections()
+    for detection in dets:
+        faces.append(shape_predictor(image, detection))
+    face_image = Image.fromarray(dlib.get_face_chip(image, faces[0], size=112))
+
+    return face_image
+
+
+def load_data(path_to_data):
+    """
+    Helper function for loading image data.  Allows user to load the input, target,
+    and test images.
+
+    Args:
+        path_to_data : Path to the given data.
+
+    Returns:
+        list : List of resized face images.
+    """
+    img_files = [f for f in os.listdir(path_to_data) if re.search(r'.*\.(jpe?g|png)', f)]
+    img_files_locs = [os.path.join(path_to_data, f) for f in img_files]
+
+    image_list = []
+
+    for loc in img_files_locs:
+        image_list.append(detect_face(loc))
+
+    return image_list
+

examples/face_adversarial_attack/example/README.md

@@ -0,0 +1,86 @@
+# 人脸识别物理对抗攻击
+
+
+
+##描述
+本项目是对人脸识别模型的物理对抗攻击，通过生成对抗口罩，使人脸佩戴后实现目标攻击和非目标攻击，并应用于mindspore平台。
+
+
+
+##模型结构
+采用华为mindspore官方训练的FaceRecognition模型
+https://www.mindspore.cn/resources/hub/details?MindSpore/1.7/facerecognition_ms1mv2
+
+
+
+
+##环境要求
+mindspore=1.7，硬件平台为GPU。
+
+
+
+##脚本说明
+├── readme.md
+├── opencv_photo
+│   ├── adv_input //对抗图像
+│   ├── input //输入图像 
+│   └── target //目标图像
+├── outputs //训练后的图像
+├── FaceRecognition //模型设置
+├── AFR.py  //训练脚本
+├── camera.py //opencv图像采集
+│── example_non-target_attack.py  //无目标攻击训练
+│── example_target_attack.py  //目标攻击训练
+│── loss_design.py  //训练优化设置
+└── test.py  //评估攻击效果
+
+
+##模型调用
+方法一:
+
+    #基于mindspore_hub库调用FaceRecognition模型
+
+    import mindspore_hub as mshub
+    from mindspore import context
+    
+    def get_net():
+        context.set_context(mode=context.GRAPH_MODE,
+                            device_target="GPU",
+                            device_id=0)
+        
+        model = "mindspore/1.7/facerecognition_ms1mv2"
+        network = mshub.load(model)
+        network.set_train(False)
+        return network
+
+方法二:
+
+    利用 MindSpore代码仓中的https://gitee.com/mindspore/models/blob/master/research/cv/FaceRecognition/eval.py的get_model函数加载模型
+
+##训练过程
+目标攻击:
+    
+    $ cd face_adversarial_attack/example/
+    $ python example_target_attack.py
+
+非目标攻击:
+    
+    $ cd face_adversarial_attack/example/
+    $ python example_non-target_attack.py
+
+
+
+
+##默认训练参数
+optimizer=adam, learning rate=0.01, weight_decay=0.0001, epoch=2000
+
+
+##评估过程
+评估方法一:
+   
+    AFR.Attack.test()
+
+评估方法二:
+    
+    $ cd face_adversarial_attack/example/
+    $ python test.py

examples/face_adversarial_attack/example/camera.py

@@ -0,0 +1,34 @@
+# Copyright 2022 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+import cv2
+
+
+capture = cv2.VideoCapture(0)
+
+
+for i in range(2):
+    while(True):
+        ret, frame = capture.read()
+        width, height = capture.get(3), capture.get(4)
+        cv2.imwrite('./opencv_photo/input/input'+str(i)+'.png', frame)
+        cv2.imshow('frame', frame)
+        if cv2.waitKey(1) == ord('q'):
+            break
+
+
+print(width, height)
+
+capture.release()
+cv2.destroyAllWindows()

examples/face_adversarial_attack/example/example_non-target_attack.py

@@ -0,0 +1,40 @@
+# Copyright 2022 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+
+import numpy as np
+import matplotlib.image as mp
+from mindspore import context
+import AFR
+
+context.set_context(mode=context.GRAPH_MODE, device_target="GPU")
+
+
+
+
+if __name__ == '__main__':
+
+    inputs = AFR.load_data('opencv_photo/input/')
+    targets = AFR.load_data('opencv_photo/target/')
+
+
+    adversarial = AFR.Attack(inputs[0], targets[0])
+    attack_method = "non-target attack"
+
+    adversarial_tensor, mask_tensor  = adversarial.train(attack_method)
+
+    mp.imsave('./outputs/对抗图像.jpg', np.transpose(adversarial_tensor.asnumpy(), (1, 2, 0)))
+    mp.imsave('./outputs/口罩.jpg', np.transpose(mask_tensor.asnumpy(), (1, 2, 0)))
+
+    adversarial.test()

examples/face_adversarial_attack/example/example_target_attack.py

@@ -0,0 +1,41 @@
+# Copyright 2022 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+
+import numpy as np
+import matplotlib.image as mp
+from mindspore import context
+import AFR
+
+context.set_context(mode=context.GRAPH_MODE, device_target="GPU")
+
+
+
+
+if __name__ == '__main__':
+
+    inputs = AFR.load_data('opencv_photo/input/')
+    targets = AFR.load_data('opencv_photo/target/')
+
+
+    adversarial = AFR.Attack(inputs[0], targets[0])
+
+    attack_method = "target_attack"
+
+    adversarial_tensor, mask_tensor  = adversarial.train(attack_method)
+
+    mp.imsave('./outputs/对抗图像.jpg', np.transpose(adversarial_tensor.asnumpy(), (1, 2, 0)))
+    mp.imsave('./outputs/口罩.jpg', np.transpose(mask_tensor.asnumpy(), (1, 2, 0)))
+
+    adversarial.test()

examples/face_adversarial_attack/example/loss_design.py

@@ -0,0 +1,160 @@
+# Copyright 2022 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+import mindspore
+from mindspore import ops, nn,Tensor
+from mindspore.dataset.vision.py_transforms import ToTensor
+import mindspore.dataset.vision.py_transforms as P
+
+
+
+
+class TrainOneStepCell(nn.Cell):
+    """
+    Encapsulation class of network training.
+
+    Append an optimizer to the training network after that the construct
+    function can be called to create the backward graph.
+
+    Args:
+        network (Cell): The training network. Note that loss function should have been added.
+        optimizer (Optimizer): Optimizer for updating the weights.
+        sens (Number): The adjust parameter. Default: 1.0.
+    """
+
+    def __init__(self, network, optimizer, sens=1.0):
+        super(TrainOneStepCell, self).__init__(auto_prefix=False)
+        self.network = network
+        self.network.set_grad()
+        self.optimizer = optimizer
+        self.weights = self.optimizer.parameters
+        self.grad = ops.composite.GradOperation(get_all=True, sens_param=False)
+
+    def construct(self,*inputs):
+        """Defines the computation performed."""
+        loss = self.network(*inputs)
+        grads = self.grad(self.network)(*inputs)
+        self.optimizer(grads)
+        return grads,loss
+
+
+
+class MyWithLossCell(nn.Cell):
+   def __init__(self,net,loss_fn,input_tensor):
+       super(MyWithLossCell, self).__init__(auto_prefix=False)
+       self.net = net
+       self._loss_fn = loss_fn
+       self.STD = Tensor([0.229, 0.224, 0.225])
+       self.MEAN = Tensor([0.485, 0.456, 0.406])
+       self.expand_dims = mindspore.ops.ExpandDims()
+       self.normalize = P.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+       self.tensorize = ToTensor()
+       self.input_tensor = input_tensor
+       self.input_emb = self.net(self.expand_dims(self.input_tensor, 0))
+
+   def construct(self,mask_tensor):
+       ref = mask_tensor
+       adversarial_tensor = mindspore.numpy.where((ref == 0), self.input_tensor, (mask_tensor - self.MEAN[:, None, None] )/ self.STD[:, None, None])
+       adversarial_emb =  self.net(self.expand_dims(adversarial_tensor, 0))
+       loss = self._loss_fn( adversarial_emb,self.input_emb,mask_tensor)
+       return loss
+
+   @property
+   def backbone_network(self):
+       return self.net
+
+
+class FaceLoss_target_attack(nn.Cell):
+    """The loss function of the target attack"""
+
+    def __init__(self,target_emb):
+        super(FaceLoss_target_attack, self).__init__()
+        self.uniformreal = ops.UniformReal(seed=2)
+        self.sum = ops.ReduceSum(keep_dims=False)
+        self.norm = nn.Norm(keep_dims=True)
+        self.zeroslike = ops.ZerosLike()
+        self.concat_op1 = ops.Concat(1)
+        self.concat_op2 = ops.Concat(2)
+        self.pow = ops.Pow()
+        self.reduce_sum = ops.operations.ReduceSum()
+        self.target_emb = target_emb
+        self.abs = ops.Abs()
+        self.reduce_mean = ops.ReduceMean()
+
+    def construct(self, adversarial_emb,input_emb,mask_tensor):
+        #像素平滑
+        # vert_diff = mask_tensor[:, 1:] - mask_tensor[:, :-1]
+        # hor_diff = mask_tensor[:, :, 1:] - mask_tensor[:, :, :-1]
+        # vert_diff_sq = self.pow(vert_diff, 2)
+        # hor_diff_sq = self.pow(hor_diff, 2)
+        # A = self.zeroslike(Tensor(self.uniformreal((3, 1, 112))))
+        # B = self.zeroslike(Tensor(self.uniformreal((3, 112, 1))))
+        # vert_pad = self.concat_op1((vert_diff_sq, A))
+        # hor_pad = self.concat_op2((hor_diff_sq, B))
+        # tv_sum = vert_pad + hor_pad
+        # tv = ops.functional.sqrt(tv_sum + 1e-5)
+        # tv_final_sum = self.sum(tv)
+        # tv_loss = (1e-4) * tv_final_sum
+        # print("tv_loss:",tv_loss)
+        prod_sum = self.reduce_sum(adversarial_emb * self.target_emb, (1,))
+        square1 = self.reduce_sum(ops.functional.square(adversarial_emb), (1,))
+        square2 = self.reduce_sum(ops.functional.square(self.target_emb), (1,))
+        denom = ops.functional.sqrt(square1) * ops.functional.sqrt(square2)
+        loss = -(prod_sum / denom)
+        print("dis_loss:", loss)
+        return loss
+
+
+class FaceLoss_no_target_attack(nn.Cell):
+    """The loss function of the non-target attack"""
+    def __init__(self, target_emb):
+        """初始化"""
+        super(FaceLoss_no_target_attack, self).__init__()
+        self.uniformreal = ops.UniformReal(seed=2)
+        self.sum = ops.ReduceSum(keep_dims=False)
+        self.norm = nn.Norm(keep_dims=True)
+        self.zeroslike = ops.ZerosLike()
+        self.concat_op1 = ops.Concat(1)
+        self.concat_op2 = ops.Concat(2)
+        self.pow = ops.Pow()
+        self.reduce_sum = ops.operations.ReduceSum()
+        self.target_emb = target_emb
+        self.abs = ops.Abs()
+        self.reduce_mean = ops.ReduceMean()
+
+    def construct(self, adversarial_emb, input_emb, mask_tensor):
+        # 像素平滑
+        # vert_diff = mask_tensor[:, 1:] - mask_tensor[:, :-1]  #
+        # hor_diff = mask_tensor[:, :, 1:] - mask_tensor[:, :, :-1]
+        # vert_diff_sq = self.pow(vert_diff, 2)
+        # hor_diff_sq = self.pow(hor_diff, 2)
+        # A = self.zeroslike(Tensor(self.uniformreal((3, 1, 112))))  #
+        # B = self.zeroslike(Tensor(self.uniformreal((3, 112, 1))))
+        # vert_pad = self.concat_op1((vert_diff_sq, A))
+        # hor_pad = self.concat_op2((hor_diff_sq, B))
+        # tv_sum = vert_pad + hor_pad
+        # tv = ops.functional.sqrt(tv_sum + 1e-5)
+        # tv_final_sum = self.sum(tv)
+        # tv_loss = (1e-4) * tv_final_sum
+        # print("tv_loss:",tv_loss)
+
+        prod_sum = self.reduce_sum(adversarial_emb * input_emb, (1,))
+        square1 = self.reduce_sum(ops.functional.square(adversarial_emb), (1,))
+        square2 = self.reduce_sum(ops.functional.square(input_emb), (1,))
+        denom = ops.functional.sqrt(square1) * ops.functional.sqrt(square2)
+        loss = prod_sum / denom
+        print("cosine_loss:", loss)
+        return loss
+
+

examples/face_adversarial_attack/example/test.py

@@ -0,0 +1,70 @@
+# Copyright 2022 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+
+import numpy as np
+import matplotlib.image as mp
+from mindspore import context,Tensor
+import mindspore
+from mindspore.dataset.vision.py_transforms import ToTensor
+import mindspore.dataset.vision.py_transforms as P
+from mindspore.dataset.vision.py_transforms import ToPIL as ToPILImage
+from FaceRecognition.eval import get_net
+import AFR
+
+context.set_context(mode=context.GRAPH_MODE, device_target="GPU")
+
+
+imageize = ToPILImage()
+
+if __name__ == '__main__':
+    """
+    The input image, target image and adversarial image are tested using the FaceRecognition model.
+    """
+
+    image = AFR.load_data('opencv_photo/adv_input')
+    inputs = AFR.load_data('opencv_photo/input/')
+    targets = AFR.load_data('opencv_photo/target/')
+
+    tensorize = ToTensor()
+    normalize = P.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+    expand_dims = mindspore.ops.ExpandDims()
+    MEAN = Tensor([0.485, 0.456, 0.406])
+    STD = Tensor([0.229, 0.224, 0.225])
+
+    resnet = get_net()
+
+    image = mp.imread("./对抗图像.jpg")
+    adv = Tensor(normalize(tensorize(image)))
+
+    input_tensor = Tensor(normalize(tensorize(inputs[0])))
+    target_tensor = Tensor(normalize(tensorize(targets[0])))
+
+    adversarial_emb = resnet(expand_dims(adv, 0))
+    input_emb = resnet(expand_dims(input_tensor, 0))
+    target_emb = resnet(expand_dims(target_tensor, 0))
+
+    adversarial = np.argmax(adversarial_emb.asnumpy())
+    target = np.argmax(target_emb.asnumpy())
+    input = np.argmax(input_emb.asnumpy())
+
+    print("input:", input)
+    print("input_confidence:", input_emb.asnumpy()[0][input])
+    print("================================")
+    print("adversarial:", adversarial)
+    print("adversarial_confidence:", adversarial_emb.asnumpy()[0][adversarial])
+    print("Confidence changes for input:", adversarial_emb.asnumpy()[0][input])
+    print("================================")
+
+    print("input:%d, target:%d, adversarial:%d" % (input, target, adversarial))