痛击肖像被恶意使用！三类deepfake检测技术总结-FlyAI

痛击肖像被恶意使用！三类deepfake检测技术总结

深度学习

最后更新 2020-12-19 15:45 阅读 138

最后更新 2020-12-19 15:45

阅读 138

深度学习

##FlyAI文献翻译

**英文原文：**[：Three Types of Deepfake Detection
](https://lionbridge.ai/articles/three-types-of-deepfake-detection/)

> **标签** ：人脸检测

![](https://static.leiphone.com/uploads/new/sns/article/202011/1606203160134934.png)
Due to advancements in machine learning, it’s now possible to create [high quality deepfakes](https://lionbridge.ai/articles/a-look-at-deepfakes-in-2020/ "high quality deepfakes") with minimal effort. So it’s no surprise that deepfakes used with malicious intent can cause great harm. For instance, the technology can be used to create large amounts of misinformation, damage reputations, and harm individuals. This makes deepfake detection more and more important.

>由于机器学习的 进展，现在用很小的功夫就炮制高质量的Deepfake图片已经变成了可能。所以不要惊讶Deepfake可能被用于某些恶意的动机。比如这项技术可能被用来制造大量的假新闻，名誉损害，以及直接伤害个人。这都使得检测Deepfake的手段变得愈加重要。

Deepfakes are images or videos that are synthetically manipulated or generated using deep learning. The term is growing in popularity, but an exact definition is difficult to pin down. It is sometimes used for “face swaps,” “expression/attribute manipulation,” and even for images entirely synthesized by a deep learning algorithm. In this article, I’ll use the term deepfake to refer to both media manipulation and media generation through deep learning.

>Deepfake是指用深度学习合成或操纵的图片或视频。这个术语变得越来越流行，但是明确的定义是很难做出的。这个词有时指“换脸“、“表情操纵”、甚至是一张完全由深度学习算法合成的图片。本文将使用Deepfake指代用深度学习对媒体进行操纵或者合成的行为。

Though it’s possible for humans to detect deepfakes, it’s getting harder as the general quality increases. Moreover, as it gets easier and faster to create high quality deepfakes, the volume of deepfake content may be too great for human detection alone. That’s why we may need to rely on algorithms to determine whether a piece of content is a deepfake or not.

>尽管人眼依然可以判断出Deepfake，但是随着总体质量的提升，这变得越来越难。而且现在想要制造高质量的Deepfake变得越来越快和容易。这就是为什么我们需要依赖算法来判断一项内容是否为 Deepfake了。

Fortunately, deepfake detection research is ongoing, and offers numerous potential benefits. Not only are algorithms automatic, they can also potentially detect cues that are hard for humans to find on their own.

>幸运的是，检测Deepfake的研究正在进行中，并且产出了几种可能的收获。这些算法不仅是自动化的，而且也可能有探测出人眼本身难以分辨的信号的潜力。

In this article, I’ll briefly explore a few techniques for detecting deepfakes. I’ll also look at the broad challenges that face deepfake detection techniques both at present and in the future.

>本文将简要地探索几个检测Deepfake的技术。我还讲广泛地讲讲现在和将来本项技术会面临的种种挑战。

###Deepfake Detection Techniques

In this section I’ll explore a few methods for detecting deepfakes. To get an idea of the various detection techniques available, I referred to DeepFakes and Beyond: A Survey of Face Manipulation and Fake Detection by Ruben Tolosana et al. Please take a look at the survey if you want to explore the techniques further.

In this article, I’ve organized deepfake detection methods into the following three broad categories:

>本节我将浅析几种检测Deepfake的方法。为了理解已有的各种检测方法的精髓我参考了Tolosana的综述文章 DeepFakes and Beyond: A Survey of Face Manipulation and Fake Detection。如果你想亲自进一步探索这些技术，请务必去看。

>本文将把各种Deepfake检测办法归到三个广泛的类当中:

- Hand-crafted Features

- Learning Based Features

- Artifacts

I have placed methods in their respective categories based on what I felt best described them, so please keep in mind that some methods could fall into multiple categories. For the sake of brevity, I’ve limited each category to a few key methods, and have kept discussion of performance to a minimum. I’ll provide links for readers who want to learn more about the contents in greater detail.

Without further ado, let’s look at our deepfake detection techniques.

>- 手工特征
- 基于学习的特征
- 图案缺陷

>我将众多办法归入最能对应的他们类型中，请记住某些方法其实属于好几类。为图简便，我已经将每种类型缩减到只有几种办法，并且尽量少地讨论它们的性能。我给想要详细了解这些内容的读者提供了链接。

>话休絮烦，我们来看看这些检测办法吧。

###Hand-crafted Features

While deepfake techniques create realistic looking content, there are often obvious flaws that a human or algorithm can spot upon closer inspection. One example is unnatural facial features. In this category, we’ll explore a couple of techniques that extract and engineer specific features to perform deepfake detection.

>虽然 deepfake 技术可以创建看上去真实的内容，但通常存在一些明显的缺陷，经过进一步检查，人类或算法可以发现这些瑕疵。一个例子就是不自然的面部特征。在这一段落中，我们将探索几种提取和设计特定特征来检测 deepfake 的技术。

[Shruti Agarwal et al.](https://openaccess.thecvf.com/content_CVPRW_2019/papers/Media%20Forensics/Agarwal_Protecting_World_Leaders_Against_Deep_Fakes_CVPRW_2019_paper.pdf "Shruti Agarwal et al.") propose a method involving a novelty detection model (here, a one-class SVM) that can distinguish a person of interest (POI) from other individuals as well as deepfake impersonators. This methodology is quite interesting as we only need authentic videos of the POI to train the model.

>Shruti Agarwal 等人提出了一种涉及新颖性检测模型（这里是一个单类支持向量机）的方法，该方法可以将感兴趣的人（POI）与其他个体以及深度假冒仿冒者区分开来。这种方法非常有趣，因为我们只需要 POI 的真实视频来训练模型。

The authors’ hypothesis is that as an individual speaks, they have distinct (but probably not unique) facial expressions and movements. The researchers use the OpenFace2 toolkit to extract facial and head movements in a given video. By collecting facial action units, head rotations about certain axes and 3D distances between certain mouth landmarks, they obtain 20 facial/head features for a given 10 second video clip.

>作者假设当一个人说话时他们有不同的面部表情和动作（但可能不是唯一的）。研究人员使用 OpenFace2 工具包提取给定视频中的面部和头部运动。通过收集面部动作单元、特定轴的头部旋转和特定口腔地标之间的 3D 距离，他们可以在给定的 10 秒视频片段中提取 20 个面部/头部特征。

![](https://static.leiphone.com/uploads/new/sns/article/202011/1606203161342195.png)

Source:https://openaccess.thecvf.com/content_CVPRW_2019/papers/Media%20Forensics/Agarwal_Protecting_World_Leaders_Against_Deep_Fakes_CVPRW_2019_paper.pdf

Finally, by computing the Pearson correlation between the above 20 facial/head features, they get a 190 dimensional feature vector representing the 10 second clip. Once the 190-D feature vector is extracted, a one-class support vector machine (SVM) is used to determine whether the 10 second video clip is a real video of the POI.

>最后，通过计算以上 20 个面部/头部特征之间的 Pearson 相关性，得到代表 10 秒片段的 190-D 特征向量。提取 190-D 特征向量后，使用单类支持向量机（SVM）判断 10 秒的视频片段是否是 POI 的真实视频。

One other example of using hand-crafted features for deepfake detection is by [Tackhyun Jung et al](https://ieeexplore.ieee.org/document/9072088 "Tackhyun Jung et al"). The authors propose a framework to analyze a person’s blinking patterns to detect deepfakes in video. They posit that blinking patterns are known to change based on physical conditions, cognitive activities, biological factors and so on. Using a collection of algorithms, the authors extract the facial region and calculate the Eye Aspect Ratio (EAR) for each frame in the video. The EAR value for a closed eye is typically smaller than the EAR value for an open eye (for the exact framework formulation, please refer to the paper). By setting an appropriate threshold, one can detect blinking events based on the EAR value and analyze a person’s blinking patterns in video.

>使用手工特征进行深度假检测的另一个例子是由 Tackhyun Jung 等人完成的。作者提出了一个框架来分析一个人的眨眼模式，以检测 deepfake 视频。他们认为，眨眼的方式会随着身体状况、认知活动、生物因素等的变化而改变。利用一组算法，作者提取了人脸区域，并计算了视频中每一帧的眼睛宽高比（EAR）。闭着眼睛的耳朵值通常比睁开眼睛的耳朵值小（确切的框架公式，请参阅论文）。通过设置适当的阈值，人们可以根据耳朵的值检测眨眼事件，并分析一个人在视频中的眨眼模式。

![](https://static.leiphone.com/uploads/new/sns/article/202011/1606203161988123.png)

As part of this framework, four attributes (gender, age, activity and time) were taken as input to describe the person in the video. Based on these attributes, the authors query a pre-configured database with typical blinking pattern data. Then they compare the measured blinking pattern data with the blinking pattern data queried from the database. Based on this they can ascertain whether a video is a deepfake.

>作为这个框架的一部分，四个属性（性别、年龄、活动和时间）作为 input 来描述视频中的人。基于这些属性，作者从一个具有典型闪烁模式数据的预配置数据库进行查询。然后他们将测量到的闪烁模式数据与数据库中查询到的闪烁模式数据进行比较。因此，他们可以确定一个视频是否是 deepfakes 的。

You can check out the [survey paper](https://arxiv.org/pdf/2001.00179.pdf "survey paper") I mentioned earlier in the text and the related work section of the[ GANprintR](https://arxiv.org/pdf/1911.05351.pdf " GANprintR") paper to learn about more hand crafted feature-based detection techniques.

>您可以查看我前面在文本中提到的调查论文和 GANprintR 论文的相关工作部分，以了解更多手工制作、基于特征的检测技术。

###Learning Based Features

In this category we’ll explore learning based methods for detecting deepfakes. Many of these methods use convolutional neural networks (CNNs) to learn the features necessary for deepfake detection.

>在此，我们将探索基于学习的方法来检测 deepfakes。这些方法中有许多使用卷积神经网络（CNNs）来学习 deepfake 检测所需的特征。

Andreas Rössler et al. in the [FaceForensics++ ](https://arxiv.org/pdf/1901.08971.pdf "FaceForensics++ ")paper evaluate several learning based methods and one hand-crafted feature extraction method on their ability to detect forgery at different video quality levels. They pre-extract the facial region from the input image (and extract a center-crop or resize for certain methods) and use the various methods to detect forgery. They noticed that all approaches achieved high accuracy on raw input data. However, performance dropped for compressed videos. Among the tested methods, the XceptionNet model achieved the highest performance across all video quality levels.

>Andreas Rossler 等人在 FaceForensics++ 的论文中评估了几种基于学习的方法，以及一种手工特征提取方法在不同视频质量水平上检测 deepfake 的能力。他们从输入图像中预先提取面部区域（并提取一个中心——根据某些方法剪切或调整大小），并使用各种方法来检测伪造。他们注意到，所有的方法都能在原始输入数据上达到高精度。然而，压缩视频的性能下降了。在测试的方法中，XceptionNet 模型在所有视频质量级别上都取得了最高的性能。

![](https://static.leiphone.com/uploads/new/sns/article/202011/1606203162807420.png)

Source: https://arxiv.org/pdf/1901.08971.pdf

Some deepfake creation methods can’t generate videos that are temporally consistent. Artifacts that arise due to this inconsistency may provide good cues for video based deepfake detection.

>一些 deepfake 的创作方法无法生成时间连续的视频。由于这种不一致性而产生的赝品可以为基于视频的 deepfake 检测提供很好的线索。

One way to incorporate temporal information along with spatial information for training a model is by using 3D CNNs. On that note, [Yaohui Wang et al.](https://hal.inria.fr/hal-02862476/document "Yaohui Wang et al.") analyze the ability of a few 3D CNNs (specifically I3D, 3D ResNet and 3D ResNeXt) to detect manipulated video. You can refer to their paper for more details on their various experiments.

>一种将时间信息和空间信息结合起来用于训练模型的方法是使用 3D 卷积神经网络。在这一点上，Yaohui Wang 等人分析了一些 3D 卷积神经网络（特别是I3D、3D ResNet 和 3D ResNeXt）检测被编辑过的视频的能力。你可以参考他们的论文了解他们各种实验的更多细节。

[Irene Amerini et al.](https://openaccess.thecvf.com/content_ICCVW_2019/papers/HBU/Amerini_Deepfake_Video_Detection_through_Optical_Flow_Based_CNN_ICCVW_2019_paper.pdf "Irene Amerini et al.") suggest using optical flow to detect deepfakes. The general idea is that deepfake videos may have discrepancies in motion across frames (such as unusual movements of facial parts) which optical flow can capture. For a given frame f(t), they use the PWC-Net model to estimate the forward optical flow OF(f(t), f(t+1)) which depicts the apparent motion of various elements in the scene. The extracted optical flow values are transformed to an RGB image format and is used by their Flow-CNN to detect deepfakes.

>Irene Amerini 等人建议使用光流检测 deepfake。这一想法是，deepfake 视频可能产生不同帧的运动差异（比如面部部分的不寻常运动），而光流可以捕捉到这些差异。对于给定的帧f(t)，他们使用 PWC-Net 模型来估计 (f(t)， f(t+1)) 的前向光流，它描述了场景中各种元素的视运动。提取的光流值转换为 RGB 图像格式，由其 Flow - CNN 用于检测 deepfake。

###Artifacts

At present, deepfake creation methods are not perfect. This means there is often evidence which we can analyze to deduce whether a piece of media was manipulated and/or is a deepfake. In this category, we will explore some detection methods that look for this evidence, often called artifacts.

>当前，deepfake生成技术还不完美。这意味着有很多证据可以拿来推断一段媒体文件是否是被操纵过或者是不是deepfake。在本类中，我们会探索查找这类证据（一般被叫做图案缺陷）的几种办法。

Some forgery detection methods work by observing the Photo Response Non Uniformity (PRNU) pattern on images. The authors of the paper Noiseprint mention that in the seminal paper by Lukas et al. they observe that each individual device leaves a specific mark on all acquired images (the PRNU pattern) due to imperfections in the device manufacturing process. The authors note that the lack of PRNU may indicate manipulation. Needless to say, PRNU-based methods can be used to detect deepfakes. However these methods do have some drawbacks, one of which is you need a large number of images to make good estimates.

>一些鉴伪办法通过观察图像的光响应非均匀性(PRNU) 图案来工作。Noiseprint这篇论文的作者提到了，在Lukas等人的研讨会文章中，他们发现由于生产过程的不完美，每个照相设备会给拍下的所有图像留下一个特定的痕迹（光响应非均匀性图案）。因此作者认为没有PRNU可能就意味着图像的伪造。不用说，基于PRNU的办法科技用来检测deepfakes。但是这些办法有缺点，比如你需要大量的图像来做出好的估计。

In another paper,[ Francesco Marra et al. ](https://arxiv.org/pdf/1812.11842.pdf " Francesco Marra et al. ")show that each GAN leaves a specific “fingerprint” in its generated images, similar to real-world cameras marking images with traces of their PRNU. The authors present an experiment showing evidence of GAN fingerprints.

> Francesco Marra et al. 的另一篇文章中展示了生成对抗网络可以给图像留下特定的、酷似真实世界中相机因PRNU给图像造成的痕迹的“指纹”。作者的一个实验展示了GAN产生指纹的证据。

For an image Xi generated by a given GAN, they mention that the fingerprint represents a disturbance unrelated with the image semantics. To obtain the fingerprint, they first calculate the noise residual Ri using Ri = Xi – f(Xi), where the function f is a denoising filter. They assume that this residual is a sum of a non-zero deterministic component (the fingerprint, F) and a random noise component (Wi). Hence, Ri = F + Wi.

>对于给定GAN产生的一张图片Xi，他们提到这个指纹其实代表了一种对图像意义无关的扰动。为了获得这种扰动，他们首先计算了噪点残值Ri ，用 Ri = Xi – f(Xi)，函数f 表示降噪滤镜。他们假定这个残值就是一个非零确定性的成分（也就是指纹，F）和随机成分（Wi）的加和。因此Ri = F + Wi。

The additive noise component mostly gets cancelled out on taking the average of residuals computed from many different images generated by the same GAN. Hence, by performing the above operation, we get an estimate of the GAN fingerprint.

>其中的加性噪声基本上会在给同一GAN生成的多张图片的残值取平均的过程中被消掉。因而通过以上一通操作，我们就得到了一个对此GAN的指纹的估计。

![](https://static.leiphone.com/uploads/new/sns/article/202011/1606203161216260.png)

Source:https://arxiv.org/pdf/1812.11842.pdf

In one of their experiments, the authors explored the ability of the fingerprints to tell apart images of different origin. They considered two different GAN architectures:

>在他们的一项实验中，作者探索了指纹区分不同来源图像的能力。他们考虑了两种不同的 GAN 架构：

- GAN A: A Cycle-GAN trained to convert orange images to apple images.

- GAN B: A Progressive-GAN trained to generate kitchen images.

>- GAN A：一个经过训练、能将橘子图像转换成苹果图像的 Cycle-GAN。

>- GAN B：一个经过训练、生成厨房图像的 Progress-GAN。

They suggest that, based on image-to-fingerprint correlation or similar indicators, meaningful fingerprints should allow one to decide which of the two GANs generated a given image. It was observed that:

>他们指出，基于图像到指纹的相关性或类似的指标，有意义指纹应该允许人们决定两个 GANs 中的哪一个生成给定的图像。有人指出：

Cross-GAN correlations (correlation between the residual of an image created by one GAN and the fingerprint of a different GAN) are evenly distributed around zero, indicating no correlation between generated images and the unrelated fingerprint.

Same-GAN correlations (correlation between the residual of an image created by one GAN and the fingerprint of the same GAN) are markedly larger than zero, testifying of a significant correlation with the correct fingerprint.

>Cycle-GAN 相关性（一种 GAN 生成的图像残差与另一种 GAN 生成的指纹之间的相关性）平均分布在零附近，说明生成的图像与不相关的指纹之间没有相关性。

>Same-GAN 相关性（生成的图像残差与相同 GAN 的指纹之间的相关系数）显著大于零，证明与正确的指纹之间存在显著的相关关系。

This result is interesting because we can potentially use it for detecting deepfakes, and even identifying the source of a deepfake. However, the authors mention that more research is necessary to assess the characteristics, usability and robustness of these fingerprints.

The work by Ning Yu et al. presents another elaborate investigation into learning and analyzing GAN fingerprints. One of their interesting claims is that even minor differences in GAN training could result in different fingerprints, which enables fine-grained model authentication. Overall, their study provides a good insight into the viability and capability of GAN fingerprints.

>这个结果很有趣，因为我们可以用它来探测 deepfake，甚至识别 deepfake 的来源。然而，作者提到更多的研究是必要的，以评估这些指纹的特征，可用性和鲁棒性。

>Ning Yu 等人的作品是对 GAN 指纹学习和分析的又一细致研究。他们的一项有趣的声明是，GAN 的训练可以产生微小差异，这可能导致不同的指纹，从而实现了细粒度模型认证。总的来说，他们的研究为 GAN 指纹图谱的可行性和能力提供了一个很好的见解。

###Challenges Facing Deepfake Detection

There are several challenges limiting the accuracy levels of deepfake detection methods. Let’s look at some below:

>目前存在着一些限制 deepfake 检测方法准确性的难题：

Degradation of Quality

Media that is shared over the internet may experience a loss in quality due to data compression, resizing, noise, and so on. This might pose problems to some deepfake detection algorithms.

As mentioned earlier, the paper FaceForensics++ compares the performance of a few forgery detection methods at different video quality levels. They noticed that while the methods had high accuracy on raw input data, performance dropped on compressed video.

It’s very important to develop deepfake detection methods that are robust to degradation of quality and content. Media items like deepfakes and forged evidence will likely have been circulated widely online in a variety of different quality levels.

>质量退化

>分享在网络上的媒体可能会由于数据压缩、调整大小、噪声等而出现质量损失。这可能会给一些深度假检测算法带来问题。

>如前所述，FaceForensics++ 一文中比较了几种不同视频质量级别的伪造检测方法的性能。他们注意到，虽然这些方法在原始输入数据上有很高的精确度，但在压缩视频上性能下降了。

>开发对质量和内容退化具有鲁棒性的深度假检测方法是非常重要的。像deepfakes和伪造证据这样的媒体报道可能会以不同的质量水平在网上广泛传播。

Artifact and Fingerprint Removal

It is possible that the artifacts or “fingerprint” information used for detection can be removed to trick deepfake detectors. For instance, the paper GANprintR introduces a simple auto-encoder-based method of removing the GAN fingerprint from synthetic fake images. This tricks facial manipulation detection systems while keeping the visual quality of the image. Strategies like this are a threat to the accuracy of certain deepfake detection methods.

Moreover, if deepfake detection methods focus on very specific traits (such as the abnormal blinking rate in a video), it’s likely people will attempt to develop improved deepfake creation methods that do not have those traits.

>伪影和指纹去除

>用于检测的伪影或“指纹”信息有可能被去除，从而欺骗 deepfake 检测器。例如，GANprintR 介绍了一种简单的基于自动编码器的方法从合成的假图像中去除 GAN 指纹。这骗面部操作检测系统，同时保持图像的视觉质量。这样的策略是对某些深度假检测方法的准确性的一种威胁。

>此外，如果 deepfake 检测方法专注于非常具体的特征（比如视频中不正常的眨眼频率），那么人们很可能会尝试开发不具备这些特征的改进的 deepfake 制作方式。

**Poor Generalization and Evolution of Deepfake Technology
**

Put simply, some deepfake detection strategies may not generalize well to content and/or manipulation techniques that differ from their training domain.

>简单地说，一些 deepfake 检测策略可能不能很好地推广到不同于其训练领域的内容和/或操作技术。

In one of their experiments (cross-manipulation techniques), Yaohui Wang et al. trained their 3D CNN detectors on real videos and fake videos created using three manipulation techniques. When they tested the trained model to detect deepfakes created by a technique that the model was not trained on, they noticed a drop in performance (as compared to their other experiments).

>在他们的一项实验（交叉操作技术）中，Yaohui Wang 等人对使用三种操作技术制作的真实视频和假视频进行了 3D-CNN 检测器的训练。当他们测试训练过的模型来检测由一种模型没有训练过的技术所产生的 deepfake 时，他们注意到（与他们的其他实验相比）性能下降了。

The ability to generalize to unknown domains is quite important in deepfake detection. For one, when applying deepfake detection in the real world, we most likely won’t know the true source of the content and the type of manipulation applied to a particular scenario, such as detecting disinformation. Also, as newer deepfake creation methods are developed, poorly generalizing strategies will need to be constantly updated to cover these new techniques.

>在 deepfake 检测中，对未知域的泛化能力是非常重要的。首先，在现实世界中应用 deepfake 检测时，我们很可能不知道内容的真实来源和应用于特定场景的操作类型，比如检测虚假信息。此外，随着较新的 deepfake 创建方法的发展，较差的泛化策略将需要不断更新，以覆盖这些新技术。

Some strategies to detect deepfakes may be able to generalize to unknown domains more easily than others. It is important to identify such strategies and create detection methods that can adapt to unconstrained settings.

>一些检测 deepfake 的策略可能比其他策略更容易推广到未知领域。识别这样的策略和创建能够适应无约束设置的检测方法是很重要的。

###Conclusion

In this article I looked at three categories of deepfake detection techniques, and explored some general challenges we may face in the future. I hope this article gives you a general idea of currently developing deepfake detection techniques. The ideas I present here are far from exhaustive, and I encourage you to dive in yourself if you want to learn more about current techniques and challenges involved in deepfake detection.

>###结论

>在本文中，我介绍了三种 deepfake 检测技术，并探讨了一些我们将来可能面临的一般性挑战。我希望这篇文章能让你对当前发展的深度假检测技术有一个大致的了解。我在这里提出的想法远远不是详尽无遗的，如果你想了解更多关于当前技术和涉及深度造假检测的挑战，我鼓励你深入研究。

本作品采用知识共享署名-非商业性使用-相同方式共享 4.0 国际许可协议进行许可，转载请附上原文出处链接和本声明。
本文链接地址：https://www.flyai.com/article/789

讨论

500字

表情

每日优质讨论奖励 20FAI

发送

每日优质讨论奖励 20FAI

删除确认

是否删除该条评论？

取消删除