Face morphing is a process in which two face images are blended in such a seamless way that it creates an entirely realistic and new image. So in other words face morphing is a technology that utilizes digital manipulation to merge one face into another. Although it can be a fun activity, face morphing potentially has major drawbacks as well. It can used to create deepfakes, which are highly realistic videos used for malicious purposes.
Working
Facial morphing software takes two images and analyzes them by pinpointing key features of the human face, like eyes, nose, and mouth. It then creates a series of images by blending the features of one face into the other. Consequently, creating a final image that is very difficult to distinguish from a genuine photograph.
Here’s a more detailed explanation of the work;
The face-morphing process involves three main steps;
Landmark identification
Key points on the face(like the mouth, nose, eyes, and chin) are identified so that a coordinate system can be created. This system outlines their locations and shapes.
Geometry Warping
In this step, the shape of the face is adjusted by positioning these key points towards an average between corresponding points in two photos. This eventually creates a blended or in other words a midway face.
Texture and Color Blending
In the third step, the color and texture of the blended face are filled in. Certain details like wrinkles and pores are smoothed. This gives the image a filtered look that is similar to Instagram. However, there are times when the post-processing methods can address any quality issues that are caused by blending. This makes the morphed faces more genuine looking as well.
How and Why Is Morphing Considered A Threat?
Face morphing may seem like a fun endeavor but unfortunately, it has become a double-edged sword. Its capability to create a realistic deepfake may pose a significant threat. Let’s discuss how and why a facial morphing attack is considered a threat;
Reputation assassination
Imagine a deepfake video of a company’s CEO making offensive remarks about another organization, or a celebrity engaging in some scandalous behavior. These videos, created via morphing techniques, can go viral and can cause immense damage to the reputation of the individual. For example, recently a deep fake video emerged online of a Bollywood actor Ranveer Singh bashing the current PM Narendra Modi amid the 2024 elections. The video, though demonstrably fake, spread widely on various social media platforms, highlighting the potential dangers of morphing.
Financial Fraud
Deepfaes can also be used to impersonate executives in videos. In 2020, there was a deepfake audio recording of a CEO wiring over $240,000. This was a fraudulent video but showcased how morphing could be used to sway investor decisions.
Social Disruption
Deepfakes can be used to create distrust among people. Imagine a fabricated video depicting a political leader inciting violence or asking his followers to declare war. The potential to create discord among diverse groups or manipulate public opinion is tremendous. Hence, one needs to be careful while dealing with morphing software.
Why Morphs Are Difficult To Detect
The Brain’s Blind Spots:
The human brain is wired to recognize faces. But in doing so it often fills details or overlooks slight inconsistencies without any effort. This makes a person susceptible to the intricate manipulations of the morphing techniques. In a well-crafted deepfake video, the actual and fabricated elements are merged so seamlessly that it becomes almost impossible for an untrained human eye to detect or pinpoint any difference.
AI-Powered Realism
Advanced AI Algorithms are often used in deepfakes. This allows the systems to analyze facial expressions, lip movements, and speech patterns with amazing accuracy. And it becomes almost impossible to differentiate between truth and fiction.
Morphing Attack Detection
Here are some methods that can assist in preventing attacks.
Multi-factor Authentication
It’s better to use multiple layers of authentication rather than solely relying on facial recognition or other forms of biometric methods. For instance, combining facial recognition with a PIN or password can be an effective strategy.
Anti-Spoofing Techniques;
Companies should try to implement anti-spoofing measures to detect fake or altered images. Using liveness detection can ascertain that the face being recognized is live and is not a face morphing attack via static morphed image or video.
Empowering Users with Knowledge:
Educating the masses about face deformation attacks and how they work is imperative. Public awareness campaigns can teach people how to distinguish between original and constructed images online. If they are encouraged to double-check the source, it will become easy to pinpoint inconsistencies. Moreover, digital literacy programs will guide people on how to identify suspicious content.
Conclusion
In recent times, the menace of image-morphing attacks has become a cause of concern for a lot of people. Hence, it is something that most definitely cannot be ignored. One needs to apply a comprehensive and robust drawdown strategy that can detect and potentially eradicate deepfakes for fraudulent activities.
