In a world where photographs can be digitally modified with just a few clicks and even produced from scratch with artificial intelligence, it is becoming increasingly difficult to believe what our eyes see.
The techniques used to manipulate images are so advanced that we have entered the age of hyper-realistic fake images.
Such images can lead to the spread of misinformation and even influence public opinion during important events such as elections.
This issue has come to the fore like never before, when a photo of Princess of Wales Kate Middleton and her children was withdrawn by news agencies due to concerns that it was “manipulated”.
So, what can be done to detect an image that has been edited or produced by artificial intelligence?
Abnormal lighting often gives away that a photo has been altered. For example, check light reflections in people’s eyes; The light source will often be reflected in the eyes.
If the size and color do not match the location or there is a difference between the eyes, you may have reason to be suspicious.
The way subjects and objects appear on reflective surfaces in an image can also provide clues.
Shadows of objects in the image may not line up if they are stitched together from multiple images, but keep in mind that some photos may be taken with more than one light source.
It’s worth examining how light appears on a subject’s face. For example, if the sun is behind them, their ears may appear red.
Artificial intelligence can also produce mismatched light and shadows, However, as algorithms improve, faces produced by artificial intelligence can often be mistaken for more real than human faces.
Another decisive method is to look for features that are difficult to copy. The AI is currently weak at rendering hands and ears; He can confuse their shapes, proportions and even the number of fingers.
These are the same features that painters find difficult. But as other aspects of AI-generated images of people become hyper-real, these inaccuracies create an unnatural and uncomfortable feeling to our eyes.
Hidden in the codes of digital images are bits of information that can help detect a fake image.
Every time a digital camera captures an image, metadata is written to the image file.
For example, time stamps have raised questions about whether President Donald Trump was actually in the White House a day after announcing he had Covid-19 in October 2020.
In every digital camera sensor, There may be minor manufacturing errors that lead to unique errors that leave a kind of “fingerprint” on the photos.
This is then associated with a specific camera and can help identify manipulated areas of a photo.
The grain of an AI-generated image may also look odd.
Internet companies like Google have introduced image verification tools that can help people identify images produced by artificial intelligence.
Facebook and Instagram have begun tagging AI-generated images from Meta’s own systems and plan to do the same for images produced by other companies’ AI tools.
This news our mobile application Download using
You can read it whenever you want (even offline):
Source link: https://www.donanimhaber.com/yapay-zekayla-olusturulmus-sahte-gorseller-nasil-anlasilir–175186