Home » Technology » Despite Prior Accusations, Samsung is Once Again Accused of Faking Moon Photos

Despite Prior Accusations, Samsung is Once Again Accused of Faking Moon Photos

Jakarta

The 100x Space Zoom feature in Samsung Galaxy S23 Ultra received criticism from a Reddit user with the account name ‘ibreakphotos’. According to him, the moon photos produced from this feature are fake.

100x Space Zoom is a feature that has existed since the S20 Ultra, which is a combination of the camera’s optical zoom capabilities and computational processing. As is known, this feature can produce very detailed moon photos for an HP camera.

“Many of us have seen stunning photos of the moon with the latest zoom lenses, starting with the S20 Ultra. However, I have always doubted their authenticity, because the results are almost too perfect,” wrote ibreakphotos.

Furthermore, he also said that the moon photo from this feature was actually not really fabricated, but also not completely original. To prove his accusations, he downloaded high-resolution photos of the moon and carried out several stages of testing.

He reduced the resolution of the image to 170 x 170 pixels and added a Gaussian blur (a feature in photo processing software to blur images) to remove some of the detail from the moon.

The moon image is processed in such a way as to be blurry. Photo: imgur/ibreakphotos Photo: undefined

The image is then displayed on his computer monitor in full screen mode, which makes the image of the moon very blurry and pixelated and does not look like the moon.

ibreakphotos then takes pictures on the computer monitor using the S23 Ultra. The result? Yes, the photo of the moon taken from the blurry image is much clearer and better, complete with details like the moon photos produced by other Space Zoom features.

A blurry image of the moon on a computer monitor is then shot using the S23 Ultra. Photo: imgur/ibreakphotos Photo: undefined
S23 Ultra moon photo results. Photo: imgur/ibreakphotos Photo: undefined
Comparison of the pictures taken with the results of the photos. Photo: imgur/ibreakphotos Photo: undefined

Then he wrote up the conclusions of the test, and initially praised Samsung for being able to use AI to produce detailed images of the moon even though the objects being photographed were very blurry. However, he criticized the statement that Samsung had issued regarding the Space Zoom feature.

“There’s a difference between processing a high-resolution image, where multiple images are combined to restore lost detail, and this, where you have an AI model trained specifically on an image of the moon, i.e. to be able to recognize the moon and replace it with a textured image of the moon to in images (where no detail can be saved, as in this experiment),” wrote ibreakphotos.

In fact, regarding this kind of accusation, Samsung apparently has explained about the 100x Space Zoom on the S20 Ultra. That, there really is an AI factor. At that time, Samsung said that this could be done because the cellphone would record 20 images at once when shooting. Then AI will evaluate and correct the resulting details in the image.

However, that explanation seems unacceptable to ibreakphotos. According to ibreakphotos, if the ‘scene optimizer’ feature is turned off, you will see real photos of the moon that can be produced by the S23 Ultra (and various Samsung cellphones with other Space Zoom features).

“This is not sharpening, this is not adding details from various photos because in this experiment, all photos come from images that have no details. None of these photos show craters on the moon because they are deliberately blurred. But it turns out that the shots can magically tell that (the crater) is there,” he concluded.

Apparently, ibreakphoto is very, very anti with AI-assisted zoom technology. In fact, many people are also helped by the combination of these two technologies. How about you, detikers?

Watch VideoTake a peek at the Epic Features of the Samsung Galaxy S23 Ultra

(asj/fay)

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.