AI Object Removal – which phone is best? | Infinium-tech

Introduction

AI is all angry these days. It has become an integral part of our smartphone and we have leaked in many of the tasks that we use – typing, talking, searching – but it actually began with photography.

Large keywords were computational photography and image algorithms, and at some point, each manufacturer began to brag about his new AI eraser, which could remove distractions from your holiday photos. Google seemed to be the first with magic eraser, but now the industry has caught, and we say that this AI has also left Pixel behind in removing AI object.

Being naturally eager, especially when technology is concerned, we decided to pits separate AI eraser against each other, which is best to see, but also to see if they are worth using.

After a lot of testing around the office, we chose four contestants as our testing subjects – using Samsung Galaxy S25 Ultra Galaxy AI, iphone 16e with Apple Intelligence, Gemini AI with Google Pixel 9 Pro XL, Gemini AI, and Xiaomi 15, Touting their own AI.

These devices represent the best efforts of their respective companies, and you can expect the same results of their other high-end phones. Generally, the midrange devices also do not even do – perhaps due to their more limited hardware.

Samsung Galaxy S25 Ultra • Apple iPhone 16e • Google Pixel 9 Pro XL • Xiaomi 15

We spent a large part of the time in choosing the right photos for the job. We finished with the selection of 20 images captured in a crowd of smartphones. In this way comparison is appropriate – every phone will use its AI on images captured with other phones.

Samples come in two groups – some distraction or general shots with others that ruin the photo – the specific example you will use the AI ​​-Operated Object Eraser. A road symbol that ruin a vista, a construction crane among buildings, people blocking a famous landmark, etc.

Then we have a classic photoboper shot where someone has crashed on your holiday photo and needs to be removed.

There are difficult objects to remove in another group of photos – a hand in front of a face, sunglasses that need to disappear, a bottle of glass cleaner among the bottles of souls at once – we designed these shots to challenge AI a challenge.

A word on the boundaries of AI

Before we look at the results, a word on each phone’s AI – what does it need to work and how it works.

Samsung

Samsung’s generative edit requires both an active internet connection and an enrolled Samsung account to work. It also shapes the image manipulated to 12MP (4,000 x 3,000px), and adds a watermark to tell you that the photo was tampered with by AI.

One of the strong suits of Galaxy AI is that it does not have restrictions to work on humans – meaning that it can edit some parts of people in a picture – some other competitors may not do so yet.

Samsung’s solution requires the least tap – open the gallery app, go to a photo, tap once, and you can draw around the obstacle you want to delete.

Despite its limitations, Samsung’s Galaxy AI object is better than the rest. This is rapid – average, especially compared to pixels and ziaomi. Samsung’s solution is also incredible in identifying what you want to find out – just draw a loose circle around anything in a picture, and 9 times out of 10, AI will recognize what you want to do and mask it completely.

Apple

Apple’s clean up is part of Apple Intellation, which was recently limited to North America. It has recently spread to more countries and regions, but it is still not available in China.

Clean Up iPhone 16E, iPhone 16, iPhone 16 Plus, iPhone 16 Pro, iPhone 16 Pro Max, iPhone 15 Pro, or iPhone 15 Pro Max works on Max; A17 Pro or M1 and later any ipad; M1 or later MAC, and iOS 18.1, iPados 18.1, or Macos Sequoia 15.1 is required.

You will only need only one active connection to download the machine learning when set apple intelligence, and it will work offline since then.

You can find clean up in the photo app, and you will need two taps, before you can draw the disturbing circle. The iPhone is not good to identify what you want to delete – an ordinary person’s silhouette may require a picture of five or six finger for the phone to mask the whole thing.

Google

Google’s magic eraser was the first of its kind, but it seems that Google has not improved much since its announcement. Which is strange to see how much airtize the feature gets in pixel advertisements.

Inside the Magic Eraser Tools section, Google is in the image editor of the photo. This means that if you want to remove an object then this three taps are far away. Google Photo will sometimes detect objects automatically to remove in a photo and inspire you to use the magic eraser as a suggestion in the first screen of the image editor, but it happened only a few times during our test.

Magic eraser does not require an active internet connection. Once you open a photo and choose the magic eraser, it will scan the photo and think automatically anything that thinks it is a distraction – it is sometimes easy – it is easy – it found the crowd of people in a photo and deleted them all, without the need to do anything.

Google offers a separate pixel studio, which, among other things, may erase objects in a picture. We tested it and saw the same results as magic eraser which are manufactured in Google photos.

If you need to choose something manually to remove, draw around it and will try magic eraser and take it out. It is only average in this regard, many inputs are required to properly mask a human silhouette.

Xiaomi

The solution of Xiaomi lives in the proprietary gallery and is called AI Eraser, AI Erase Pro, or AI Erase 2.0. It is the most difficult to achieve – before an entire four tap you can eventually start drawing around subjects or objects. This also requires an active internet connection, and it appears that it depends on the server in China because it is sometimes very slow, and it also refused to work a few times.

Xiaomi’s object recognition is not good. Drawing around a person achieved poor results – AI chose only a small part of the entire size, and we needed to attract additional circles up to 10 – sometimes up to 10 to cover a simple human shape.

Choosing many objects is a tiring task, and the app would simply clean all masks at one point, and we needed to start with scratches.