The cameras come out when my family gets together for the holidays to capture moments to share with others who live far away, as well as pictures of food and cousins having a good time. The strange distractions that sometimes appear in my photos, such a spilled napkin or a person’s shoe at the edge of a frame, are what I dislike photographing.
After analyzing a picture, Clean Up fills in the areas that have been erased and identifies objects you may wish to remove, such people or cars in the background. Sometimes the majority of viewers cannot see the modification, and other times the outcomes are quite subpar. I’ve developed a few basic tips to assist you achieve the finest cleaned-up photographs after testing the program on a variety of photo formats.
A brick structure on an uphill street is seen in two pictures. A number of vertical traffic posts are distracting in the first. The posts in the second have been taken down.
Distractions may be eliminated using the Clean Up tool.
CNET/Jeff Carlson
Unexpectedly, there has never been a program like Clean Up for removing little distractions from iPhone or iPad photos.A simple Retouch tool that can fix some sections is included in the Mac version; on compatible Macs, Clean Up takes its place.
However, it’s crucial to keep in mind that Clean Up is an Apple Intelligence feature, which means you can only view it if you have access to the Apple Intelligence beta and are using a compatible device. This includes Macs with M-series processors running MacOS Sequoia 15.1, iPads with M-series processors (including the iPad mini with the A17 Pro chip) running iPadOS 18.1, and iPhones running iOS 18.1.
See which features I believe you’ll utilize most and where Apple Intelligence’s alerts need to be improved for additional information.
What distinguishes Clean Up from other retouching programs?
In order to fill in the area where you’re performing a remedy, the repair tools in the majority of picture editing software duplicate adjacent or comparable pixels. For instance, they are excellent in eliminating dust spots or lens flares against a sky.
In order to determine what should occupy the region you’ve chosen, the Clean Up tool employs generative AI, which examines the whole picture. For instance, generative AI uses its knowledge of background foliage and tree texture to generate a substitute if you want to eliminate a dog standing in front of a tree. It also considers the orientation and lighting in the image.
The way generative AI generates the picture is what gives it its “generative” quality. The area’s pixels are essentially made of nothing: The program begins with a random dot pattern and rapidly iterates to produce what it thinks should show up in the same area.
Remember that generative AI-based retouching tools are the ultimate YMMV, or “your mileage may vary.” In challenging compositions, I’ve had excellent results; in regions I assumed the software would handle easily, I’ve received bad results.
Apple Intelligence Impressions: Don’t Expect Radical Change is now playing.
09:05
How to use Apple’s Clean Up tool to eliminate distractions
Two methods are used by the Clean Up tool to fix photographs. It makes suggestions for potential objects to eliminate, including people or cars in the backdrop, using machine learning. Alternatively, you may tell Photos to work on the region you want to erase by dragging it over. This is how the procedure is broken down:
1. Tap the Edit button after opening a picture. (On MacOS, hit the Return key or click the Edit button.)
2. Select Clean Up. Depending on your Internet connection, it may take a minute or more for Photos to download Clean Up materials the first time you use the feature. After analyzing the picture, Photos uses a translucent shimmer to identify any possible objects that should be eliminated.
Two screen grabs of a bearded guy taking a selfie from an iPhone. Cars and people may be seen in the background. The Clean Up interface of the Photos app is seen in the picture on the right, where highlighted things are indicated by arrows.
Tap Clean Up once the Photos edit interface is up. Images provide recommendations on what should be eliminated.
Jeff Carlson/screenshot. Tap a recommended item to delete it. Or, around any non-glowing object with a circle.
4. You may need to draw over the remaining spots to do further cleaning, so don’t be shocked if the area isn’t completely cleaned on the first try. You may press the Undo button if you’re not satisfied with a repair.
A close-up of a guy taking a selfie with others removed from the background. Except for one person’s legs, everyone is highlighted in the picture on the left. The same picture with a selection made to tidy the legs is shown on the right.
Use Clean Up again to continue cleaning the area if it doesn’t catch everything (see that the person’s legs aren’t highlighted in the picture on the left).
Jeff Carlson/ screenshot. Tap Done when you’re done. If you wish to start anew, you may go back to the original, just as with any other change in Photos: To get back to the original, use the More (…) button.
A surprising and awesome feature: Safety Filter
The Clean Up tool is mostly used to remove distracting objects from a scene, but it also has the ability to conceal the identity of a person in the picture.
Around their face, draw a circle. You don’t need to fill it out; a simple swipe would suffice. To hide the subject’s face, the photographer uses a blocky mosaic design.
A guy taking a selfie in two pictures. His face is surrounded by a circle selection on the left. A mosaic grid takes the place of the face at the right.
One ingenious use of the Clean Up tool is the Safety Filter.
Image courtesy of Jeff.
Where Clean Up will be most successful
It’s helpful to know where to concentrate your efforts since Clean Up works better in certain settings and locations.
These broad kinds of fixes have shown the most success in my testing:
minor diversions. Things like dust and threads on people’s clothes or garbage on the ground always end up looking good.
textures of the background. Tree leaves, grass, and stone are examples of areas that are easily recreated.
flare of the lens. Light bouncing between camera lens components may generate lens flare, provided it isn’t too big.
backdrop cars or bystanders that don’t take up much space.
regions with little history or description.
Examples of clean-up at work include deleting a lens flare from a sunset photograph of a ship in port, painting a bag between two people seated on enormous pumpkins, and deleting an out-of-focus dog from the background of a close-up of a flower.
With the originals at the top and the altered ones at the bottom, Clean Up may sometimes work successfully.
Generally speaking, while dragging around an area, be careful to capture any shadows or reflections that the object you need to delete casts. Thankfully, Photos often detects them and incorporates them into its choices.
Three iPhone displays using the Photos app’s Clean Up feature. at front of a rainbow-painted rock wall at a Stockholm metro station, a couple is posing for a picture. They are highlighted, chosen by the program, and then eliminated.
Make sure you choose reflections and shadows (left). Clean Up uses the wide selection (middle) to determine what should be eliminated. The remaining reflection may be eliminated with one more tool swipe (right).
Image courtesy of Jeff Carlson.
Avoid these areas while attempting to apply Clean Up
When you attempt to eliminate some Clean Up targets, you will get frustrated. For instance:
enormous regions. Photos objects and informs you to mark a smaller area if it’s too large, otherwise it ruins the region. In terms of imagining what might logically fit in such a vast area, it is also incongruous.
crowded places with distinct qualities. In general, distant tree leaves function well, but not when there are identifiable objects or buildings nearby. For example, it doesn’t work effectively to remove a noticeable leaf from a bunch of leaves or to remove individuals from well-known sites.
A mom and kid in an outdoor market, in two pictures. There is an orange traffic cone next to the child’s seat. The lady stands with her back to the camera. Attempts to remove the lady have left a visual jumble in the picture on the right.
Large items in the picture create a mess of pixels when they are removed.
Image courtesy of Jeff Carlson.
Where additional effort is needed for cleanup
Keep in mind that while Clean Up and the other Apple Intelligence capabilities are accessible to anybody with a compatible device who enrolls in the beta program, they are still officially in beta. (In general, I have certain opinions on installing beta software.)
Even if you may achieve some positive outcomes, there are still a few things I hope Apple will work on in next updates. Specifically, the replacement portions have inconsistent quality and sometimes resemble non-AI repair equipment. Apple’s algorithms should have been more adept at identifying what’s in a scene and creating substitute sections, in my opinion.
Your only choices in terms of user experience are to undo or reset the modification if you’re unhappy with what Clean Up provides for a removal. Additionally, you receive the same previously processed results if you undo and attempt again. In contrast, Adobe Lightroom provides three options for each patch, and if you’re not happy with the results, you may create a new set.
Three images of Lightroom taking out a sack that was adjacent to a huge pumpkin. A new replacement choice is shown on each screen.
You have three choices for a deleted region when using Lightroom (the iPhone app that is seen above).
Image courtesy of Jeff Carlson.
The anticipated expectations of Clean Up and other comparable AI-based cleanup programs are similarly problematic: Its potential has been shown, which increases the standard for what we believe every modification should accomplish. We anticipate the tool to perform better when it becomes confused and displays a disorganized collection of different pixels. In the next releases, maybe.
Take a look at the visual intelligence feature to learn more about what Apple Intelligence offers your Apple devices.