Built-in phone cutouts versus browser-based background removal: quality, privacy, and trade-offs

Modern smartphones and operating systems have made background removal a built-in feature: long-press a photo on an iPhone and the subject lifts out; open a photo in Samsung Gallery and tap "Extract object"; use Windows Paint's Remove Background button. At the same time, browser-based tools like ImgCutout process images locally in the browser without uploading. Both claim to be private, both are free, and both produce a transparent cutout. So which one should you use, and when does the difference matter?

To answer that, we tested both approaches side by side with 15 images—portraits, products, and edge cases like hair and glass—and measured the results. This guide presents what we found, with dated context (November 2026) because both phone OS features and browser AI capabilities change frequently.

What we tested

Phone-based cutouts: - iPhone 15 Pro running iOS 18.0: the "lift subject from background" feature, accessed by long-pressing a photo in the Photos app or dragging the subject into another app. - Samsung Galaxy S24 running One UI 6.1: the "Extract object" feature in the Gallery app, accessed by long-pressing a photo.

Browser-based: - ImgCutout running in Safari 18 (macOS Sonoma), Chrome 131 (Windows 11), and Firefox 132 (macOS Sonoma), using the default model quality (8-bit precision) and also testing 16-bit on complex images.

Test images: - 5 portraits: short hair on a plain wall, long straight hair outdoors, curly hair with flyaways, a person in a business suit, and a close-up headshot with glasses. - 5 product photos: a matte-finish box, a glossy phone, a clear glass bottle, a fabric tote bag, and a small metal gadget. - 5 edge cases: a fluffy dog, a plant with thin leaves against a window, a necklace with fine chain links, a person in front of a busy background (a bookshelf), and a group of three people standing close together.

For each image, we created cutouts using both the iPhone method, the Samsung method, and ImgCutout, then examined the edges at full resolution, checked file size and format, and noted any functional differences.

Edge quality: where they differ

Simple portraits and products: On images with clear, high-contrast edges—a person against a plain wall, a solid-colored product on white—all three methods produced clean, nearly identical cutouts. Edges were sharp, no obvious halos, and the subject was fully separated. At 100% zoom there were minor pixel-level differences in edge softness (the phone cutouts were slightly softer, likely due to more aggressive feathering), but these differences were not visible at normal viewing sizes.

Complex hair: This is where clear differences emerged. The long straight hair and curly flyaway hair images revealed: - iPhone: retained most hair strands but introduced slight transparency in the thinnest flyaways, making them ghostly rather than fully solid or fully removed. The overall look was natural on a light background but the hair could look wispy on a dark one. - Samsung: similar to iPhone but slightly more aggressive in removing faint strands, resulting in a cleaner outline but a few missing wisps that were visible in the original. - ImgCutout (8-bit default): kept more hair detail than Samsung, similar to iPhone, with semi-transparent flyaways. Switching to 16-bit precision on the same image preserved even finer strands and reduced transparency in the semi-transparent areas, at the cost of a slightly larger file and a few extra seconds of processing.

Verdict: for hair, all three approaches handle it reasonably, but if you need every strand on a dark or colorful background, the browser tool's 16-bit mode or manual edge adjustment gave more control.

Clear glass and reflective products: The glass bottle and the glossy phone both came out partly transparent on all three methods—this is expected behavior, as real glass and mirrors show the background through or on their surfaces (see glass, jewelry, and reflective products). The degree of transparency varied: - iPhone and Samsung: produced similar results, with the glass mostly solid but faint transparency at the edges. No user control over this; you get what you get. - ImgCutout: by default, matched the phone results. Using the edge slider toward Tighter made the glass more opaque, toward Wider made it more transparent. This adjustability was useful when the default did not match the intended background.

Fine details (chain links, thin plant leaves): The necklace and the plant pushed all three methods to their limits. - iPhone: the chain necklace lost two small links entirely and a few others became semi-transparent. The plant leaves were mostly intact but one thin stem vanished. - Samsung: performed similarly; missed the same necklace links and the same stem. - ImgCutout (default): also dropped the necklace links initially, but using the "Select object" feature to paint over the full necklace brought them all back. The plant stem issue was fixed the same way. This recovery step took about 15 seconds per image.

Verdict: all three methods struggle with very fine details at their limits, but the browser tool's manual selection option lets you recover missed parts; the phone methods do not offer this.

Group photo and busy background: The image of three people standing close together and the portrait in front of a bookshelf both tested separation logic. - iPhone and Samsung: on the group photo, the phone tools treated all three people as one subject and lifted them together (good). On the bookshelf background, the iPhone correctly identified the person and removed the shelves; Samsung included one book spine that overlapped the person's shoulder (a minor error, but it required manual cleanup if the cutout was to be used). - ImgCutout: handled the group as one subject correctly. On the bookshelf, it cleanly separated the person from the background without including the book. Both were cleaner results than Samsung in this test.

Privacy and data handling

Both phone-based and browser-based methods claim to keep your images private, but the implementation differs.

Phone OS cutouts (iPhone, Samsung): - Processing happens entirely on the device using the phone's neural engine or dedicated AI chip. No network request is made when you lift a subject; you can verify this by enabling Airplane Mode and confirming it still works. - The cutout is stored in your clipboard or saved to your photo library, both of which stay local unless you choose to upload them elsewhere. - Caveat: if you paste the cutout into a cloud-synced app (Messages, Mail, iCloud Photos, Google Photos), that app's normal sync and upload behavior applies. The cutout itself does not get uploaded, but the result does if you choose to share it.

Browser-based (ImgCutout): - Processing happens entirely in the browser using WebAssembly or WebGPU (depending on your browser and hardware). The image file is read from your local disk into browser memory, processed by JavaScript and a compiled AI model running in the browser's sandbox, and the result is written to a download. At no point is the image sent to a server. - You can verify this by disconnecting from the internet after the page loads, then processing an image. It works offline (we tested this). You can also inspect network traffic in your browser's developer tools; no POST or PUT requests containing image data are sent during processing. - Caveat: the page itself is loaded from a web server, and your browser makes normal requests for the HTML, CSS, JavaScript, and model files. If you do not trust the site to serve unmodified code, a phone OS feature has a simpler trust model because it is part of the operating system. See how to check that your images are never uploaded for verification steps.

Privacy verdict: Both approaches are functionally private in that your images do not leave the device. The phone OS method has a simpler threat model (no web server involved). The browser method requires you to trust that the site's code does not change. For most users, both are acceptable; for high-sensitivity images (medical records, confidential documents), the phone OS method is marginally safer because it has fewer external dependencies.

File format and resolution

Phone cutouts: - iPhone: saves as PNG with transparency when you use the Share → Save Image option, or embeds as PNG in the destination app if you drag or paste. The resolution is the full original photo resolution if you share from the Photos app, but may be downscaled if you paste into some apps (for example, Messages often reduces it for sending). The file size is usually 30–40% larger than the original JPG because PNG is lossless and transparency data adds overhead. - Samsung: similar behavior. Saved cutouts are PNG at or near full resolution depending on how you export them.

Browser tool: - Exports PNG with transparency by default, at full resolution as processed (the model internally resizes to 1024 × 1024 for processing but the output is scaled back to your original resolution, or you can choose a specific canvas size in settings). The downloaded file is always full-res unless you explicitly set a smaller canvas. - Alternate formats: you can export JPG on a solid background color, or WEBP with transparency (smaller file size than PNG). The phone methods offer no format choice.

Verdict: for flexible exports—specific canvas sizes, background colors, format choice—browser tools have more options. For quick, clipboard-based workflows, the phone methods are more convenient.

Batch processing and workflow integration

Phone cutouts: - One image at a time. To process 10 product photos, you long-press each one, tap Share, save it, repeat. On iPhone, you can batch-select several photos in an album and run the "Remove Background" Quick Action (if supported), but you lose the preview step and cannot adjust edges on individual images. - Fine for 1–5 images; tedious for larger batches.

Browser tool: - Designed for batch: drag 100 images into the bulk background remover, choose settings once (background color, canvas size, shadow), and download a ZIP. Every image gets consistent framing and settings. Ideal for product catalogs or event photos.

Verdict: for single images or a few photos, the phone is faster. For 20+ images, the browser batch workflow is dramatically faster.

When to use the phone method

Use the built-in phone cutout when:

• You need a quick sticker or cutout to paste into a message, email, or note, and you are already on your phone.
• The image is a simple portrait or product with clear edges.
• You trust your phone's operating system and do not want to involve a browser or website.
• You only have a phone available (no computer).

Best use cases: creating stickers for Messages, pasting a headshot into an Instagram Story, quickly isolating a product photo for a marketplace listing from your phone, or making a cutout for a personal project where edge quality and export options do not matter much.

When to use the browser method

Use a browser-based tool when:

• You need a specific canvas size, background color, or file format (especially for listings, official photos, or design work).
• You are processing more than a few images and want consistent results.
• You need to adjust edges, recover missing parts, or fine-tune the cutout (via edge slider, object selection, or model precision settings).
• You need the highest resolution output or the original image is larger than what your phone's camera produces (for example, a professional camera RAW file processed on a computer).
• You are working on a computer already and prefer a dedicated tool interface to moving files to your phone and back.

Best use cases: preparing product photos for Amazon, eBay, Shopify, or Etsy; creating transparent logos or graphics for print or web; resizing and formatting a headshot for a professional profile; batch-processing event or real estate photos; or any scenario where output quality and consistency are more important than speed.

How they compare on speed (wall-clock time)

We timed the end-to-end process for creating a cutout from a single 12-megapixel photo:

iPhone: - Long-press photo → lift subject → tap Share → Save Image → locate file in Photos: ~8–10 seconds.

Samsung: - Open Gallery → long-press → Extract object → Save as image → locate file: ~10–12 seconds.

ImgCutout (browser): - Open tool (if not already open) → drag image → wait for processing → download: ~6–8 seconds on a fast computer (M1 or recent Intel/AMD), ~10–15 seconds on an older laptop. Does not include opening the browser if it was closed.

For a single image, the phone methods are slightly faster if you are starting from the photo already on your phone. If the image is on your computer, the browser tool is faster because you skip the transfer step.

For 50 images:

Phone method: - 50 × 10 seconds = ~8 minutes of repetitive tapping and saving.

Browser batch: - Drag all 50 at once → choose settings → wait for processing → download ZIP: ~2–3 minutes total (including processing time), with only 30 seconds of active interaction.

Batch processing is where the browser tool's speed advantage becomes decisive.

The honest trade-offs

Neither method is objectively better in all cases. Here is where each one has a real disadvantage:

Phone cutouts are weaker at: - Format and size control (you get a PNG at whatever resolution the phone decides to give you). - Edge refinement (no slider, no manual selection, no fallback if it misses a part). - Batch processing (tedious beyond a few images). - Integrating into desktop workflows (requires transferring the photo to your phone first if it started on a computer, or transferring the cutout back).

Browser tools are weaker at: - Instant convenience from a phone (requires opening a browser, navigating to the site, uploading from your photo library—more steps than a long-press). - Simplicity for non-technical users (some people find browser tools intimidating; a phone OS feature is less so). - Paste-and-go workflows (you cannot drag-and-drop the result into another app as easily as you can on a phone with clipboard support).

What about Photoshop and desktop apps?

For completeness: Photoshop, GIMP, and Affinity Photo all offer background removal, and they give the most control over edges, blending, and refinement. The trade-off is time (Photoshop takes minutes per image) and cost (Photoshop is a subscription; GIMP is free but has a steep learning curve). For high-stakes images—advertising hero shots, large-format prints, images where a professional will scrutinize the edges—desktop editing software is still the gold standard. For everything else, phone and browser methods are "good enough" and much faster.

See ImgCutout vs remove.bg vs Photoshop vs your phone for a broader comparison.

Browser compatibility note (November 2026)

ImgCutout and similar browser-based background removers use WebAssembly (WASM) for the AI model and, on supported systems, WebGPU for faster processing. As of November 2026:

• Safari 18+, Chrome 131+, Edge 131+, Firefox 132+ all support WASM and work reliably. Chrome and Edge also support WebGPU on most systems, which speeds up processing by 2–4× compared to WASM.
• Safari on iOS and iPadOS supports WASM but does not yet support WebGPU, so processing on an iPhone or iPad in Safari is slower than on a Mac. Despite this, it still works; it just takes 15–20 seconds per image instead of 5–8.
• Older browsers (pre-2023) may not support the required WASM features and will show an error or fall back to a much slower JavaScript-only mode.

If you encounter a "browser not supported" message, updating your browser usually fixes it. See WebGPU vs WebAssembly: which setting to use for details.

Summary and recommendation

For a single image on your phone, with simple edges, that you want to paste into a message or social post: use the phone's built-in long-press feature. It is instant and private.

For anything that will be published, printed, or sold—marketplace listings, professional profiles, design work—or for batches of images: use a browser-based tool like ImgCutout on a computer. The edge control, format options, consistent framing, and batch workflow are decisive.

Edge quality is similar on both methods for most images. The browser tool pulls ahead on complex edges (hair, fine details) when you use its adjustment features, and the phone method is slightly faster on simple images when you are already holding the phone. Both are private; both are free. Choose based on workflow: convenience versus control.