Remove dust from photos

Get rid of the dust specks dotting your scanned family photos, the sensor blobs on your landscape shot, the surface dust on your product photo, the lint on your studio backdrop. Magic Eraser cleans hundreds of dust specks in seconds — no clone-stamping each speck one by one, no manual spot-healing pass.

Last updated May 16, 2026

Why dust specks defeat manual editing at scale — Magic Eraser

A single dust speck is easy to clone-stamp. A scanned old photo with 300 specks across the sky and another 200 across a uniform background is a different problem — each speck needs its own clone source, each source needs to match the surrounding texture, and the work scales linearly with speck count. Magic Eraser handles the entire affected area as a region: brush over the sky, the wall, the studio sweep, or the affected portion of the scan, and the AI removes all the specks in one pass while preserving the underlying texture. Works for digital sensor dust on landscape sky, scanner glass dust on old photos, surface dust on product photos, and lint on studio backdrops.

ステップバイステップ手順

1. 1

   Upload your photo

   Open Magic Eraser on the web, iOS, or Android and drop in the scanned old photo, digital landscape, product shot, or studio image. JPEG, PNG, HEIC, WebP, and TIFF (where supported) are accepted.

2. 2

   Brush over the dusty region

   Paint over the affected area — the full sky, the uniform wall, the studio backdrop, the scanned negative's clear area — rather than dabbing at individual specks. A broad brush over the region tells the AI 'clean this entire surface' and runs as a single pass; dabbing speck-by-speck is slower and produces less consistent texture rebuilds. For scans with dust spread across multiple regions (sky AND foreground), brush each region in turn.

3. 3

   Tap Erase and inspect

   Magic Eraser cleans all the specks in the brushed region in seconds, preserving the underlying sky gradient, wall texture, or studio sweep. Scan at 100% zoom for any residual specks, run a second pass on missed areas, then export at full resolution for the photo album, product catalog, portfolio, or archive.

こんな方におすすめ

• Scanned old family photos with scratches and dust from the scanner glass or original print surface
• Vintage film scans (35mm, medium-format, large-format) where dust accumulated on negatives or transparencies
• Digital landscape and astrophotography shots with sensor dust visible in the sky
• Product photography where surface dust is visible on glossy or dark backgrounds
• Studio backdrop shots with lint, hair, or dust on the sweep paper
• Documents and book-page scans where dust on the scanner glass shows in the margins
• Slide-film digitizations with archival dust from years of storage
• Microscopy and macro photos where dust on lens elements is visible at high magnification

注意事項

Dust removal is one of the cleanest AI photo edits because specks are usually small, dark, and against a relatively uniform background. Two edge cases benefit from extra care. First: when a 'speck' is actually fine subject detail (a distant bird in the sky, a small flower in a landscape, a faint star in astrophotography), the AI may treat it as dust and remove it; mask only the regions where you're confident the dust is genuinely dust, and inspect carefully at 100% before exporting. Second: when scratches and dust both appear together on a scanned old photo, brush over both — the AI handles scratch repair and dust removal in the same pass, but allow extra time on the inspection step because the scratch repair sometimes needs a touch-up pass to match the surrounding texture. For archival and restoration work, keep the original scan as the master and save the cleaned version as a separate file; archives standards (IFLA, NEDCC guidelines for photographic preservation) consistently recommend preserving the original alongside any cleaned derivative.

よくある質問

Is it free to remove dust from a photo?

Yes. Magic Eraser's free tier covers dust removal with daily usage limits. Upgrading to Premium ($29.99/year) removes the limits and unlocks higher-resolution exports — useful for printed family-photo restoration, product catalogs, and archival digitization projects.

Can I clean an old scanned photo with hundreds of dust specks?

Yes. Brush over each affected region (sky, uniform background, faces if dust is over the cheeks, walls) and erase. The AI cleans all the specks in the masked region in one pass; for scans with dust spread across many regions of the photo, run the brush-and-erase pattern for each region. This is far faster than clone-stamping speck by speck and produces more consistent texture rebuilds.

What about scratches on the same scan?

Brush over scratches and dust together. Magic Eraser handles both in the same pass — scratches read to the AI as elongated dust, and the rebuild logic is similar (preserve underlying texture, remove the linear or point defect). Allow extra inspection time at 100% zoom because scratch repair occasionally needs a touch-up to match the texture direction of the surrounding area, especially on textured surfaces like fabric, brick, or grass.

Will the cleaned sky, wall, or backdrop look obviously edited?

On most photos, no — the AI preserves the underlying gradient, grain pattern, and color. Two cases where it can show: skies with fine cloud texture or directional grain where the rebuild may slightly soften the pattern (run a second pass with a smaller brush to restore character); and product photos with deliberate subtle texture (a matte material that has its own grain) where the AI may smooth more than intended (mask more tightly and run a finer-brush touch-up).

Should I keep the original scan as well as the cleaned version?

Yes. For archival, restoration, and family-history projects, IFLA and NEDCC photographic preservation guidelines recommend keeping the original digital master alongside any cleaned derivative. Use a clear file naming convention — `IMG_1942_original.tif` and `IMG_1942_cleaned.tif` — so the relationship is unambiguous decades from now. The original is the historical record; the cleaned version is the working print.