How to Unflatten an Image
in Photoshop

This undoes the last action. If flatten was the last thing you did, a single press restores all layers immediately. The Layers panel will repopulate with every layer exactly as it was before the flatten.

If you made other edits after flattening and need to step further back, hold Alt (Win) or Option (Mac) while pressing Ctrl/Cmd + Z to undo multiple steps in sequence.

The moment your layers are back, go to File → Save As → Photoshop (.PSD) before doing anything else. This locks your recovered layers into a file that will survive closing Photoshop.

💡 By default Photoshop stores 50 history states. If you've made more than 50 changes since flattening, earlier states may have been pushed out. Increase this to 200 in Edit → Preferences → Performance → History States.

Go to Window → History. You'll see a chronological list of every action taken in the current session, from top (oldest) to bottom (most recent). The panel is usually docked alongside Layers and Channels.

Scroll through the history states until you see the entry labelled "Flatten Image". This marks the exact moment all your layers were merged into a single pixel layer.

Click on the history state directly above (before) the Flatten Image entry. Photoshop instantly reverts to that point — your layers reappear in the Layers panel exactly as they were.

⚠️ Any actions you took after flattening will be lost when you revert. If you did significant editing after the flatten, decide carefully whether undoing the flatten is worth losing that work.

Go to File → Save As → Photoshop (.PSD) and save. This preserves your recovered layers in a format that survives closing Photoshop.

In the Layers panel (Window → Layers), double-click the thumbnail image of the Smart Object layer — not the layer name. This opens the Smart Object's contents in a separate document window (it will have a .psb extension in the title bar).

In the new .psb document window, click the top layer in the Layers panel, then Shift-click the bottom layer to select them all. You'll see all your original layers listed exactly as they were before the merge.

Click the three-dot panel menu in the top-right corner of the Layers panel and choose "Duplicate Layers…". In the dialog, select your original document from the Document dropdown. Click OK — all layers are copied across.

Switch back to your original document. All layers from the Smart Object are now present. Close the Smart Object window without saving (no changes needed inside it). Delete the original Smart Object layer — it's no longer needed. Save as PSD.

💡 Going forward, use Convert to Smart Object (right-click layers → Convert to Smart Object) instead of Flatten Image — it compresses without destroying layer recoverability.

Open your flat JPEG, PNG, or PSD in Photoshop. In the Layers panel, the image will appear as a locked Background layer. Double-click it to convert it to a regular unlocked layer (rename it "Original" and click OK). Immediately duplicate this layer (Ctrl/⌘ + J) as a safety backup before touching anything.

Go to Select → Subject. Photoshop's AI analyses the image and automatically selects the primary subject. For a person against a background, this usually produces an excellent selection in seconds. Alternatively, use Edit → Remove Background (Photoshop 2024+) — one click, transparent background output, no manual selection needed.

With the selection active, go to Select → Select and Mask (or press Alt + Ctrl/⌘ + R). Use the Refine Edge Brush to paint over hair, fur, or complex edges. Set Output to "New Layer with Layer Mask" — non-destructive, fully adjustable later.

With the subject selected, press Ctrl/⌘ + J to copy the selection to a new layer. Rename it descriptively — "Subject", "Person", "Product". Repeat this process for each distinct element you need to separate.

On the original layer, the area where the subject was removed is still showing the underlying pixels. Ctrl/⌘-click the subject layer thumbnail to load its selection, expand it slightly (Select → Modify → Expand → 5–10px), then on the original layer go to Edit → Fill → Content-Aware. Photoshop intelligently fills the gap.

💡 For complex backgrounds, open Edit → Content-Aware Fill workspace for a sampling brush and live preview — much better results than the basic fill.

When painting or cloning to fill gaps, press I to activate the Eyedropper Tool. In the Options bar, set Sample Size to "3 by 3 Average" for the most representative colour sample rather than a single pixel value.

Use the Rectangular Marquee Tool (M) to draw a selection around the text you want to identify. Include only one line at a time — multi-line selections confuse the AI. The cleaner and higher-resolution the text, the better the result.

Go to Type → Match Font. Photoshop analyses the selection and returns a ranked list of matching fonts — both installed on your system and available via Adobe Fonts. Select the closest match and click OK.

Select the Type Tool (T), click on the canvas, and type the text. Use the Character panel (Window → Type → Character) to match font size, tracking, and leading. Use the Eyedropper (I) to sample the exact text colour from the flattened image. Position the new text layer over the flat version and mask or hide the original.

💡 If Match Font doesn't find an exact match, try WhatFontIs.com or Font Squirrel's Matcherator — upload a screenshot of the text for external identification including free alternatives.