Quick answer: To tell if a photo is photoshopped, look for warped backgrounds near the subject, shadows and lighting that do not match, jagged or glowing edges, and repeated textures from cloning. Then run a reverse image search to find the unedited original. One sign only suggests editing; two or more independent signs, plus a matching original, settle it.
Learning how to tell if a photo is photoshopped is one of the most useful verification skills you can build. Edited photos turn up everywhere now: in dating profiles, marketplace listings, news feeds, and messages from strangers asking you to trust them. Some edits are harmless. Some change what the photo means. This guide shows you how to spot a fake edited photo with your own eyes, how to prove it by finding the original, and how to judge what you found, without any paid software.
A “photoshopped” photo is a real photograph that has been altered with editing software (named after Adobe Photoshop, though any editor can do the same work) to add, remove, move, or reshape something in the frame. That definition matters, because it draws the line this whole guide is built on: an edited photo is not automatically a dishonest one. Cropping, brightening, and removing a spot are edits too. The question is never only “was this edited?” It is “does the edit change what the photo claims?”
What does “photoshopped” actually mean, and when does editing become deception?
Most photos you see online have been edited in some way. Phones sharpen, brighten, and straighten every shot before you see it. Magazines smooth skin. Sellers brighten product photos. None of that is deception, and treating every edit as a lie will leave you doubting genuine photos. The signs a photo is edited are only half the job; the other half is judging what the edit does.
Think of edits on a scale. At the harmless end sit the edits that change how a photo looks but not what it shows: a crop, a brighter exposure, a colour correction, a blemish removed. In the middle sit edits that tidy a scene: a stranger removed from the background of a holiday photo, a bin moved out of frame. At the far end sit the edits that change the facts: a person added to a place they never visited, an object removed so a damaged item looks new, a sign reworded, a body reshaped to sell a product that does not exist. Only that far end is deception, but the middle of the scale is where most arguments start, because a “tidy” edit can hide something the viewer needed to see.
Three kinds of edit should always raise your guard. The first is an edit that removes context: a tight crop that hides the crowd, the warning sign, or the second person in the frame. The second is an edit that adds false evidence: a face pasted into a location photo, a document with retyped details, a product photo with a feature the real item lacks. The third is an edit that changes a person: reshaped bodies and faces in dating profiles, job applications, and before-and-after adverts. AAP FactCheck, Australia’s fact-checking service, notes that re-purposing a real image for a different context is “perhaps the most common way images are used to mislead”, and a crop or a removal is often how that re-purposing is done. AAP’s own cases make the point concrete: its fact-checkers exposed a photo of Egypt’s Giza pyramids with snow added to it by finding the identical image with the same shadows and no snow, and debunked a viral “pink peacock” photo by matching it to an image of an albino peacock.
So before you start hunting for pixels, ask what the photo is being used to prove. A heavily retouched portrait on a photographer’s portfolio proves nothing about the person in it. The same portrait on a dating profile, presented as “me, last week,” is a claim about identity, and the standard of proof is higher. Keep that question in mind through every check below.
How to tell if a photo is photoshopped just by looking at it?
Your eyes catch more manipulation than you might expect, if you know where to look. Open the photo at full size, zoom in, and work through the signs below slowly. No single sign is proof on its own. What matters is the pattern: several signs pointing at the same spot in the same image.
Warped backgrounds: the liquify giveaway
The liquify tool pushes pixels around to reshape a body or a face: slimmer waist, longer legs, narrower jaw. The tool does not know where the person ends and the background begins, so straight lines behind the subject bend. Look at door frames, floor tiles, fences, and walls close to the person’s outline. If a straight edge curves or ripples exactly where it passes the body, that is one of the strongest visual signs a photo is edited. The same warping appears around enlarged muscles and reshaped faces.
Lighting and shadows that disagree
Light behaves the same way across a real scene. Every shadow should fall in the same direction from the same light source, and every surface should catch the light the same way. In a composite (two photos joined into one) the pasted person is often lit from a different direction than the background. Check the shadows under feet, chins, and objects: shadows pointing in different directions, a shadow missing where one should fall, or a face brightly lit against a dull, overcast background are all signs of editing. Skin that looks unnaturally smooth and plastic, with no pores or fine lines, usually means heavy retouching rather than a composite, but it still tells you the image has been worked on.
Edges, halos, and cut-out outlines
When an element is cut from one photo and pasted into another, its outline rarely blends perfectly. Zoom in on hair, shoulders, and hands and look for a thin bright or dark halo, jagged steps where a smooth curve should be, or a faint line of colour bleeding at the edge. Hair is the hardest part to cut out cleanly, so stray hairs that end abruptly, or a fringe that looks too sharp against a soft background, are worth a close look. The same check applies to text and signs pasted into a photo: the lettering often sits at a slightly wrong angle, with lighting and sharpness that do not match the surface it claims to be printed on.
Repeated textures: the clone stamp
The clone stamp copies pixels from one part of an image to cover another, most often to remove an object or a person, or to make a crowd or a garden look fuller. Look for the same patch of texture appearing twice: identical leaves, bricks, sand, or fabric patterns in places they should not repeat. Also look for smeared or oddly blurred patches in the background where something was painted over. If a background looks like it has been rubbed smooth in exactly one area, something was probably removed there.
Reflections, text, and orientation
Reflections are easy to forget and hard to fake well. A person standing beside water, a mirror, or a polished floor should have a reflection that matches their pose; a missing reflection, or one that shows a different pose, is a warning sign. Text is another classic slip: when a pasted element is flipped to fit, any writing on it flips too, so reversed or mirrored lettering is a near-certain sign of manipulation. The same applies to logos, badges, and number plates. If the text reads backwards, the image was flipped.
Mismatched sharpness and quality
A genuine photo is usually consistent in quality from one side to the other, allowing for depth of field. In a composite, the pasted element often comes from a different camera, so one part of the image looks sharp and high-resolution while the rest looks soft or noisy, or the other way round. A face that is much sharper than the clothing and background around it, or text that is crisper than everything else in the frame, suggests the parts were not photographed together. The same applies to colour: if one object is far more saturated than its surroundings, it probably came from a brighter photo.
Which signs prove a photo was edited, and which only suggest it?
Not all signs carry the same weight, and treating them as equal is the most common mistake people make. A single weak sign is a reason to look closer, not a verdict. Use the decision table below (the ranking asset for this guide) to weigh what you found before you accuse anyone of faking a photo.
| Sign | Strength | What it means |
|---|---|---|
| Reversed or mirrored text, logos, or number plates | Proves editing | Real cameras do not mirror text. The image was flipped or the element was pasted in. |
| Straight background lines bending around a body | Proves editing | Warping from a liquify-style tool. Walls and door frames do not bend on their own. |
| Identical texture patches repeated in the background | Proves editing | Clone stamping copies pixels. The same leaves or bricks cannot occur twice naturally. |
| A reverse search finds the unedited original | Proves editing | An earlier, different version of the same image is direct evidence of a change. |
| Editing software named in the metadata | Proves editing | The file records its own history. It proves an edit happened, not what it changed. |
| Shadows falling in different directions | Suggests editing | Strong evidence of a composite, but mirrors and multiple light sources can confuse it. |
| Halos or jagged edges around a person | Suggests editing | Typical of cut-and-paste work, but heavy compression can create similar artefacts. |
| Plastic-smooth skin with no texture | Suggests editing | Heavy retouching is likely, but phone beauty modes do this to genuine photos too. |
| One part sharper than the rest | Suggests editing | Points to a composite, but depth of field and focus can explain some of it. |
| Missing reflection where one should exist | Suggests editing | Worth checking, but reflections can be faint, cropped out, or hidden by angle. |
| No metadata in the file | Proves nothing | Social apps and messaging services strip metadata from ordinary photos. |
| Everything in focus at every distance | Proves nothing | Small phone sensors keep most of a scene sharp. It is normal, not suspicious. |
| The photo simply “looks odd” | Proves nothing | Wide lenses, flash, and unusual light make genuine photos look strange. |
Read the table as a rule of thumb: one “proves editing” sign is enough to say the photo was changed, but you still need the context to say whether the change deceives. One “suggests editing” sign is a prompt to run the reverse search below. And anything in the “proves nothing” row should never be used as evidence on its own. It only becomes meaningful alongside stronger signs.
How do you find the unedited original with a reverse image search?
Visual signs tell you a photo may have been edited. A reverse image search tells you what it looked like before. If the same image exists online in an earlier or cleaner version, the comparison settles the question in seconds, and it is the check the competitor guides describe only briefly. Run it before you spend money on any “photoshopped photo checker” app, because no detector beats seeing the original with your own eyes.
Here is the method, step by step:
- Save the best copy you can get. Use the largest, clearest version of the image. A small, re-compressed screenshot matches far worse than the original file, so download the photo itself where possible instead of screenshotting it.
- Search the whole image first. Upload it to Google Lens (our reverse image search guide walks through the desktop and phone steps) then check the “Exact matches” tab. An exact match on another site, dated earlier, is usually the original or a copy of it.
- Try a second engine. Upload the same image to TinEye and sort the results by oldest first, which surfaces the earliest indexed copy. Different engines index different parts of the web, so a photo Google misses may appear on TinEye or Yandex.
- Check the image’s history. Open Google About this image from the Lens results and read when Google first saw the image and which pages used it earliest. Google launched About this image in English in 2023 and expanded it to 40 additional languages in April 2024, according to Google, so the history panel is now widely available.
- Compare the versions side by side. Open the earliest copy and the version you were sent. Look for what changed: a person removed, a background swapped, a sign reworded, a body reshaped. Screenshot both, with their page addresses visible, so you have a record.
- Search a tight crop if the full image finds nothing. Crop to the most distinctive detail (a face, a sign, a building) and search that on its own. A crop sometimes matches when the full image does not, because it removes the edited area. But as the worked example below shows, a crop can also fail, so treat it as a second attempt, not a verdict.
Worked example: what actually happened when we tested this
We ran this method on 9 November 2026 with a real photograph, to show how the tools actually behave rather than how they should behave. The test photo was “Starting Point: Oeschinensee Lake Hike,” a turquoise alpine lake scene above Kandersteg, Switzerland, photographed by Nanda Gopal Lakshman and published on Pexels on 9 August 2024. We searched it on Google Lens, first as the whole photo and then as a tight crop, and cross-checked on TinEye.
From the whole photo, Lens found the original straight away. The exact Pexels photo page appeared in the results (as the third entry under “Exact matches,” in the visual matches, and in the web results dated 10 August 2024) and the results page identified the scene as Oeschinensee Lake above Kandersteg. That is the outcome you want: the original exists, it is dated, and you can compare it with the copy you were sent.
From a tight crop, Lens did not find it. We used Lens’s own crop tool to isolate a single mountain peak in the upper-right of the frame, excluding the lake, the shoreline trees, and the other peaks. The “Exact matches” tab then read “No exact matches found,” and the top results were only generic Swiss mountain photos from other sites. No Pexels page anywhere in the visible results. The location was still identified correctly, but the photo itself was not matched. The lesson is practical: search the whole image first, and crop only as a second attempt. A tight crop throws away the context the matching engine needs, so a “no match” on a crop proves nothing about the photo.
TinEye was the weaker check this time. Searching the same image, TinEye reported that it had searched 86.6 billion images and found one result (a sponsored stock photo on Unsplash, first indexed on 4 April 2026) and did not return the Pexels page at all. That does not make TinEye useless; its oldest-first sorting is still the best way to date a copy when it does match. It is a reminder to use more than one engine and to read what each one actually returned, rather than assuming a single “no results” means the photo is original. If you need to pin down when a photo first appeared, our guide on how to work out when a photo was taken covers the dating methods in full, and our step-by-step reverse image search guide covers the phone and desktop steps for every engine mentioned here.
What can the photo’s metadata tell you, and what can it not?
Every digital photo carries a block of hidden information called EXIF metadata: the camera model, the date and time, sometimes the location, and (usefully here) the software that last saved the file. If the metadata names Photoshop, Lightroom, or another editor, the photo was edited. Where the edit used Photoshop’s Generative Fill, the file carries a different record again: our guide to telling if an image was made with Adobe Firefly explains what that record shows. That is a fact, not a guess. If the metadata is missing entirely, it proves nothing at all: Facebook, Instagram, WhatsApp, and most websites strip metadata automatically when a photo is uploaded, so an ordinary genuine photo from social media will usually show nothing.
Three cautions apply. First, metadata is easy to edit, so a file that claims to be untouched may have been rewritten. Treat it as one piece of evidence, never the whole case. Second, editing software in the metadata tells you an edit happened, not what changed; a photo can pass through Photoshop for a simple crop and still be honest. Third, never rely on the file’s “modified” date as the date the photo was taken. Every save rewrites it. For the full method, including how to read the date and camera fields properly, see our guide to telling whether a photo is real or fake, which puts the metadata check alongside the other six verification methods.
What is Error Level Analysis, and when should you use it?
Error Level Analysis, or ELA, is a forensic technique that highlights parts of a JPEG that were compressed differently from the rest, which is what happens when one area of a photo is edited and the file is saved again. The free tool FotoForensics produces an ELA map of any uploaded photo, and areas that stand out brightly against an otherwise even map deserve a closer look. That is the whole method in one paragraph, deliberately: ELA maps are easy to misread (high-contrast edges, phone portrait processing, and heavy recompression all create bright patches that look like edits but are not) so reading one properly needs its own guide, and a dedicated FotoForensics guide in this series covers exactly that. Use ELA only after the visual checks and the reverse search above, never instead of them, and never treat a bright patch as proof on its own.
How do you decide what to believe?
Once you have looked, searched, and checked the metadata, decide on the balance of the evidence rather than on any single test. That is the real skill in learning how to tell if a photo is photoshopped: weighing several small pieces of evidence together instead of trusting one test. Two or more independent signs pointing the same way (for example, a bent door frame plus an earlier unedited version found by reverse search) is a solid basis for saying the photo was edited. One weak sign, with a clean reverse search and no original found, is a reason to stay cautious, not a reason to call the photo fake. And if the reverse search finds the photo under a different name, on a stock site, or in an old scam report, the question stops being about editing at all: the photo is stolen, and the person using it is not who they claim to be.
That last case is the one that matters most in practice. Most fake profile photos are not photoshopped at all. They are real photos taken from someone else. A reverse search exposes those in seconds, which is why it comes before every other check in this guide. For the wider picture, including AI-generated images and photos used out of context, see our guide to how to tell if a photo is real or fake, and AAP FactCheck’s guide to fact-checking photographs, which covers the same three stages (creation, editing, and distribution) from a fact-checker’s point of view.
Finally, remember what the checks cannot do. No method in this guide can prove a photo is genuine. A skilled editor leaves no visible signs, a private photo has no history to search, and a clean metadata record can be faked. What the method can do is far more useful: it catches most edited photos, it finds the original when one exists, and it stops you from trusting (or sharing) a photo that fails its checks. When in doubt, do not share it, do not pay based on it, and do not accept it as proof of identity.
FAQs
Is this picture photoshopped or just filtered?
A filter changes the whole image at once (colour, brightness, contrast) while Photoshop-style editing changes one part of it, like removing an object or reshaping a body. If only one area looks different from the rest, or the background bends near the subject, it is edited, not just filtered. A filter alone rarely deceives; a local edit can.
Can you tell if a photo is photoshopped from a screenshot?
It is harder, because screenshots compress the image and strip its metadata, which blurs the fine edges and textures the visual checks rely on. You can still look for warped backgrounds, mismatched lighting, and repeated textures, but the reverse search is weaker too. Ask for the original file whenever the photo matters.
Does a photo with no signs of editing prove it is genuine?
No. A skilled editor can make changes that leave no visible trace, and some edits (a simple crop, a colour change) leave no forensic trace at all. Absence of signs only means you found nothing, not that nothing was done. Always check the photo’s history and source as well, and treat a photo with no history as unverified.
What is the fastest free way to check a photo?
Run a reverse image search first: upload the photo to Google Lens and check the exact matches, then try TinEye sorted by oldest. That takes about two minutes and finds the original when one exists. Then zoom in and check the background lines, shadows, and edges. Those two steps catch most edited photos without any paid tool.







