· 4 min read
How to Remove an Unwanted Object From a Photo
Heshan Fernando
Co-founder & COO
A good landscape photo with a bin in the corner. A family picture with a stranger walking through the edge of frame. A product shot with a cable you didn’t notice until you were back at the desk.
Removing them is genuinely achievable now, and the results range from invisible to obviously wrong depending on one thing: what’s behind the object.
What removal actually does
There’s nothing behind the object in the file. The pixels simply don’t exist. So removal means inventing them — sampling the surrounding area and generating something plausible to fill the gap.
That works well when the surroundings are consistent. Grass, sky, sand, a plain wall: the fill only needs to match texture and colour, and the eye has no reference to check it against.
It works badly when the region crosses structure. A straight edge that should continue through the gap, a repeating pattern like brickwork or tiles, text, or a face — these have rules the fill has to satisfy, and any error is immediately visible because the viewer knows what should be there.
Size compounds it. A small gap has lots of adjacent context and little area to invent. A large gap has the opposite, and the fill gets progressively less convincing as the region grows.
Why people get stuck here
- Oversized selections. Painting generously over the object gives the fill more area to invent than it needs.
- Structured backgrounds. Removing something from a tiled floor or against a horizon line exposes the fill immediately.
- One big pass. Attempting a large removal in a single step, rather than working inward in stages.
- Edges. The boundary between original and filled pixels is where the eye lands first.
What good removal looks like
A tight selection
Cover the object and a few pixels beyond, no more. Every extra pixel selected is a pixel that has to be invented.
Several small passes
Removing a large object in stages — working inward from the edges — consistently beats one large selection, because each pass has more real context to sample from.
Checked along the structure
After the fill, follow any straight line, horizon, or repeating pattern through the region. That’s where a fill gives itself away, and where a small manual correction is usually enough.
| Background | Result | Why |
|---|---|---|
| Grass, sky, sand | Usually invisible | Texture only, no structure |
| Plain wall | Good | Consistent colour and grain |
| Brick, tiles, railings | Difficult | Pattern must line up |
| Faces, text | Poor | Viewer knows what’s missing |
Common mistakes to avoid
- Selecting far more than the object, which makes the fill’s job harder for no benefit.
- Removing something that overlaps the subject, where an error is right where the eye goes.
- Working on a heavily compressed JPEG, where the fill inherits and spreads the existing artefacts.
- Accepting the result at screen size without checking at 100%.
- Removing something from a documentary or journalistic image, where alteration is a different question entirely.
How to do it with Photo Object Eraser
The Photo Object Eraser processes the image in your browser, so the photo stays on your device.
- Upload the photo and paint over the object you want gone.
- Keep the selection tight — a few pixels beyond the object is enough.
- For a large object, work in several smaller passes rather than one.
- Check any straight lines or patterns running through the region before accepting it.
Other browser-only photo tools are in the tools directory.
Frequently asked questions
What kinds of object remove cleanly?
Small objects against uniform backgrounds — a bin on grass, a bird in sky. Anything crossing a straight edge, a repeating pattern, or a face is much harder.
Why does the fill look smudged?
Because it’s invented from surrounding pixels. The bigger the region and the more structured the background, the more obvious that becomes. Removing in several small passes usually beats one large one.
Is my photo uploaded?
No. Image handling happens in the browser, so the file stays on your device.
Final thought
Look at what’s behind the object before deciding whether to remove it. Plain texture removes invisibly; anything with structure will need a human eye afterwards regardless of the tool.