Eraslino

Eraslino › Guides

AI object remover test: how big an object can you erase?

By Eraslino · Published 2026-11-15 · Updated 2026-11-15

Short answer: an AI object remover works best on objects that cover a small part of the frame. We erased the same red ball at four sizes from a lake photo. Under about 1.5% of the frame the AI fill was hard to tell from the original; at 5% it was plausible but wrong in the details; at 10% it redrew the reflection in the wrong place.

The error grew steadily with size: 11.9 → 17.4 → 23.3 → 28.3 (0–255 scale) as the selection grew from 0.27% to 10.3% of the photo.

"Remove any object" is the promise on most object-remover pages. What they rarely show is how quality changes with the size of the thing you remove. We tested exactly that, with an answer key: we placed the ball on the photo ourselves, so we know what the lake looks like underneath.

How we tested

Eraslino with the lake photo and a large red ball covered by a round red selection
The 256 px ball selected in the eraser. One dab of a brush slightly bigger than the ball is the whole selection.

Results

BallSelected pixelsShare of photoSmart fill errorQuick patch errorDetail (smart / quick)
48 px3,2760.27%11.913.70.69 / 0.11
128 px16,4461.34%17.433.70.55 / 0.07
256 px58,3924.75%23.327.20.67 / 0.05
384 px126,15610.3%28.327.40.70 / 0.04

Fill time: 0.8 s for every smart fill (model cached) and at most 0.1 s for quick patch, Chrome on an Apple M4 Pro. The model pipeline crops around the hole and works at 512 px internally, which is why its time barely changes with size.

What each size looks like

The 48 px ball: before, smart fill, quick patch, original
0.27% of the frame: both methods are close. At normal viewing size you would not find the spot.
The 128 px ball: before, smart fill, quick patch, original
1.34%: smart fill continues the tree reflections downward and is hard to spot; quick patch turns into a pinwheel of streaks, its worst result in the test (33.7).
Open the 128 px example →
The 256 px ball: before, smart fill, quick patch, original
4.75%: smart fill still produces lake and reflections, but the edge between the dark reflection and the bright water is softer and in a slightly different place.
The 384 px ball: before, smart fill, quick patch, original
10.3%: smart fill draws mountain-like reflection where the forest reflection should be and breaks the shoreline. From a distance it passes; up close it is a different lake.
Open the 384 px example →

Five things the test shows

  1. Size is the main factor. Error more than doubled from the smallest to the largest ball, with the same photo and the same position.
  2. Small objects do not need AI. At 0.27% of the frame quick patch was within 2 points of smart fill and needs no model download.
  3. Straight lines through the hole give a large fill away. The shoreline and the reflection edge are what broke first. Before you erase, look at what lines cross the object.
  4. A good error score is not the same as a good look. At 384 px the two methods scored about the same (28.3 and 27.4), but quick patch is an obvious smear and smart fill a believable lake.
  5. The model's time is flat. Because the MI-GAN pipeline works on the area around the hole at 512 px internally, a large object cost the same 0.8 s as a small one; what grows with size is the amount the model has to invent, not the time.

Removing people

A person in a photo is just a larger object with more lines around it, so the same rules apply: a distant figure covering 1% of the frame is an easy erase; someone filling a tenth of the frame leaves the model to invent everything behind them. Erase the person and their shadow together, and if they overlap the main subject, the background behind the overlap will be guessed. Only edit photos of people in ways they would be fine with.

Sources

Keep going