When to use it
The registration system wants a one-inch photo, a CV needs a two-inch photo, and social security and exams want a white-background photo, but all you have is a phone selfie — this is the most common entry point. The tool finds the face first, then crops to the correct aspect ratio for the specification, and you only fine-tune the head and shoulders.
It is also worth using when the photo already exists but the background colour is wrong. If yours is blue and the system wants white, swapping the background beats retaking it; if the photo is acceptable but the head is too small or sits too low, zoom and vertical offset can rescue it. It cannot help with wrong exposure, closed eyes or a face hidden by hair — retaking is faster.
Before you start, confirm the requirements of the receiving authority. Size, background colour, head proportion and whether glasses are allowed differ by country and agency, and they change. This tool only crops by millimetres and swaps the background; it does not judge compliance, so follow the receiving party's latest requirements.
How to choose the settings
- One inch is 25×35 mm, which is 295×413 pixels at 300 DPI, the most common size. Two inch is 35×49 mm, or 413×579 pixels; small one inch is 22×32 mm, or 260×378 pixels; small two inch is 35×45 mm, or 413×531 pixels; three inch is 55×84 mm, or 650×992 pixels.
- Resolution of at least 300 DPI; below 200 DPI you can see grain when printing enlarged. But the pixel count is the hard measure: DPI is only a convention for converting millimetres to pixels, and the same 295×413 file gains no detail whether it is labelled 300 or 350 DPI.
- The zoom slider runs from 60% to 180%. ID photos usually require the head to be 60–70% of the frame height; too large crops hair or shoulders, too small leaves lots of space around and makes the head look small. Start at 100% and adjust in steps of 10%.
- A vertical offset within plus or minus 40 is enough, and placing the eyes about 40% down from the top is stable. The face box is roughly from the eyebrows to the chin, and the tool estimates head and shoulders by multiplying it by 3; when the forehead is covered by a fringe or a cap brim, automatic framing drifts, so pull it up manually.
- Choose white, blue, red or grey for the background according to the receiving party's requirements. What really affects the result is the original background: the cleaner it is and the more it contrasts with the clothing, the cleaner the background swap; a patterned wall, backlighting and shadows on the wall all stay at the edges.
- The export is PNG, and the size is the pixel values above with no further compression. If the system limits file size, look at the cap it gives first and then use an image compression tool; do not lower quality at the crop-and-replace-background step.
Compared with other approaches
Compared with a photo studio, the difference is the lighting. They have softboxes, backdrop paper and a retoucher, so the output is consistent; the cost is a trip and a wait. If you just need one electronic photo that passes review, cropping it yourself is faster. The shortcomings should be stated too: no beautification, no blemish retouching, no fill light.
Compared with professional image software, controllability is a step behind. Photoshop and the like can trace edges by hand and rescue hair with masks. This tool strings face detection, proportional cropping, cutout, background swapping and millimetre-based export into one pipeline that gives a result in one step, at the cost of only being able to adjust zoom and vertical offset.
Compared with online ID photo services, there are trade-offs on both sides. Those services have many templates, and some include manual review and printing; the cost is that the photo has to be uploaded to their server. This tool computes entirely in the browser, the photo is not uploaded and the result stays on this machine, which is one less place for a face-containing ID photo to be retained; the cost is no manual review or compliance check, and no server to help, so the cutout takes a few seconds longer on an old machine.
Additional Q&A
What size should I output for printing?
Export at the pixel values above, that is 300 DPI, and hand it to a photo lab to output at the actual size. Printing several on 6-inch paper is layout work; this tool only produces a single photo. If the pixel count is too low and you force it bigger for printing, the face goes blurry.
Why do phone selfies need headroom above the head?
ID photos require the head to take a fixed proportion of the frame, with background above the top of the head. Selfies tend to fill the frame with the face, so after cropping there is no headroom and the head is too large. When shooting, hold the phone level with your eyes and step back half a pace so there is more background above the head; too close also makes the nose look big.
After swapping the background, why do the edges look grey and carry the old background colour?
Light from the original background reflects onto the face outline and the gaps between hairs, and the cutout marks those pixels as semi-transparent, so the new background shows through greyed; the model also computes on a small 320×320 image and scales it up, leaving a one- or two-pixel transition band at the edge, most visible when a dark original background is replaced with white. To reduce it, keep the subject farther from the background, light evenly, and wear clothing with strong contrast; for photos already taken, switching back to a similar colour family is safest.
Are the requirements the same for an electronic photo and a printed one?
No. Registration systems mostly check only pixel dimensions and file size, and some cap it at a few tens of KB; printing depends on millimetre size and resolution. For the same specification you may need to export once for upload and once for print.