What shoes are these?

ShirtMath does not identify shoes from a photograph, and it will not pretend otherwise. What does work is a short chain of checks, in order: reverse image search, then the style code printed inside the shoe, then the sockliner and outsole markings, then a community identification forum when the first three fail.

Can a photograph identify a shoe on its own?

Sometimes, and only when the picture is good enough to match against an indexed image of the same product. A clear side-on shot of a current mass-market model in daylight is the easy case. A cropped, low-light photo of a worn pair in a regional colourway from several years ago is the hard case. Reverse image search will still return something for the hard case, and that something is often a visually similar shoe rather than the right one. Treat the first result as a candidate and nothing more until a printed code on the shoe itself agrees with it.

The reason is worth stating plainly. Image matching compares your photo to pictures that already exist online. It has no access to the label, the article number, the factory code or the year, so it can only ever tell you which product photos look like your photo. Two shoes can share a silhouette, an outsole pattern and a colour blocking and still be different models, different generations of the same model, or a licensed regional variant.

What actually works, in order?

1. Reverse image search the cleanest photo you have — crop to the shoe, shoot side-on against a plain background, and search the image rather than describing it in words. Note the candidate name, do not commit to it.
2. Read the size label inside the shoe — on most branded footwear the label stitched inside the tongue or along the collar lining carries a style or article code as well as the size. That code is the single most reliable identifier on the shoe.
3. Check the box if you still have it — the box end label normally repeats the same style code, the colourway name and the size run. Boxes get separated from shoes, which is exactly why the inside label matters more.
4. Photograph the sockliner and the outsole — the printed insole often carries the model name, and the outsole carries mould markings, size stamps and sometimes a factory code. These survive when the tongue label has worn away.
5. Post to a community identification forum last — enthusiast communities identify obscure and discontinued models well, but only if you give them the outsole, the sockliner, the heel and the inside label, not one blurry side shot.

Where is the style code printed, and what does it look like?

Formats differ by brand and change over time, so read the code off the shoe rather than assuming a pattern. Nike prints a style code on the inside size label, written as a block of digits, a hyphen, then a short suffix; the suffix identifies the colourway rather than the model. Adidas prints an article number on the label and repeats it on the box. New Balance builds the model number into the name itself and prints a fuller code on the label. Whatever the brand, the code on the label is what a retailer or the brand's own product pages can be searched against.

Two cautions. A code identifies the product as it left the factory, so a general release and a shop-specific make-up can look identical in photos and differ in code. And counterfeits copy labels, including codes, which means a code that returns nothing on the brand's own site is a signal in itself.

What do the sockliner and outsole tell you that the outside does not?

The exterior of a shoe is the part designers change most often, and it is the part copied most often. The parts nobody styles are more diagnostic. Pull the insole out and photograph both faces: model names, technology names and sometimes a size or region code are printed there. Then photograph the outsole flat. Mould geometry, the position of flex grooves and the stamped size block are all specific to a generation of a model, so they separate a current version from the one that came before it. Those two photographs plus the inside label are what an identification forum actually needs.

Can ShirtMath identify or size a shoe from a photograph?

No, and that boundary is deliberate. ShirtMath measures garments laid flat and compares those measurements against published garment specifications. A photograph alone cannot produce a verified shoe size or a fit claim, because shoe fit depends on the length and width of your foot inside a last whose internal shape is not visible from outside the shoe. Getting that right needs one of three things: a size you already own and trust, a foot measurement you have taken yourself, or verified sizing published by the retailer.

What ShirtMath does today is narrower and verifiable. Photograph a shirt laid flat with a scale card and it returns that shirt's flat measurements, then matches it against published garment measurements across the catalogue. Roughly half of scanner readings land within an inch of a tape measure, sleeve opening and bicep are the least reliable readings, and anything the scanner cannot read confidently is left blank rather than filled with a guess.

How do you get a shoe size you can trust?

Measure the foot, then read the brand's own chart rather than a generic conversion. Our walkthrough for doing that at home is in how to measure shoe size at home, and the conversions between regional systems are set out in the shoe size chart. If your problem is a shoe that is the right length and still wrong across the forefoot, width is the variable to check next, which is covered in the shoe width chart.

Two practical notes. Feet are usually not the same size as each other, so fit the larger one. And foot length changes across the day, so measure in the evening rather than first thing.

Why does the same size label mean different things?

Footwear is not unusual here. The same drift shows up in clothing, where it is easier to demonstrate with published numbers. A unisex Gildan G500 in Medium publishes a 20.0" flat chest, while the ladies' cut G500L in Medium publishes 17.5" — a two and a half inch gap on the same letter, widening to a clean 3.0" at Large, where the same two styles publish 22.0" and 19.0". The label is a name for a garment, not a measurement of your body.

The same logic is why a shoe identified correctly still needs a size checked against that brand's chart. For the clothing version of this problem, see vanity sizing.

Common questions

Where to go next