Creating ghost mannequin photos with AI: a practical guide

Quick answer: To make a ghost mannequin photo with AI, shoot the garment once, on a mannequin, on a hanger or laid flat, and upload that photo to an AI ghost mannequin tool. Photta's AI Ghost Mannequin removes the mannequin and returns the garment in 3D, as if an invisible body were wearing it, or as a flat lay, in a front, side or back view (a back view needs a photo of the back as well). You can try it free on your own garment photo, no credit card needed: photta.app/tools/ai-ghost-mannequin-free.

What the ghost mannequin effect is

A ghost mannequin image, also called invisible mannequin or hollow man, shows a garment as if an invisible person were wearing it. The shoulders and sleeves keep their shape and the inside of the neckline is visible, without a plastic mannequin or a model in the picture.

Traditionally the effect is built in retouching: Pixelz describes an invisible mannequin image as "a combination of two or more images", with the garment shot on a mannequin, the hidden interior shot separately, the mannequin masked out and the two composited (Pixelz, updated September 2025). With AI, one photo of the garment is enough to start, and a second photo of the back improves the interior. This guide walks through it step by step with Photta, and every setting it names is taken from Photta's public documentation.

Illustration of the idea: a garment photographed on a mannequin, and the same garment presented as a ghost mannequin image.

What you need

The garment, steamed and styled the way it should be shown.
A mannequin, a hanger or a flat surface. A standard mannequin works; so does a hanger, or laying the garment flat.
A phone or camera and soft, even light.
A Photta account to run the ghost mannequin tool.

Why shoppers need more than a flat lay

Shoppers look at the pictures before anything else: Baymard Institute found that 56% of users' first action on a product page is to explore the images (Baymard, 2 April 2020). For clothing, those images have to show shape and fit.

Flat lay photography, where clothes are laid on a table and shot from above, suits social media, but it cannot show fit and drape. When a hoodie looks flat, a shopper cannot picture how it sits on the shoulders or how the fabric falls at the waist. The ghost mannequin effect shows that structure without a model.

Illustration comparing a flat lay with a ghost mannequin presentation of the same garment.

Step 1: Prepare and capture your product

The quality of the output depends on the photo you start from. A few preparation steps make the biggest difference.

Use a mannequin or a flat surface

You have a few options. You can use a physical mannequin, even a cheap plastic one, hang the garment on a hanger, or lay it flat. Photta's tool takes all three; its API documentation describes the garment photo as "on a mannequin or flat" (Photta API docs). On a mannequin, make sure the garment is centred. For a flat lay, style the garment by slightly tucking the waist to give it a more natural shape.

Lighting is key

Avoid your camera's direct flash. Place the product near a large window for soft, natural light, or use two simple LED softboxes. The goal is to avoid harsh shadows that hide the texture of the fabric.

Shoot the back as well

The hardest part of the effect is the inside of the garment, which the mannequin hides. Photta takes an optional photo of the opposite side, which its documentation says "helps the AI reconstruct the interior of the garment" (Photta API docs). Take one photo of the front and one of the back, with the same framing and light.

The ideal setup for a ghost mannequin photo: centred, well lit, and on a neutral background.

Step 2: Upload your photos

Open the ghost mannequin tool in Photta and upload the front photo, then the back photo if you have one. According to the documentation, the main photo can show either the front or the back of the garment.

Step 3: Choose your settings

These are the options Photta documents for a ghost mannequin image (Photta API docs, checked 24 September 2026):

Style: 3d keeps the garment's three-dimensional shape; flat_lay lays it flat with the mannequin removed.
Angle: straight, back or side. A side view is reconstructed from the front photo, plus the back photo when you provide it.
Background: white, gray, beige or black.
Resolution: 2K or 4K. Choose upfront: 4K is generated natively and cannot be changed afterwards, and it cannot be combined with a square 1:1 image.
Aspect ratio: 3:4, 1:1, 4:3, 9:16 or 16:9.
A short note, up to 400 characters, for ambiguous choices such as "closed front" or "cream interior lining".
Brand label: a saved label can be applied to the garment.

Many marketplaces want the main image on a plain white background, but not all: Zalando, for example, accepts plain, lifestyle or ambient settings and applies its own light grey background to primary packshots submitted on white (Zalando image guidelines, updated 1 October 2026, checked 18 October 2026). Check each channel's own rules before you choose.

Cost: each image costs 4 credits at 2K and 5 at 4K (Photta API docs, checked 24 September 2026).

Step 4: Generate and review

Generation runs in the background; the documentation lists a typical completion time of 1.5–4 minutes per image (Photta API docs, checked 24 September 2026), so you can prepare the next garment meanwhile.

Then compare the result with the real garment at full size:

Neckline and interior: the inside of the collar looks like the real one, with no doubled or missing parts.
Shape: shoulders even, sleeves straight, hem where it should be.
Details: buttons, pockets, seams and zips are all present and in the right place.
Labels and prints: identical in shape, text and position.
Colour: unchanged compared with the product under neutral light.

If the garment itself has changed, generate again rather than retouching the output: a missing or invented detail is a product error.

Step 5: Upscale and export

Hero images are the ones shoppers zoom into. Baymard found that low resolution and weak zoom led to product abandonment in its testing (Baymard, 2 April 2020). For those images, generate at 4K from the start. For images already generated at 2K, Photta's AI Upscale tool can enlarge them before export.

Illustration of the ghost mannequin and upscale steps.

Common mistakes to avoid in ghost mannequin photography

A few mistakes in the photo hold back any result:

Wrinkled clothing: deep wrinkles look messy in a ghost mannequin shot. Use a handheld steamer on the garment after it is on the mannequin to get a smooth silhouette.
Tilted angles: shooting from too high or too low makes the perspective look off. Keep the lens at chest level for shirts and waist level for trousers.
Cluttered backgrounds: a busy or strongly coloured background can reflect onto the edges of the garment. Shoot against a plain white or light grey wall.
Incorrect sizing: a mannequin that is too large stretches the fabric. A slightly smaller mannequin, with clips at the back to pull the fabric taut, looks more natural.

Beyond the ghost mannequin

The same garment photos can be used for other image types:

On-model images: Photta's apparel try-on puts the garment on an AI model you choose, including your own models. An image costs 5 credits at 2K (Photta API docs). Check it against the garment like any generated image.
Your own model: Model Maker builds a model from attributes such as age, body type and hair, for 4 credits (Photta API docs), and the same model can be reused across a collection.
Batches: for a rack of garments, the batch tool takes one set of settings for all of them: 5 garments per batch on Free and Hobby, 10 on Seller, 25 on Brand and 50 on Studio (Photta API docs, checked 24 September 2026).

Illustration of on-model images made from garment photos.

Cost: AI versus traditional

Compare the two routes with published figures rather than estimates.

Traditional route: Shopify's 2026 pricing guide puts professional product photography at $50 to $350 per image, and photographer day rates at $500 to $3,000 (Shopify, 27 January 2026, checked 24 September 2026). A ghost mannequin image also needs a retoucher to build the composite.

The AI route: 50 ghost mannequin images at 2K use 200 credits. That fits within the Seller plan's 500 monthly credits, listed at $29 a month in US dollars (Photta pricing, checked 24 September 2026). You still take the photos yourself, and every output needs the review in Step 4.

Whichever route you choose, compare the cost per approved image, including your own time and any retries.

Frequently asked questions

What is the ghost mannequin effect?

It is a way of presenting clothing so that it looks worn by an invisible person: the mannequin is removed and the inside of the garment is shown. Traditionally it is a composite of two or more photos; AI tools generate it from one photo, and a second photo of the back improves the interior.

Do I need a special ghost mannequin?

No. Mannequins with removable parts make the traditional route easier, but Photta's tool accepts a garment photographed on a standard mannequin, on a hanger or laid flat.

Can I use a real person instead of a mannequin?

Photta's tool is meant for a garment on a mannequin, on a hanger or laid flat, so use one of those. If you want the garment on a person, use apparel try-on instead.

How many credits does one ghost mannequin image cost?

4 credits at 2K and 5 at 4K. Credits are charged when the generation starts and refunded automatically if it fails.

Does it work with patterned or sheer fabric?

Patterned garments work, but check that the pattern lines up across seams in every output. Sheer, lace and heavily embellished fabric is harder for any tool, because the material is ambiguous to reconstruct; review those images especially closely.

How do I get a white background for marketplaces?

Choose the white background in the settings. Then check the marketplace's own image rules, which also cover size and framing; our guide to product image requirements by marketplace collects the published rules with their sources.

Is there a free AI tool for ghost mannequin photos?

You can try Photta's AI Ghost Mannequin free on your own garment photo, with no credit card needed: upload one photo, choose 3D ghost or flat lay, and generate. See the pricing page for the plans once you want to make more.

Can I make a ghost mannequin photo from a single photo?

Yes. One clear photo of the garment, on a mannequin, on a hanger or laid flat, is enough to start with Photta's AI Ghost Mannequin. A photo of the back and one of the inside are optional: the back photo is needed for an accurate back view and helps when the back differs from the front, and the inside photo helps when the inside differs from the outside.

How do I check a ghost mannequin photo before I publish it?

Hold it next to the real garment. Check the inside of the neckline, that the sleeves and hems match on both sides, that the print runs across the seams, that the label sits where it should and that the colours are true. If something is off, add or replace the photo of the back or the inside and generate again.

Illustration of a consistent ghost mannequin gallery for a collection.