Use case: on-model photos for a fashion store, without a photo shoot

By Manuel Mattar4 min read

In short

This is an example scenario. A fashion store with flat-lay photos of its garments creates a fictional model in the Studio, generates photos where the model wears each garment based on the real photos, reviews them one by one and keeps the good ones. Every generated image is labeled as created with AI.

This case is an example, not a customer: an imaginary store with scenario numbers to show how the Studio is used from start to finish. The credit figures are the ones in today's help guide; the catalog figures are made up for the example.

The situation

An online fashion store sells 120 garments: blouses, dresses and trousers. Its suppliers send flat-lay photos, well lit, on a white background. They show the print, but not how the fabric falls, how long a dress is when worn or how the neckline looks. An on-model photo shoot for 120 garments means a studio, a photographer, a model, make-up and weeks of waiting, and every season it starts over.

The team wants on-model photos for the product pages, without inventing a product that does not exist and without hiding that the photos are generated.

Step 1: create the model once

An administrator goes to Media → The Studio and describes the model through a casting: gender, apparent age, origin, build, skin tone, eye color, hair, make-up and look, plus a few details in text. Bruxel generates a five-view session —front face, three-quarter, profile, half body and full body— to check face and hands calmly.

The first proposal does not convince her, so she asks for another; she can ask for up to two at no extra cost. When she likes it, she uses "Accept this model" and the model is ready for every photo. Creating a model costs 100 credits, once. It is a fictional person: when creating it, the store confirms it is not trying to recreate anyone real.

Step 2: the photo session for each garment

On each product page, below its images, there is the Photo session. An operator picks the model, one to three real photos of the garment as reference and, above the references, the "Clothing" option: the model then wears the garment instead of holding it, and fidelity is kept in the fabric, the print, the cut and the length. Bruxel preselects it when the name, the category or a size attribute gives it away.

Then the operator picks the expression —natural, smiling, laughing, neutral, serious or playful— and, optionally, the hairstyle and make-up for that photo. In "Clothing" there is no outfit to choose: the garment is the outfit. The team asks for two proposals per session. Each proposal costs 10 credits, the price is shown on the button before generating and, if the session fails, the credits come back.

To avoid going page by page, the team selects several garments in the catalog and generates the clothing photos in a batch, with one model and one expression for the whole set. They can leave the page: the bell notifies them when the photos are ready.

Step 3: review before saving

Proposals arrive in "Proposals from your sessions". The team rule is to check two things before saving: that the print and the cut match the real garment, and that the hands look right. Each proposal has two actions:

  • Save to the product, which links it to the product page.
  • Discard, which sends it to the image trash for 30 days.

Discarding does not return credits: the cost was for generating. That is why the team first tries ten garments and adjusts the expression and the references before launching the rest.

Step 4: the main photo stays real

An on-model photo cannot be the main image. For garments whose supplier photo has a messy background, the team uses the "Main photo" tab: it picks a single real photo and Bruxel generates its studio version, with a clean background and no people, for 5 credits. It is the only proposal that offers "Use as main".

The scenario's numbers

With 120 garments, two proposals per garment and a main photo for half the catalog, the scenario adds up to: 100 credits for the model, 2,400 for the on-model photos and 300 for the main photos. These are numbers for this example; Studio spending counts toward the organization's spending cap per period, just like descriptions. The credits guide explains how to buy them and how to set that cap.

What gets solved and what does not

Solved: the store has on-model photos for its garments, the same model across the whole catalog, and every image declared as "Created with AI" in the app and in the metadata of the original file, with the synthetic person mark.

Not everything is solved:

  • One garment per photo. For an outfit you generate one photo per garment or start from a real photo with the full outfit.
  • Accessories take another path. Glasses, bags, jewelry and shoes go under "Product", with a direction such as "wearing it".
  • Quality depends on the real photo. The better the references, the more faithful the garment; a generated photo cannot be used as a reference.
  • The Studio is enabled per account. If it does not appear in the menu, you ask support for it.

The full Studio guide covers five-second videos, extended video and 3D objects. To see where it fits in the product, check the Studio and catalog pages.

Frequently asked questions

Can I ask for the model to look like a famous person?

No. Studio models are fictional people. When you create one you confirm you are not trying to recreate anyone real, and a description that names a real person is rejected before anything is charged.

Can an on-model photo be the product's main image?

No. The main image always comes from a real photo. That is what the Main photo tab is for: it generates a studio version with no people from a single real photo.

How does my customer know the photo was generated?

Every image Bruxel generates carries the Created with AI badge in the app and metadata embedded in the original file that declares it as generated and, when a model appears, as a synthetic person.

Sources

  1. Synthetic media (Wikipedia)
  2. Fashion photography (Wikipedia)

About the author

Manuel Mattar

Founder of Bruxel.AI

In 2014 he founded Wolf Sellers, an e-commerce agency in Mexico City, and later Bruxel.AI, the AI-powered PIM.

See all posts by Manuel Mattar

Try it with your catalog

Create your free account and try Bruxel with your own products using the welcome credits.

Keep reading

Use case: on-model photos for a fashion store, without a photo shoot