Fashion production glossary

Learn the language of fashion production

Read 24 direct definitions for photography, AI production, image quality, ecommerce workflows, motion, and provenance.

01

Choose a photography method

Compare the standard ways fashion teams document a garment for catalogs and product pages.

  1. 01Ghost mannequin photographyGhost mannequin photography is a product-imaging method that shows a garment in a filled, three-dimensional shape while removing the mannequin or model from the final image. Separate frames are often combined so inner details such as the neckline, collar, or waistband remain visible.
  2. 02Flat lay photographyFlat lay photography is an overhead imaging method in which a garment or product is arranged on a flat surface and photographed from above. It presents the item without a visible model or mannequin, making silhouette, color, and surface details easy to compare.
  3. 03On-model photographyOn-model photography is fashion imagery that presents a garment worn by a human model. It gives viewers visual context for fit, proportion, drape, styling, and scale that product-only photography cannot show on its own.
  4. 04Pack shotA pack shot is a clear product-only image made to document an item accurately for a catalog, product page, marketplace, or sales material. It usually uses a controlled background, consistent framing, and minimal styling so the product remains the primary subject.
  5. 05Hollow man effectThe hollow man effect is the finished visual result of removing a mannequin or body support from an apparel image while preserving the garment's worn shape and visible interior. It is also commonly called the ghost mannequin or invisible mannequin effect.
  6. 06Ghost mannequin vs on-modelGhost mannequin imagery shows a garment in a three-dimensional worn shape without a visible person, while on-model imagery shows the garment on a human model. The first isolates product construction; the second adds fit, proportion, styling, and human context.

02

Direct AI image production

Understand the reusable inputs and workflows behind consistent generated fashion imagery.

  1. 01AI fashion modelAn AI fashion model is a synthetic human likeness created or rendered with generative AI for use in fashion imagery. The model can be directed through references and production instructions, while the clothing still needs separate checks for fit, construction, color, and detail.
  2. 02Virtual modelA virtual model is a digitally created or digitally represented person used to present clothing or accessories. The model may be a 2D composite, a 3D avatar, a reusable digital identity, or an AI-generated likeness, so the term is broader than AI fashion model.
  3. 03AI art directionAI art direction is the practice of turning a creative brief into explicit, reusable instructions for image or video generation. It can define the model, framing, pose, lighting, background, styling, references, product constraints, and review criteria for a production run.
  4. 04Colorway generationColorway generation is the creation of visual variants that show the same product in different intended colors. The process changes specified color regions while aiming to preserve the garment's design, construction, material appearance, fit, and surrounding scene.
  5. 05Batch image generationBatch image generation is the creation of many images through one structured production job rather than a series of unrelated manual requests. Each row or record can supply product inputs and variables while sharing art direction, output settings, and review rules.
  6. 06Production-scale image generationProduction-scale image generation is the repeatable creation of approved visual assets across many products, variants, or channels using structured inputs and controlled workflows. It combines reusable direction, generation, optional automatic QA, retries, approvals, and traceable delivery.

03

Measure image quality

Name the checks that protect garment details and make review criteria explicit.

  1. 01Automatic image QAAutomatic image QA is an agent-driven review process that checks generated or edited images against configured requirements. When enabled, it can flag exceptions, assign review states, or trigger retries; teams can turn it off and keep final approval human.
  2. 02QA pass rateQA pass rate is the share of evaluated outputs that satisfy the defined quality criteria on a specified review pass. It is calculated by dividing the number of passing outputs by the number evaluated, using the same scope, rules, and decision point.
  3. 03Fabric fidelityFabric fidelity is the degree to which an image preserves the visible material properties of the referenced garment. Relevant properties can include weave or knit, texture, sheen, transparency, thickness, print, color, weight, and the way the material folds.
  4. 04Print driftPrint drift is an unwanted change in a garment print or graphic between the product reference and an output, or between related outputs. The drift may affect position, scale, orientation, repeat spacing, shape, color, or continuity across seams and folds.
  5. 05Garment drapeGarment drape is the way fabric hangs, folds, bends, and falls over a body or support. It is shaped by the material's weight and stiffness, the garment's cut and construction, fit, pose, contact points, movement, and gravity.

04

Move products into commerce

Follow the production steps that turn product references into launch-ready visual sets.

  1. 01PDP image requirementsPDP image requirements are the visual, technical, and content specifications that product images must meet for a product detail page. They can define required views, crop, aspect ratio, dimensions, background, color accuracy, file format, naming, accessibility text, and channel-specific rules.
  2. 02Sample-to-PDP workflowA sample-to-PDP workflow is the sequence that turns a physical or digital garment sample into an approved set of product-detail-page images. It connects source capture, product data, art direction, image creation, quality review, formatting, and delivery to the correct SKU.
  3. 03Lookbook productionLookbook production is the planning and creation of a curated image sequence that presents a fashion collection, season, or styling point of view. Unlike a SKU-complete PDP set, a lookbook prioritizes visual narrative, outfit relationships, pacing, and collection coherence.

05

Extend stills into experiences

See how product imagery becomes motion or an interactive shopper view.

  1. 01Still-to-videoStill-to-video is a generation method that creates a moving sequence from one or more static images. The source image can act as the first frame or a visual reference while the system infers motion, camera behavior, and additional frames that were not present in the still.
  2. 02Virtual try-onVirtual try-on is a digital experience that visualizes how a garment may look on a person, uploaded image, live camera view, or avatar. It creates an appearance preview; it does not by itself predict physical fit, comfort, or size unless those capabilities are modeled separately.

06

Track content origins

Learn the records and standards used to communicate where a digital asset came from.

  1. 01Image provenanceImage provenance is information about the origin and history of a digital image, including who or what created it, which source assets or tools were involved, and what edits were recorded. It may be carried through metadata, signed manifests, watermarks, hashes, or external logs.
  2. 02C2PAC2PA is an open technical standard for binding signed provenance information, called a C2PA Manifest or Content Credential, to a digital asset. A credential can record assertions about creation and editing, while cryptographic validation can reveal whether the bound data or asset has changed.