Understanding the wardrobe as a system.

A person photographs their clothes. Muse tries to understand them.

That understanding can then become the foundation for a set of capabilities — each at a different level of maturity.

• organizing a wardrobe
• describing individual garments
• evaluating pieces
• finding compatible items
• creating outfits
• recommending what to wear
• understanding personal style
• identifying gaps or opportunities within the wardrobe

The exact scope and behaviour of these capabilities is still evolving. See Features for the status of each.

Make personal style understandable.

The physical wardrobe is rich in information, but most of that information stays implicit. Muse asks whether technology can make it explicit.

If clothing can be photographed, recognized, structured and understood, a different kind of personal-style tool becomes possible — an intelligent layer between a person and the clothes they already own.

From a photograph to a wardrobe intelligence layer.

The initial interaction is deliberately simple. A person has clothing; a person photographs clothing; Muse tries to understand it. From there the product can progressively build a digital wardrobe.

• 01

  Capture

  Core concept

  Bring physical clothing into Muse through images. The image becomes the raw material from which the system can attempt to understand the garment.

• 02

  Understand

  Explored

  Use computer vision, image processing and AI to identify and describe the garment. The exact recognition pipeline remains an area of technical development.

• 03

  Structure

  In development

  Turn visual information into wardrobe data that can be searched, compared, filtered and used by other parts of the product.

• 04

  Evaluate

  Proposed

  Allow garments to be evaluated or rated, adding a layer of information beyond objective visual characteristics.

• 05

  Combine

  Proposed

  With multiple garments represented together, explore compatibility between pieces.

• 06

  Outfit

  Planned

  Use the wardrobe as the input for outfit creation.

• 07

  Recommend

  Future

  An intelligent personal stylist using the wardrobe and the individual's preferences to make useful recommendations.

A garment can be described along many dimensions.

One of the fundamental ideas behind Muse is that a wardrobe can be represented digitally. A single piece might be described through:

The exact data model remains part of the technical and product exploration — see open questions on data.

The more interesting question is what happens once the clothes are understood together.

Muse is not only about recognizing individual garments. A wardrobe representation could allow the system to reason about relationships:

These relationships are central to the longer-term product direction.

The evolving interaction model.

The less effort required to create a useful digital wardrobe, the more valuable the resulting system can become. The experience is being explored around several stages.

• Bring clothing in

  The user captures clothing through images; the system attempts to understand what it sees.

• Review

  The user inspects what Muse understood and corrects or refines it. This relationship between automation and user control is a central design question.

• Build the wardrobe

  As garments are added, Muse gradually develops a representation of the user's wardrobe.

• Explore

  The wardrobe becomes something the user can browse and work with.

• Create outfits

  The wardrobe becomes the basis for outfit combinations.

• Get recommendations

  Moving from browsing toward assistance — an answer to "what should I wear today?" based on the person's actual wardrobe rather than generic fashion advice.

• Personalization

  A future direction in which Muse learns from the user's preferences and choices. The personalization model is an open product and AI question.