Virtual Models
About Our Product Photography and Virtual Models
Every Kiku garment is one of a kind. Each piece begins with vintage silk kimono textiles and is transformed into wearable art that honors the craftsmanship, history, and beauty of the original fabric. Because every garment is unique, photographing our collection presents challenges very different from those of traditional fashion brands that produce a single design in large quantities. Each piece exists only once, making extensive photography on multiple live models difficult and impractical.
Our first approach to product pages was to use female and male mannequins to show how a garment would look when worn. We soon moved to the more complicated process of translating the mannequin images into “ghost mannequin” images, where the mannequin was digitally removed while preserving the human shape in the garment. While more aesthetically attractive, ghost mannequin images can be difficult to interpret.
We are now transitioning to AI-generated virtual models created from photographs of the garment on a mannequin. We believe these virtual representations provide a more natural and intuitive way to understand the scale, silhouette, and spirit of a garment. Throughout our website, you will see four recurring virtual models intended to help illustrate how garments may appear on different body types and generations. These figures represent a young woman in her 20’s who is a size large, an older woman in her 40s who is a size medium, a senior woman who is a size small, and a young man who is a men’s medium size. By showing our pieces on multiple figures, we hope to give customers a broader and more realistic sense of proportion, styling, and versatility.
At the same time, AI-generated imagery has limitations that customers should understand. The software extrapolates information from the garment, mannequin, and virtual figure to create a realistic interpretation, but it does not know the garment’s actual dimensions and does not know the virtual model’s attributes, such as assumed size. As a result, sleeve length, hem length, garment volume, and overall fit may not always be represented accurately relative to the size of the virtual model. We attempt to show these limitations by the images on this page where all of the models, live, virtual and ghost, are wearing the same jacket.
In the example shown above, the AI-generated image interprets the male virtual model as being significantly smaller relative to the garment than the live model, as can be seen by the comparative sleeve and hem lengths. For this reason, we strongly encourage customers to rely on the measurements and written descriptions provided rather than the perceived fit shown on the virtual model.
Because Kiku garments originate from traditional Japanese kimono, AI software can also misinterpret certain construction details. Sleeves, in particular, present challenges, as the software may recognize the original kimono form without fully understanding the modifications made during the refashioning process.
Despite these limitations, we believe virtual models offer a clearer and more engaging representation of our garments than ghost mannequin imagery alone. Our goal is to provide the most informative and transparent presentation possible while celebrating the individuality of each handmade piece.