Virtual try-on combines an image of a person with an outfit reference and asks a generative model to produce a plausible composite. The result can be useful for exploring an aesthetic direction. It can also alter identity, anatomy, fabric, colour, pose, background, and details—even when the instruction asks the model to preserve them.
Generation is not a digital dressing room measurement
A generative system learns statistical visual relationships and synthesises pixels. It does not physically place the exact manufactured garment on a measured 3D body. Research in fashion computing covers recommendation, compatibility, retrieval, and image generation, each with different objectives and evaluation limits.
A convincing picture can still contain incorrect seams, closures, texture, opacity, logos, proportions, hands, or drape. Visual realism is not evidence of product accuracy.
What happens to a photo in Lace Riot
The app sends the selected photo of the user and the chosen outfit reference to the Lace Riot API, which forwards the required inputs to fal.ai for generation. Lace Riot's server is designed not to save the uploaded source photo as a media file or database record. Provider processing still occurs and is governed by the provider arrangement described in the Privacy Policy.
Product search is a separate flow. It sends the catalogue outfit image—not the user's personal photo—to OpenAI with storage disabled for that request, then returns cited external links. Always verify each seller independently.
Use the output for direction, never diagnosis or fit certainty
The image cannot reliably determine body measurements, clothing size, compression, support, coverage, transparency, fibre behaviour, comfort, health, attractiveness, or personal value. Do not use it for medical, psychological, legal, employment, or other high-stakes decisions.
Upload only a current photo of yourself, showing one clearly adult person. Do not upload another person, a public figure, a group, or anyone under 18. Review every result critically and check the actual product details, measurements, price, availability, and return terms with the retailer.
- Treat every output as synthetic and potentially wrong.
- Check identity, anatomy, fabric, and garment details before saving or sharing.
- Never infer size, health, attractiveness, or worth from the generation.
- Use only your own current photo and only if you are 18 or older.
THREE THINGS TO REMEMBER
- Virtual try-on creates synthetic pixels; it does not measure a body or physical garment.
- Lace Riot separates personal-photo generation from outfit-only product search.
- Outputs can be wrong and must not guide sizing, health, or other high-stakes decisions.
Questions and answers
Does AI virtual try-on show my correct clothing size?
No. Lace Riot does not measure your body or reliably model the exact product's construction and material behaviour.
Can I upload a photo of someone who gave permission?
No. Lace Riot's rule is stricter: upload only a current image of yourself, with one clearly adult person visible.
Does product search use my personal photo?
No. The product-search step uses the selected catalogue outfit image rather than the user's personal photograph.
Research sources
- Yujuan Ding, Zhihui Lai, P. Y. Mok, Tat-Seng Chua (2024). Computational Technologies for Fashion Recommendation: A Survey. ACM Computing Surveys. DOI: 10.1145/3627100
Limitation: Benchmark performance across fashion-computing systems does not guarantee an accurate or satisfying result for an individual user.
- Tangwei Ye, Liang Hu, Qi Zhang, Zhong Yuan Lai, Usman Naseem, Dora D. Liu (2023). Show Me The Best Outfit for A Certain Scene: A Scene-aware Fashion Recommender System. Proceedings of the ACM Web Conference 2023. DOI: 10.1145/3543507.3583435
Limitation: The paper evaluates a scene-aware recommender. It provides research context but does not describe a feature that Lace Riot claims to offer.