AI Shoot Input Requirements

This guide outlines the specifications your images must meet to ensure optimal quality and timely delivery when using AutoRetouch’s GenAI features, such as AI Shoot. Providing high-quality input is essential for achieving market-leading results.

General input requirements

• File Format: All images must be in one of AutoRetouch’s supported file formats.
• Single Garment Focus: Each image must feature only one SKU (stock-keeping unit).
• Maximize Garment Area: The garment should occupy the largest possible proportion of the image (e.g., cropping the mannequin slightly is acceptable). This ensures the AI can operate at the highest possible resolution on the product itself.
• Avoid Complex Details: Garments should generally not have complex, busy patterns, transparent sections, or overly intricate/unusual shapes. These elements can reduce final quality.
• Minimum Resolution: A minimum resolution of 1.5K is required for all inputs.
• Required views: front (required), back (required for back views generation)

Preferred garment input types

The method of preparation significantly affects the final quality. We highly recommend using Ghost Mannequins as inputs.

Input Type	Quality	Description
1. Ghost Mannequins	⭐⭐⭐⭐⭐ (Best)	A composite image of the garment, typically showing the inner collar/neck-join (ghosted effect). This is the best input type as it provides a clean, 3D-like, background-free view of the garment.
2. Ghosted Garment	⭐⭐⭐⭐ (Excellent)	An image of the garment with the background removed, often taken on-model or on a mannequin initially. This is a very clean input that works well.
3. Mannequin Shot	⭐⭐⭐ (Good)	A picture of the garment on a plastic mannequin. If you submit this, AutoRetouch will need to remove the background first, which may be billed as pre-processing.
4. Flat Lay	⭐⭐ (Acceptable)	A picture of the garment laid flat. This can sometimes result in the final generated model wearing the garment in an overly flat manner, leading to less realistic fit and volume.