User Guide

Request process

1The first referral process

contact

To make your first AI model image request, please submit a request on the [Contact] page with a brief description of your request. Please be sure to include your email address.

A representative will review your request and contact you with availability and a request form.

Go to Contact
contact

2Request by email

contact

If you have experience with generative AI, please download and fill out the request form. Click the Download button to download a request form.

If you have an image of the model you want, please enter the model code name.
(We'll create a new model based on the image you have)

Download

Please send the completed request form to aimodel@lghnh.com. We'll get back to you within 2 business days with production availability, scheduling, and any additional questions.

contact

Model creation range

The User Guide provides guidance on the precautions to take when creating model images and the scope of image generation. Please refer to the following items and place your request.

Cautions for using AI model people

Instead of using the AI model as a specific, fixed face, we recommend using it as a 'reference' for your artwork. Models are not trained by the AI, so it's not possible to lock onto the same face as the model.

Generated as close as possible to the model you selected, but changes in appearance occurred

Importance of reference images

Providing multiple reference images of the desired atmosphere will effectively assist AI in creating similar images, as AI learns a significant portion of the reference images when aiming to produce images with a similar atmospher

AI images generated to reflect the skin expression, brow texture, and stray hairs of the reference.

Face-focused image

When a face and hands are shown together, there is a higher probability of errors occurring in the hands. We recommend generating images that focus on the model's face.

Face-focused image

Creating detailed hands (number of fingers, hand to face ratio)

Limitations of detailed image generation

As images are generated through prompts, we recommend the most common styles. There are limitations to detailed expressions. (Specific clothing brands are not recommended.)

Create images with universal words (long hair, white shirt, etc.)

Specific clothing brands, hairstyles detailed volumisation, etc.

Limitations on identical face image generation

Even with the same prompt settings, the model faces can be slightly different. (The images below were created with the same prompts).

Created with the same prompt, but the look changes depending on the pose

Unique masks & Specific celebrity

Creating AI-generated faces with universally appealing features is generally easier, while producing models resembling specific individuals can pose challenges due to potential portrait rights issues.

Limitations of creating unique looks

Can't create celebrity-like images

Limited diversity and inclusivity

The quality of images generated by AI can vary depending on the amount of training data. Children, men, and East Asians have less training data compared to adults, men, and Westerners. The features of Korean faces may not be fully reflected compared to Japanese or Chinese faces.

Western women

Korean baby boy

Multi-person model

AI can generate model images of various appearances, races, and genders. However, there's a higher chance of errors occurring when creating complex poses, poses involving hands, or images where models are positioned close together.

Diversity of appearance, race, and gender. Moderate distance between models

Close distance between models, Similar model appearance

Limitations of Model Directed Poses

To create an image of a model holding a product, create an Ai model and then composite the product. We only provide an image of the model pose for you to use to help you assemble your product, and it is not possible to create detailed model poses.

face is away from the product, lightly holding the product

Pose using both hands to hold and showcase product

Product formulation limitations

Only normal formulations can be displayed, including how thin the formulation is, where it is applied to the face and how it is presented. There are limitations in terms of detail.

Applying a Cream Texture to the Model's Cheeks look (normal formulation)

Limitations of fine-tuning (formulation, colour, etc.)

Image Usage Precautions

Image Information

  1. 1 Basic image size is 512px. Images can be upscaled to a maximum of 8192px.
  2. 2 We deliver image creations in jpg / png extensions. The layout of the model and background cannot be separated.

License Information

  1. 1 Please clearly indicate that the image was created using AI when using it.
  2. *This image was created using AI
  3. 2 When training AI models using images or data of specific people, legal consent for portrait rights must be obtained.