Granted Patent B2

US 12,725,332 · App. 18/642,119 · Granted Sep 1, 2026

Inventors: Zeinab Bagheri (Toronto, CA); Suraya Shivji (New York, NY); Yinghua David Xia (Toronto, CA); Matthew Saunders (Santa Monica, CA)

Assignee: SNAP INC.

G06T11/60G06F3/04817G06T11/10A63F13/53A63F13/795A63F13/87G06T2200/24G06T2210/16

Abstract

Claims (61)

1 . A method comprising:

presenting, by one or more processors, a graphical user interface (GUI) comprising a fashion item selection region that includes icons representing different types of avatar fashion items, each type of avatar fashion item defining a category of garment or accessory wearable by an avatar;

receiving a first input via the fashion item selection region that selects an individual icon corresponding to an individual avatar fashion item of a first type;

in response to receiving the first input, presenting a texture selection region separate from the fashion item selection region, the texture selection region comprising a first set of predefined textures and an option associated with a second set of textures generated based on a prompt by a generative machine learning model, each texture in the first set of predefined textures and the second set of textures defining visual attributes comprising at least one of a pattern, a material appearance, or a color scheme applicable to the individual avatar fashion item;

receiving a second input via the texture selection region that selects an individual texture;

applying the individual texture to the individual avatar fashion item by modifying visual attributes of the individual avatar fashion item according to the individual texture, the individual avatar fashion item is overlaid on an avatar depicted in the GUI;

receiving a third input via the fashion item selection region that selects a different avatar fashion item of a second type different from the first type; and

in response to receiving the third input, automatically applying the individual texture to the different avatar fashion item without receiving additional input via the texture selection region to maintain the visual attributes defined by the individual texture across changing from the first type to the second type of the avatar fashion item.

2 . The method of claim 1 , further comprising:

selecting a predefined texture from the first set of predefined textures;

applying the selected predefined texture to the individual avatar fashion item; and

overlaying, on the avatar, the individual avatar fashion item to which the selected predefined texture has been applied.

3 . The method of claim 2 , further comprising:

receiving input that adjusts a size of the individual avatar fashion item with the individual texture relative to the avatar.

4 . The method of claim 2 , further comprising:

receiving input that adjusts a position of the individual texture on the individual avatar fashion item that is applied to the avatar.

5 . The method of claim 2 , further comprising:

presenting a text input region associated with the prompt for the generative machine learning model together with the first set of predefined textures;

in response to selecting the predefined texture, accessing a predefined prompt associated with the predefined texture; and

populating the text input region automatically with the predefined prompt.

6 . The method of claim 1 , further comprising:

presenting a text input region associated with the prompt for the generative machine learning model together with the first set of predefined textures;

receiving input that selects the text input region; and

in response to receiving the input that selects the text input region, presenting a list of suggested prompts for the generative machine learning model.

7 . The method of claim 6 , further comprising:

selecting an individual prompt from the list of suggested prompts;

generating an enhanced prompt using the individual prompt having more words than words of the individual prompt; and

processing the enhanced prompt by the generative machine learning model to generate the second set of textures having attributes matching text of the enhanced prompt.

8 . The method of claim 7 , further comprising:

processing the enhanced prompt by a first instance of the generative machine learning model to generate a first subset of textures of the second set of textures; and

processing the enhanced prompt by a second instance of the generative machine learning model to generate a second subset of textures of the second set of textures.

9 . The method of claim 8 , wherein the first and second instances of the generative machine learning model process the individual prompt in parallel.

10 . The method of claim 8 , wherein the first instance of the generative machine learning model comprises a first model trained based on a first set of training data, and wherein the second instance of the generative machine learning model comprises a second model trained based on a second set of training data.

11 . The method of claim 8 , wherein the first and second instances of the generative machine learning model each comprise a respective stable diffusion model.

12 . The method of claim 6 , further comprising:

receiving text from a user in the text input region to generate an individual prompt;

generating an enhanced prompt using the individual prompt having more words than words of the individual prompt; and

processing the individual prompt by the generative machine learning model to generate the second set of textures having attributes matching text of the individual prompt.

13 . The method of claim 12 , further comprising:

presenting the second set of textures in the GUI;

selecting an artificial texture from the second set of textures;

applying the selected artificial texture to the individual avatar fashion item; and

overlaying, on the avatar, the individual avatar fashion item to which the selected artificial texture has been applied.

14 . A system comprising:

at least one processor; and

at least one memory component having instructions stored thereon that, when executed by the at least one processor, cause the at least one processor to perform operations comprising:

presenting a graphical user interface (GUI) comprising a fashion item selection region that includes icons representing different types of avatar fashion items, each type of avatar fashion item defining a category of garment or accessory wearable by an avatar;

receiving a first input via the fashion item selection region that selects an individual icon corresponding to an individual avatar fashion item of a first type;

in response to receiving the first input, presenting a texture selection region separate from the fashion item selection region, the texture selection region comprising a first set of predefined textures and an option associated with a second set of textures generated based on a prompt by a generative machine learning model, each texture in the first set of predefined textures and the second set of textures defining visual attributes comprising at least one of a pattern, a material appearance, or a color scheme applicable to the individual avatar fashion item;

receiving a second input via the texture selection region that selects an individual texture;

applying the individual texture to the individual avatar fashion item by modifying visual attributes of the individual avatar fashion item according to the individual texture, the individual avatar fashion item is overlaid on an avatar depicted in the GUI;

receiving a third input via the fashion item selection region that selects a different avatar fashion item of a second type different from the first type; and

in response to receiving the third input, automatically applying the individual texture to the different avatar fashion item without receiving additional input via the texture selection region to maintain the visual attributes defined by the individual texture across changing from the first type to the second type of the avatar fashion item.

15 . A non-transitory computer-readable storage medium having stored thereon instructions that, when executed by at least one processor, cause the at least one processor to perform operations comprising:

presenting a graphical user interface (GUI) comprising a fashion item selection region that includes icons representing different types of avatar fashion items, each type of avatar fashion item defining a category of garment or accessory wearable by an avatar;

receiving a first input via the fashion item selection region that selects an individual icon corresponding to an individual avatar fashion item of a first type;

in response to receiving the first input, presenting a texture selection region separate from the fashion item selection region, the texture selection region comprising a first set of predefined textures and an option associated with a second set of textures generated based on a prompt by a generative machine learning model, each texture in the first set of predefined textures and the second set of textures defining visual attributes comprising at least one of a pattern, a material appearance, or a color scheme applicable to the individual avatar fashion item;

receiving a second input via the texture selection region that selects an individual texture;

applying the individual texture to the individual avatar fashion item by modifying visual attributes of the individual avatar fashion item according to the individual texture, the individual avatar fashion item is overlaid on an avatar depicted in the GUI;

receiving a third input via the fashion item selection region that selects a different avatar fashion item of a second type different from the first type; and

in response to receiving the third input, automatically applying the individual texture to the different avatar fashion item without receiving additional input via the texture selection region to maintain the visual attributes defined by the individual texture across changing from the first type to the second type of the avatar fashion item.

Assignments (1)

Continuity (1)

References Cited (272)