IP Library › Granted Patent US 12,681,962
Granted Patent B2
US 12,681,962 · App. 18/958,801 · Granted Jul 14, 2026

Prompt personalization for generative models

Inventors: Anand Khanna (San Jose, CA); Abhishek Tandon (San Jose, CA); Nikolaos Vlassis (San Jose, CA); Oliver Brdiczka (San Jose, CA)
Assignee: ADOBE INC.
G06F16/3322G06T11/00G06T2200/24
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,681,962
App. No.
18/958,801
Granted
Jul 14, 2026
Kind
B2
Abstract

A method, apparatus, non-transitory computer readable medium, and system for generating suggested prompts include obtaining a sequence of text prompts associated with a user and determining a session concept for the user based on the sequence of text prompts. Embodiments then generate, using a prompt generation model, an image generation prompt based on the sequence of text prompts and the session concept. Subsequently, embodiments generate, using an image generation model, a synthetic image based on the image generation prompt.

Claims (56)

1 . A method comprising:

obtaining a sequence of text prompts associated with a user, wherein each of the sequence of text prompts comprises an input for an image generation model;

determining a session concept text for the user based on the sequence of text prompts, wherein the session concept text describes an image element common to a plurality of the sequence of text prompts;

generating, using a prompt generation model, an image generation prompt based on the session concept text, wherein the image generation prompt describes a scene including the image element from the session concept text; and

generating, using the image generation model, a synthetic image based on the image generation prompt, wherein the synthetic image depicts the scene including the image element from the session concept text.

2 . The method of claim 1 , wherein generating the image generation prompt comprises:

concatenating a token representing the session concept text to a sequence of tokens representing the sequence of text prompts to obtain an augmented input, wherein the image generation prompt is generated based on the augmented input.

3 . The method of claim 1 , further comprising:

receiving feedback on the image generation prompt;

generating an additional image generation prompt based on the feedback and the sequence of text prompts; and

generating an additional synthetic image based on the additional image generation prompt.

4 . The method of claim 1 , wherein obtaining a sequence of text prompts comprises:

identifying a user session; and

receiving the sequence of text prompts from a user during the user session.

5 . The method of claim 1 , wherein determining the session concept text comprises:

generating the session concept text using a concept generation model based on the sequence of text prompts.

6 . The method of claim 1 , further comprising:

presenting the synthetic image to the user via a graphical user interface (GUI).

7 . The method of claim 1 , wherein:

the prompt generation model is trained to generate image generation prompts using training data including a training sequence of text prompts and a training concept corresponding to the training sequence of text prompts.

8 . The method of claim 1 , wherein:

the prompt generation model is trained to generate image generation prompts using training data including a plurality of image captions and a training concept corresponding to the plurality of image captions.

9 . A non-transitory computer readable medium storing code for prompt generation, the code comprising instructions that, when executed by at least one processor, cause the at least one processor to perform operations comprising:

identifying a user session;

receiving a sequence of text prompts during the user session, wherein each of the sequence of text prompts comprises an input for an image generation model;

determining a session concept text based on the sequence of text prompts;

generating, using a prompt generation model, an image generation prompt based on the session concept text, wherein the image generation prompt describes a scene including the image element from the session concept text; and

generating, using the image generation model, a synthetic image based on the image generation prompt, wherein the synthetic image depicts the scene including the image element from the session concept text.

10 . The non-transitory computer readable medium of claim 9 , the operations further comprising:

presenting the synthetic image to the user via a graphical user interface (GUI).

11 . The non-transitory computer readable medium of claim 9 , the operations further comprising:

generating the session concept text using a concept generation model based on the sequence of text prompts.

12 . The non-transitory computer readable medium of claim 9 , wherein:

the prompt generation model is trained to generate image generation prompts using training data including a training sequence of text prompts and a training concept corresponding to the training sequence of text prompts.

13 . The non-transitory computer readable medium of claim 12 , wherein:

the training sequence of text prompts is filtered based on prompt consistency.

14 . The non-transitory computer readable medium of claim 12 , the operations further comprising:

generating sentence-level embeddings of the text prompts of the training sequence of text prompts; and

computing pairwise distances between the sentence-level embeddings.

15 . The non-transitory computer readable medium of claim 9 , wherein:

the prompt generation model is trained to generate image generation prompts using training data including a plurality of image captions and a training concept corresponding to the plurality of image captions.

16 . A system comprising:

a memory component;

a processing device coupled to the memory component, the processing device configured to perform operations comprising:

obtaining a sequence of text prompts associated with a user, wherein each of the sequence of text prompts comprises an input for an image generation model;

determining a session concept text for the user based on the sequence of text prompts, wherein the session context text describes an image element common to a plurality of the sequence of text prompts;

generating, using a prompt generation model, an image generation prompt based on the session concept text, wherein the image generation prompt describes a scene including the image element from the session concept text; and

generating, using the image generation model, a synthetic image based on the image generation prompt, wherein the synthetic image depicts the scene including the image element from the session concept text.

17 . The system of claim 16 , the system further comprising:

a concept generation model trained to generate the session concept text based on the sequence of text prompts.

18 . The system of claim 16 , wherein:

the image generation model comprises a diffusion model.

19 . The system of claim 16 , the system further comprising:

a caption generation model configured to generate a caption based on an image, wherein the prompt generation model is trained using the caption.

20 . The system of claim 16 , wherein:

the prompt generation model comprises a transformer architecture.