Granted Patent B1

US 12,499,191 · App. 18/596,498 · Granted Dec 16, 2025

Inventor: Seaton Gras (Sequim, WA)

G06F21/31G06F16/958G07C9/00563

Abstract

Claims (25)

1 . A computing system for allowing access to a secure area through a determination process, the computing system comprising: a memory and one or more processors, wherein when creating an account, a user contributes one or more self provided content, wherein when logging in on a computing device, the user is presented with a multiple array of content within which the one or more self provided content is included as one or more elements that are presented and artificially created decoy images using an adversarial network that are similar in general appearance based on predefined parameters within the computing system, wherein the user selects their content element, which they recognize as their own because they made a content contribution, wherein each user selection, a score value increases when correct or decreases when incorrect, calculated by the one or more processors of the computing system, wherein the score value is greater than a predefined level, the user is granted access, by the one or more processors of the computing system, to the secure area on the computing device.

2 . The computing system of claim 1 wherein the user can contribute their own images or text, wherein the user can also select images from a source, wherein for text string contribution, as an example, the user can use a quote from their favorite literature or movie quote.

3 . The computing system of claim 1 wherein the content contribution of the user is in a form of text strings, wherein the computing system presents an array of similar text strings with comparable length and language, wherein for example, the array of similar text strings are configured to be pulled from a library of similar text strings, which are tagged with the comparable length and language and other attributes, wherein a close appearance of presented text strings makes it harder for anyone to guess a correct selection.

4 . The computing system of claim 1 , wherein the user can modify the score value level for concluding an authentication determination.

5 . The computing system of claim 1 , wherein the user is presented an array where none of the content was submitted by the user, such that any user selection merely reloads with a next array while making no change in the score value, wherein when in this mode, outside parties watching conducted steps will be misled to think this false step is part of login requirements.

6 . A computer-implemented method for secure user authentication, comprising:

receiving, by one or more processors, user-contributed content for creating an authentication pattern;

generating, using a machine learning model, a plurality of artificially created images based on the user-contributed content, wherein the artificially created images are visually similar to but distinct from the user-contributed content; storing the user-contributed content and the artificially created images in a secure database;

presenting, during an authentication attempt, a dynamic array of images including at least one piece of the user-contributed content and multiple artificially created images;

receiving user selections from the presented array; calculating, in real-time, an authentication score based on the user selections, wherein the authentication score increases for correct selections and decreases for incorrect selections; and

dynamically adjusting the presented array based on the calculated authentication score; granting access to a secure area when the authentication score exceeds a predefined threshold.

7 . The computer-implemented method of claim 6 , wherein

generating the artificially created images comprises:

analyzing, using computer vision algorithms, visual features of the user-contributed content; inputting the analyzed features into a generative adversarial network (GAN);

generating, by the GAN, the artificially created images that maintain key visual characteristics of the user-contributed content while introducing controlled variations.

8 . The computer-implemented method of claim 6 , further comprising: adaptively modifying a difficulty of distinguishing between the user-contributed content and artificially created images based on a user-defined or system-determined security level.

9 . The computer-implemented method of claim 6 , wherein the user-contributed content comprises user-contributed text strings, and generating artificially created images comprises:

analyzing semantic and syntactic features of the user-contributed text strings;

generating, using a natural language processing model, artificial text strings that are semantically and syntactically similar to the user-contributed text strings; and

converting the artificial text strings into image format for presentation in the dynamic array.

10 . A system for secure user authentication, comprising: a memory and one or more processors that cause the system to perform operations comprising:

receiving user-contributed content for creating an authentication pattern; generating, using an artificial intelligence model, a plurality of decoy elements based on the user-contributed content, wherein the plurality of decoy elements are designed to be visually or semantically similar to the user-contributed content;

generating parameters for predefined limits of difference between the plurality of decoy elements and the user-contributed content;

presenting, during an authentication session, a dynamic interface containing a mixture of the user-contributed content and decoy elements; and

analyzing user interactions with the presented interface to determine authentication success without providing explicit feedback on individual selections.

Continuity (3)

References Cited (44)