EP-4738282-A1 - MEDIA ANALYSIS FOR REGION PREDICTION

EP4738282A1EP 4738282 A1EP4738282 A1EP 4738282A1EP-4738282-A1

Abstract

Systems and methods are disclosed for content analysis and/or processing. In some examples, a system receive content (e.g., an image) that includes a representation of an object (e.g., a person) in a scene. The system analyzes the content to detect a first portion of the content (e.g., subset of pixels) that represents the object. A second portion of the content represents the scene. The system obfuscates (e.g., masks, blurs, pixelates, replaces) the first portion of the content to generate obfuscated content. The obfuscated content retains the second portion of the content that represents the scene. The system analyzes the obfuscated content using at least one trained machine learning model to generate a prediction of a geographic region (e.g., at least a portion of a continent) that the scene is in. The system outputs an indicator of the prediction of the geographic region.

Inventors

RAZA, Shahzad
JABLONSKI, Mitchell
MACHAK, Christina
JOSHI, Aditya

Assignees

Block, Inc.

Dates

Publication Date: 20260506
Application Date: 20251029

Claims (15)

A computer-implemented method comprising: receiving content, wherein the content includes a representation of an object in a scene; analyzing the content to detect a first portion of the content that represents the object, wherein a second portion of the content represents the scene; obfuscating the first portion of the content to generate obfuscated content, wherein the obfuscated content retains the second portion of the content that represents the scene; analyzing the obfuscated content using at least one artificial intelligence model to generate a prediction of a geographic region that the scene is in, wherein the geographic region is at least a portion of a continent; and outputting an indicator of the prediction of the geographic region.
The computer-implemented method of claim 1, wherein the object includes at least a portion of a person.
The computer-implemented method of claim 1 or claim 2, wherein the object includes personally identifiable information (PII) associated with a person.
The computer-implemented method of any of claims 1 to 3, wherein the at least one artificial intelligence model includes at least one trained machine learning model, further comprising: receiving further information indicating a level of accuracy of the prediction of the geographic region; and updating the at least one trained machine learning model based on the further information to improve an accuracy of the trained machine learning model at geographic region prediction.
The computer-implemented method of any of claims 1 to 4, wherein analyzing the content to detect the first portion of the content that represents the object includes processing the content using a second artificial intelligence model that detects the first portion of the content that represents the object.
The computer-implemented method of claim 5, wherein the second artificial intelligence model includes a second trained machine learning model, further comprising: receiving further information indicating a level of accuracy of the detection of the first portion; and updating the second trained machine learning model based on the further information to improve an accuracy of the second trained machine learning model at object detection.
The computer-implemented method of any of claims 1 to 6, wherein obfuscating the first portion of the content includes one or more of: masking the first portion of the content; blurring the first portion of the content; or pixelating the first portion of the content.
The computer-implemented method of any of claims 1 to 7, wherein obfuscating the first portion of the content includes interpolating based on the second portion of the content to replace the first portion of the content.
The computer-implemented method of any of claims 1 to 8, further comprising: comparing the prediction of the geographic region with a second indicator of geographic region of the object to identify a conflict between the prediction and the second indicator, wherein the indicator of the prediction of the geographic region is an indicator of the conflict.
The computer-implemented method of claim 9, further comprising: automatically disabling access to at least one of an asset or a function of a device in response to identifying the conflict.
The computer-implemented method of any of claims 1 to 10, further comprising: processing the prediction of the geographic region and a second indicator of geographic region of the object using a second artificial intelligence model to determine a level of confidence in an identity of a user, wherein the object is the user; and using the level of confidence in the identity of the user for at least one of a fraud check, an identity verification, or a transaction authorization.
The computer-implemented method of claim 11, wherein the second artificial intelligence model includes a second trained machine learning model, further comprising: receiving further information about the identity of the user; and updating the second trained machine learning model based on the further information to improve an accuracy of the second trained machine learning model at identity determination.
The computer-implemented method of any of claims 1 to 12, wherein outputting the indicator of the prediction of the geographic region includes associating the indicator of the prediction of the geographic region with an identifier of the object in a data structure, and/or, wherein outputting the indicator of the prediction of the geographic region includes using the prediction of the geographic region to make a decision, wherein the decision includes at least one of a decision for which region to input for a region field of a form, a decision whether a functionality of a device is to be enabled, a decision as to what currency to use in association with a transaction, a decision as to what price to use in association with a transaction, a decision as to what language to use for a string of text, or a decision as to which of a set of locations to select.
A system comprising: a memory that stores instructions; and a processor coupled to the memory, wherein execution of the instructions by the processor causes the processor to: receive content, wherein the content includes a representation of an object in a scene; analyze the content to detect a first portion of the content that represents the object, wherein a second portion of the content represents the scene; obfuscate the first portion of the content to generate obfuscated content, wherein the obfuscated content retains the second portion of the content that represents the scene; analyze the obfuscated content using at least one artificial intelligence model to generate a prediction of a geographic region that the scene is in, wherein the geographic region is at least a portion of a continent; and output an indicator of the prediction of the geographic region.
A computer-implemented method for image processing, the computer-implemented method comprising: receiving an image of an object in a scene, wherein the image includes a plurality of pixels; analyzing the image to detect pixels representing the object within the plurality of pixels of the image, wherein the pixels depicting the object are a first subset of the plurality of pixels of the image, wherein a second subset of the plurality of pixels of the image represent the scene; masking the pixels representing the object to generate a masked image, wherein the masked image retains the second subset of the plurality of pixels that represent the scene; analyzing the masked image using at least one trained machine learning model to generate a prediction of a geographic region that the scene is in, wherein the geographic region is at least a portion of a continent; and associating an indicator of the prediction of the geographic region with an identifier of the object in a data structure.

Description

TECHNICAL FIELD Image sensors are devices that capture images of a scene. In an illustrative example, an image may include a depiction of a person in a scene. In some cases, certain systems can use images of users for a user profile, for identity verification, for biometric verification, for sharing in a social media context, or for other purposes. BRIEF DESCRIPTION OF THE DRAWINGS The detailed description is set forth with reference to the accompanying figures. In the figures, the left-most digit(s) of a reference number identifies the figure in which the reference number first appears. The use of the same reference numbers in different figures indicates similar or identical items or features. Moreover, multiple instances of the same part are designated by a common prefix, in some cases separated from the instance number by a dash and/or parentheses. The drawings are not to scale. FIG. 1 is a block diagram illustrating a process, performed by a content processing system, in which user data (e.g., including an image and a stated region) is received by platform server(s) from user device(s) and is analyzed using analysis engine(s) to generate a region prediction, in accordance with some examples;FIG. 2 is a block diagram illustrating a process, performed by an analysis engine, of obfuscating an image and analyzing the obfuscated image to perform region prediction, in accordance with some examples;FIG. 3 is a block diagram illustrating a process through which a region prediction is generated by analyzing an obfuscated image, including reasons for the region prediction, in accordance with some examples;FIG. 4 is a block diagram illustrating a process through which a region prediction is generated by analyzing an obfuscated image, including reasons for the region prediction, in accordance with some examples;FIG. 5 is a block diagram illustrating examples of different types of obfuscated images that can be generated by the obfuscation engine from an image, in accordance with some examples;FIG. 6 is a block diagram illustrating a process, performed by an analysis engine, of obfuscating an image, analyzing the obfuscated image to perform region prediction, and using the region prediction for further decisions, in accordance with some examples;FIG. 7 is a block diagram illustrating an example of a machine learning system for training, use of, and/or updating of one or more machine learning model(s) that are used for content processing, region prediction, and/or identity verification, in accordance with some examples;FIG. 8 is a flow diagram illustrating a process for content analysis and/or processing, in accordance with some examples;FIG. 9 is a flow diagram illustrating a process for image analysis and/or processing, in accordance with some examples;FIG. 10 is a block diagram illustrating an example environment for providing an application and/or for customizing the application for different platforms, in accordance with some examples;FIG. 11 is a block diagram illustrating an example environment including a service provider system which may be associated with the server(s) of FIG. 10, in accordance with some examples; andFIG. 12 is a block diagram illustrating a system for performing techniques described herein, in accordance with some examples. DETAILED DESCRIPTION Platforms that provide services to users rely heavily on security to maintain trust with users and to protect the platform from malicious parties. Security can be particularly important for a platform that provides payment services, transaction services, and/or asset management services for users. If security is compromised for such platforms, bad actors can transact on such platforms, users can lose access to their assets (e.g., fiat currency, cryptocurrency, securities, and/or other assets), can lose their assets, can lose the ability to make a purchase, can lose the ability to perform another type of transaction, and the like. A high level of security is also important to maintaining users' trust in a platform. Thus, maintaining and improving security is both technically important for platform systems and practically important to users of platforms. In some cases, a platform can receive an image or other media content from a user, for instance for an image for the user's profile or account, for identity verification, for biometric verification, for other purposes, or a combination thereof. In some cases, a platform can also receive an indication, from the user, as to what region the user is located in, what region the user resides in, what country (region) the user is a citizen of, or another indication of geographic region. In some cases, a platform can receive another type of media content for such purposes, such as a video, a 3D representation of a scene (e.g., captured using a depth sensor and/or stereoscopic camera arrangement), another type of media content, or a combination thereof. In an example, systems and methods are described for proce