Image-Derived Advertising Targeting Metrics

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional advertising methods lack precision in targeting users, as they cannot utilize user-specific information derived from interactions with computing devices, limiting the relevance of advertisements.

Innovation Solution

Analyzing publicly and non-publicly available images to derive advertising targeting metrics, including physical attributes and preferences, which can be used to select and display tailored advertisements, with options for local processing to enhance privacy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional advertising methods are used, then advertising can be displayed to users, but the targeting precision is low and relevance to users is limited

Engineering Contradiction:
Improvetargeting precisionVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments user information into multiple dimensions including publicly available information, non-publicly available information, and image-derived metrics. This segmentation allows the system to process and utilize different types of user data separately, thereby improving targeting precision without overwhelming system complexity

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a new dimension for advertising targeting by deriving metrics from user images (physical attributes, demographic information). This adds a visual dimension to traditional text-based user profiling, significantly enhancing targeting precision beyond conventional methods

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If non-public images are processed remotely to derive advertising metrics, then more precise user information can be obtained, but user privacy is compromised

Engineering Contradiction:
Improveuser information accuracyVSAvoidprivacy risk
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent extracts the image processing function from remote servers and relocates it to the user's local computing device. This extraction allows precise user information to be derived locally without transmitting sensitive non-public images to external servers, thereby maintaining privacy while achieving accurate metrics

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The user's computing device acts as an intermediary that processes non-public images locally. Instead of directly transmitting images to advertising servers, the local device derives metrics and sends only the processed information, serving as a privacy-protecting intermediary between the user and the advertising system

Inventive Principle:
Principle #24Intermediary (Mediator)

3Adaptability or versatility

If more user-specific information is collected for advertising targeting, then advertisement relevance increases, but user privacy concerns increase

Engineering Contradiction:
Improveadvertising relevanceVSAvoidprivacy concern
Core Design Contradiction:
Adaptability or versatilityVSObject-affected harmful factors

Solution Approach 1:

The patent creates a local copy of the image processing capability on the user's computing device. Instead of sending original images to servers for analysis, the system uses local processing to generate advertising metrics, thereby collecting necessary user-specific information while keeping the actual image data private and local

Inventive Principle:
Principle #26Copying

Data Source

PatentUS9183557B2Advertising targeting based on image-derived metrics
Publication Date: 2015.11.10 MICROSOFT TECHNOLOGY LICENSING LLC
  • US9183557B2 patent drawing
  • US9183557B2 patent drawing
  • US9183557B2 patent drawing

AI summary

Advertising targeting metrics for individuals can be derived from images associated with those individuals. Such advertising targeting metrics can include physical attributes, as well as preferences based on prior activity, or history. Public images associated with specific user identities can be processed and advertising targeting metrics can be derived therefrom to more accurately tailor the advertisements displayed to the individuals associated with those user identities. Additionally, non-public images can be likewise processed, either remotely or, for greater privacy, locally if so allowed by the user. Advertisers can then utilize the greater breath of advertising targeting metrics that can be derived from images to more accurately target advertisements to specific groups. In doing so, advertisers can submit their own images of exemplary targeted users, and the targeting metrics for those advertisers' advertisements can be automatically derived from the submitted images utilizing equivalent algorithms.