Partial Facial Recognition for Obstructed User Access

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current facial recognition systems cannot identify users with partially obstructed faces, leading to inconvenience or security risks for users wearing eyewear or masks, as they are unable to access computing devices without removing obstructions or using alternative authentication methods.

Innovation Solution

A computing device equipped with an image sensor and AI model that generates real-time images of users, compares unobstructed facial features to stored reference images, and allows access based on customizable security settings, including location, time, and secondary authentication methods, even if the face is partially obstructed.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If traditional facial recognition systems are used, then security is maintained through complete face verification, but users with obstructed faces cannot access the device

Engineering Contradiction:
Improveaccess convenienceVSAvoidauthentication reliability
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The system performs facial recognition on only the unobstructed portion of the face rather than requiring complete face verification. The processor identifies that portions of the face are unobstructed and performs recognition based solely on those visible portions, allowing users with partially obstructed faces to access the device while maintaining security through the recognition of unique facial features in the visible areas.

Inventive Principle:
Principle #16Partial or excessive action

2Reliability

If complete face verification is required, then authentication security is ensured, but users must remove obstructions or use alternative methods

Engineering Contradiction:
Improveauthentication securityVSAvoidaccess convenience
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The system performs facial recognition on only the unobstructed portion of the face rather than requiring complete face verification. The processor identifies that portions of the face are unobstructed and performs recognition based solely on those visible portions, allowing users with partially obstructed faces to access the device while maintaining security through the recognition of unique facial features in the visible areas.

Inventive Principle:
Principle #16Partial or excessive action

3Productivity

If facial recognition is implemented, then access speed is improved, but the system fails when faces are obstructed

Engineering Contradiction:
Improveaccess speedVSAvoidsystem robustness
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The system performs facial recognition on only the unobstructed portion of the face rather than requiring complete face verification. The processor identifies that portions of the face are unobstructed and performs recognition based solely on those visible portions, allowing users with partially obstructed faces to access the device while maintaining security through the recognition of unique facial features in the visible areas.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The system dynamically adjusts the recognition parameters based on the degree of facial obstruction. When portions of the face are obstructed, the system modifies its recognition algorithm to focus on and weigh the unobstructed portions more heavily, allowing flexible adaptation to different obstruction scenarios while maintaining fast access speeds.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS12197550B2Providing access to a computing device
Publication Date: 2025.01.14 MICRON TECHNOLOGY INC
  • US12197550B2 patent drawing
  • US12197550B2 patent drawing
  • US12197550B2 patent drawing

AI summary

Methods and devices associated with providing access to a computing device are described. A method can include generating a real-time image including an attempted user, comparing the real-time image to a reference image of an authorized user, identifying an obstructed portion of the attempted user, comparing an unobstructed portion of the attempted user to a portion of the authorized user corresponding to the unobstructed portion of the attempted user, providing access to a computing device in response to the unobstructed portion of the attempted user matching the portion of the authorized user, and rejecting access to the computing device in response to the unobstructed portion of the attempted user being different from the portion of the authorized user.