Facial Recognition Security for Unauthorized Access Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional device security approaches face challenges in efficiently identifying and preventing unauthorized individuals from viewing sensitive data through shoulder surfing, particularly in public spaces or when users are engaged in sensitive tasks.

Innovation Solution

The implementation of machine learning facial recognition techniques on user devices to automatically detect and compare faces within a given proximity, triggering security actions if the detected face is not recognized as trusted, thereby generating a secure view or taking other security-related measures.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional device security approaches are used, then device security is maintained, but unauthorized shoulder surfers cannot be efficiently identified and prevented

Engineering Contradiction:

Improvedevice securityVSAvoidefficiency in identifying unauthorized individuals

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent replaces conventional mechanical/security-based device protection methods with an automated optical recognition system. A camera captures images of individuals near the device, and machine learning algorithms automatically analyze facial features to identify unauthorized persons, substituting manual security monitoring with automated optical-mechanical recognition.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The device performs self-monitoring by automatically detecting faces in its vicinity, comparing them against a database of authorized users, and triggering security responses without external intervention. The system serves its own security needs through autonomous facial recognition and automated alert generation.

Inventive Principle:

Principle #25Self-service

2Extent of automation

If machine learning facial recognition techniques are implemented, then unauthorized individuals are automatically detected, but device complexity increases

Engineering Contradiction:

Improveautomatic detection of unauthorized individualsVSAvoidsecurity system complexity

Core Design Contradiction:

Extent of automationVSDevice complexity

Solution Approach 1:

The patent integrates multiple functions into a unified security system: the camera serves both as a standard device component and as a facial recognition sensor; the processor handles both normal device operations and security analysis; the system provides both authorized user confirmation and unauthorized user detection through a single integrated mechanism.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent introduces an intermediary layer of machine learning software that bridges the gap between simple camera input and complex security decision-making. This intermediary processing layer analyzes facial features, compares them against databases, and translates raw image data into actionable security responses, managing system complexity through modular software architecture.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Measurement precision

If facial comparison is performed continuously, then detection accuracy is improved, but energy consumption increases

Engineering Contradiction:

Improvefacial recognition accuracyVSAvoidenergy consumption for continuous monitoring

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The system performs facial recognition analysis periodically or on-demand rather than continuously. The camera captures images at intervals or when triggered by specific conditions (such as detecting motion near the device), and machine learning analysis is performed only on captured frames, reducing energy consumption while maintaining detection accuracy through strategic sampling.

Inventive Principle:

Principle #19Periodic action

Solution Approach 2:

The system performs partial facial analysis by focusing computational resources on key facial features rather than analyzing entire images in full detail. The machine learning model extracts and compares critical biometric data points (such as facial geometry and distinctive features) without processing every pixel, achieving sufficient accuracy with reduced computational and energy expenditure.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS11244148B2Automated detection of unauthorized access to device screens using machine learning facial recognition techniques

Publication Date: 2022.02.08 DELL PROD LP

AI summary

Methods, apparatus, and processor-readable storage media for automated detection of unauthorized access to device screens using machine learning facial recognition techniques are provided herein. An example computer-implemented method includes automatically detecting, using one or more machine learning facial recognition techniques, at least a portion of at least one face within a given proximity of a user device; automatically performing a comparison, using one or more machine learning facial recognition techniques, of the detected portion of the at least one face with facial images attributed to a set of trusted individuals; and automatically performing security-related actions with respect to the user device upon determining, based at least in part on the comparison, that the detected portion of the at least one face exhibits at least a threshold level of distinctiveness relative to each of the facial images attributed to the set of trusted individuals.