Anomaly Detection Using Aggregated Eye-Tracking Data

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current anomaly detection methods fail to effectively utilize eye-tracking data to identify user-specific and user-independent changes over time and location, leading to incomplete understanding of spatial familiarity and environmental awareness.

Innovation Solution

An apparatus and method that obtain and compare eye-tracking data from multiple users across various locations and times, using machine learning algorithms to detect anomalies by aggregating user-specific changes and determining location states based on these changes, while considering additional data such as fatigue, alcohol consumption, and social media data to refine familiarity measures.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If eye-tracking data from multiple users across multiple locations and times is collected and analyzed, then anomaly detection accuracy and spatial familiarity understanding are improved, but system complexity and data processing requirements increase

Engineering Contradiction:
Improveanomaly detection accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system segments the analysis by separating user-specific changes from user-independent changes. It processes eye-tracking data for each user individually to identify their specific patterns, then aggregates these to detect anomalies that affect multiple users. This segmentation allows accurate anomaly detection without requiring complex simultaneous analysis of all user data together.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system adds temporal and spatial dimensions to the analysis by collecting eye-tracking data across multiple times and multiple user locations. Instead of analyzing single-point data, it examines changes over time and across different locations, enabling detection of patterns that would be invisible in static analysis. This multi-dimensional approach improves anomaly detection accuracy.

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

2Measurement precision

If additional data such as fatigue indicators, alcohol consumption indicators, and social media data are integrated, then user familiarity measurement accuracy is improved, but data processing complexity and computational requirements increase

Engineering Contradiction:
Improveuser familiarity measurement accuracyVSAvoiddata processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system merges multiple data sources including eye-tracking data, fatigue indicators, alcohol consumption indicators, and social media data into a unified analysis framework. By combining these diverse data types, it creates a comprehensive view of user familiarity and environmental awareness, improving measurement accuracy through multi-factor analysis.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The system uses machine learning algorithms as intermediaries to process and integrate the diverse data sources. These algorithms automatically weigh and combine the different data types (eye-tracking metrics, physiological indicators, social media information) to derive meaningful familiarity measures, reducing the manual complexity of data integration.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP4390626A1Apparatus, system, method and computer program for anomaly detection
Publication Date: 2024.06.26 NOKIA TECHNOLOGIES OY
  • EP4390626A1 patent drawingFigure 1
  • EP4390626A1 patent drawingFigure 2
  • EP4390626A1 patent drawingFigure 3A~3B

AI summary

An apparatus comprising means for: for multiple users: obtaining eye-tracking data for at least one of: multiple times and multiple user locations; comparing the eye-tracking data for at least one of: the multiple times and the multiple user locations to identify a user-specific change with respect to at least one of time or user location; and automatically detecting an anomaly, comprising aggregating the user-specific changes to identify a user-independent change with respect to at least one of: time or location.