Eye-Gaze Data Capture Layout for Screen Edge Accuracy

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional eye-gaze tracking systems face accuracy degradation at screen edges and corners due to insufficient training data and non-uniform distribution, leading to less accurate eye-gaze predictions, particularly when interactive system icons and buttons are located there.

Innovation Solution

A grid-based approach is employed on the screen with regions of varying sizes and a non-overlapping scan path that passes through the centers of these regions, capturing eye-gaze data uniformly across the screen, including more data at edges and corners, using multiple cameras for enhanced accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If evenly distributed eye-gaze training data is collected across the screen, then the calibration process covers the entire screen area, but the prediction accuracy degrades at screen edges and corners due to insufficient neighboring data points

Engineering Contradiction:
Improveeye-gaze prediction accuracyVSAvoidtraining data density at edges and corners
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent applies local quality by creating overlapping regions centered at each calibration point, where each region has a defined radius that ensures sufficient neighboring data points. This local densification around calibration points (particularly at edges and corners) improves prediction accuracy in these critical areas without requiring uniform oversampling across the entire screen, thus resolving the contradiction between overall coverage and local data density.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent performs preliminary action by pre-defining a set of calibration points across the screen before data collection begins. These calibration points are strategically positioned to ensure that edges and corners are adequately represented. By establishing this predetermined grid of calibration points with associated overlapping regions, the system prepares the data collection framework in advance to capture sufficient training data at critical screen boundaries.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If more training data is collected to improve accuracy at screen boundaries, then the prediction accuracy improves, but the calibration time increases causing stress on the operator

Engineering Contradiction:
Improveeye-gaze prediction accuracy at edges and cornersVSAvoidcalibration time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent segments the screen into multiple overlapping regions, each centered at a calibration point. This segmentation allows the system to collect and process eye-gaze data in localized zones rather than requiring comprehensive coverage of every screen location. The overlapping nature ensures edge and corner regions are adequately covered while limiting the total number of calibration points needed, thus reducing calibration time while maintaining accuracy at boundaries.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies partial action by collecting eye-gaze data only within defined overlapping regions around calibration points, rather than uniformly across the entire screen. This focused data collection approach concentrates sampling effort where it matters most (at and near calibration points, especially at edges and corners) while avoiding redundant collection in areas already well-covered, thereby reducing overall calibration time while maintaining or improving accuracy at critical boundaries.

Inventive Principle:
Principle #16Partial or excessive action

3Quantity of substance

If calibration points are densely distributed across the screen, then more training data is available for accurate prediction, but the complexity of the calibration process increases

Engineering Contradiction:
Improvetraining data availabilityVSAvoidcalibration process complexity
Core Design Contradiction:
Quantity of substanceVSDevice complexity

Solution Approach 1:

The patent segments the calibration process into discrete, manageable calibration points arranged in a grid pattern across the screen. Each calibration point has an associated overlapping region with a defined radius. This segmentation transforms the complex task of dense uniform sampling into a structured process with a limited number of predetermined locations, simplifying the calibration workflow while ensuring adequate data collection coverage.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary action by pre-calculating and storing the positions of calibration points and their associated overlapping regions before the calibration process begins. This predetermined structure eliminates the need for real-time complex calculations during data collection, reducing computational complexity and making the calibration process more efficient and easier to implement while still providing sufficient training data for accurate prediction.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4327184B1Systems and methods of capturing eye-gaze data
Publication Date: 2025.11.26 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP4327184B1 patent drawingFigure 1
  • EP4327184B1 patent drawingFigure 2A~2B
  • EP4327184B1 patent drawingFigure 2C~2D

AI summary

Systems and methods are provided for collecting eye-gaze data for training an eye-gaze prediction model. The collecting includes selecting a scan path that passes through a series of regions of a grid on a screen of a computing device, moving a symbol as an eye-gaze target along the scan path, and receiving facial images at eye-gaze points. The eye-gaze points are uniformly distributed within the respective regions. Areas of the regions that are adjacent to edges and corners of the screen are smaller than other regions. The difference in areas shifts centers of the regions toward the edges, density of data closer to the edges. The scan path passes through locations in proximity to the edges and corners of the screen for capturing more eye-gaze points in the proximity. The methods interactively enhance variations of facial images by displaying instructions to the user to make specific actions associated with the face.