Eye Gaze Estimation Using Attention Maps and Pixel Distance

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing gaze estimation technologies face challenges in accurately estimating and understanding user attention, particularly in scenarios such as augmented reality (AR) and human-computer interaction, where existing technologies fail to provide efficient and accurate methods for gaze estimation. Existing methods fail to provide efficient and accurate methods for gaze estimation, especially in complex environments.

Innovation Solution

The method involves obtaining target information from an image, including an eye, and using a target feature map to perform gaze estimation. This includes using attention information, temporal relationship information, and distance between pixels, along with frontal facial features, to enhance the accuracy of gaze estimation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional gaze estimation methods are used, then the system is simple to implement, but the gaze estimation accuracy is insufficient in complex environments

Engineering Contradiction:
Improvegaze estimation accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the gaze estimation process into multiple independent modules: face detection module, eye region extraction module, feature point detection module, and gaze calculation module. Each module processes specific aspects of the image independently, allowing the system to achieve high accuracy through coordinated processing of multiple features while maintaining manageable system complexity through modular architecture.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transitions from traditional 2D image processing to 3D spatial reasoning by detecting feature points in three-dimensional space and calculating gaze direction based on spatial relationships between pupils, eye centers, and face geometry. This dimensional enhancement enables more accurate gaze estimation in complex environments by incorporating depth and spatial orientation information.

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

2Measurement precision

If more features are used for gaze estimation, then the accuracy improves, but the processing time increases

Engineering Contradiction:
Improvegaze estimation accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary face detection and eye region extraction before detailed feature analysis. By pre-identifying the eye regions and key facial landmarks in advance, the system prepares processed data structures that accelerate subsequent gaze calculation, reducing the processing time required for the computationally intensive feature analysis and gaze direction computation.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent applies different processing strategies to different regions of the image: full-face processing for face detection, focused eye region extraction for pupil detection, and selective feature point analysis for gaze calculation. This localized processing approach ensures high accuracy in critical regions while reducing unnecessary computations in less important areas, optimizing the balance between accuracy and processing time.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP4191545B1Device and method with gaze estimating
Publication Date: 2025.12.31 SAMSUNG ELECTRONICS CO LTD
  • EP4191545B1 patent drawingFigure 1
  • EP4191545B1 patent drawingFigure 2
  • EP4191545B1 patent drawingFigure 3

AI summary

An electronic device and method with gaze estimating are disclosed. The method includes obtaining target information of an image, the image including an eye, obtaining a target feature map representing information on the eye in the image based on the target information, and estimating a gaze for the eye in the image based on the target feature map. The target information includes either attention information on the image, or a distance between pixels in the image, or both.