Pupil Identification via Connected Graph Weighting

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for pupil identification in images, such as those used in eyeball tracking systems, face challenges in accurately locating the pupil without manual judgment or external apparatus, and lack flexibility in adapting to varying image conditions.

Innovation Solution

A method and device that preprocess images to build a grey-scale image, detect edges, and identify pupils by categorizing connected graph pairs based on grey-scale values, calculating centroid locations and axis ratios, and selecting the pair with the maximum weighted value, allowing for flexible adaptation to image variations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional pupil identification methods are used, then the system can locate pupil features, but the accuracy is insufficient without manual judgment or external apparatus

Engineering Contradiction:
Improvepupil location accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The method segments the image processing into distinct stages: grey-scale conversion, edge detection, connected component analysis, and pupil identification. By dividing the complex task of pupil localization into these manageable segments, the system achieves high accuracy without requiring manual intervention or external apparatus, resolving the contradiction between measurement precision and device complexity

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The method performs preliminary actions by pre-processing the image through grey-scale conversion and edge detection before pupil identification. Connected graphs are pre-marked and categorized by grey-scale value ranges (100-200 and over 200), creating a structured foundation that enables accurate pupil localization without complex real-time processing or external devices

Inventive Principle:
Principle #10Preliminary action

2Adaptability or versatility

If conventional pupil identification methods are used, then basic pupil detection is possible, but the adaptability to varying image conditions is limited

Engineering Contradiction:
Improveadaptability to image variationsVSAvoidpupil identification accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The method adapts to varying image conditions by changing parameters: it categorizes connected graphs into two grey-scale value ranges (100-200 and over 200), calculates multiple geometric parameters (centroid location, major axis length, minor axis length), and uses a composite evaluation formula W=L-m*R-n*P that incorporates distance, ratio, and positional information. This multi-parameter approach enables the system to maintain high precision across different image conditions

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The method introduces dynamic adaptability through the weighted evaluation formula where m and n are adjustable weight factors. This allows the system to dynamically adjust the importance of different parameters (distance L, ratio R, position P) based on specific image conditions, enabling both high adaptability and maintained precision

Inventive Principle:
Principle #15Dynamics

3Measurement precision

If detailed image processing is performed to improve pupil identification accuracy, then the location precision increases, but the processing time increases

Engineering Contradiction:
Improvepupil location precisionVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The method extracts only the essential features needed for pupil identification: edge pixels are extracted to form connected graphs, and only relevant geometric parameters (centroid, major axis, minor axis) are calculated. By extracting and focusing on these key features rather than processing the entire image in detail, the system achieves high precision while minimizing processing time

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3376431B1Method and apparatus for identifying pupil in image
Publication Date: 2023.04.19 CHINA UNIONPAY
  • EP3376431B1 patent drawingFigure 1
  • EP3376431B1 patent drawingFigure 2

AI summary

A method and device for identifying a pupil in an image is provided. The method comprises four steps including: A. image preprocessing, B. edge detecting, C. building connected graphs, and D. screening connected graph pairs.