Left Right Eye Image Differentiation Using Neural Network Corner Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing eye image processing techniques are computationally expensive and inefficient, particularly in determining the type of eye (left or right) due to the use of brute force methods and significant resource consumption.
Innovation Solution
A system and method that uses neural network classifiers to detect inner and outer eye corners and determine the horizontal displacement between the iris and pupil centers to differentiate between left and right eye images, reducing computational complexity and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If brute force methods are used for eye corner and iris detection, then detection accuracy is maintained, but computational resources and processing time increase significantly
Solution Approach 1:
The patent segments the eye image processing into distinct functional modules: eye corner detection using neural networks, iris center detection, pupil center detection, and horizontal displacement calculation. Each module processes specific features independently, reducing the overall computational burden compared to brute force methods that analyze the entire image uniformly.
Solution Approach 2:
The patent changes the processing parameters by using neural network classifiers with optimized architectures and training data specifically designed for eye corner detection. This transforms the problem from general image processing to a specialized parameter-optimized task, achieving both accuracy and efficiency.
2Reliability
If multiple likely eye corner location templates are used, then detection reliability is improved, but computational complexity and resource consumption increase
Solution Approach 1:
The patent replaces the mechanical template-matching system with a neural network-based detection system. The neural network learns optimal detection patterns during training, eliminating the need for multiple pre-defined templates and reducing system complexity while maintaining or improving reliability.
Solution Approach 2:
The patent uses training data that copies realistic eye corner patterns from multiple sources, allowing the neural network to learn from diverse examples without requiring actual multiple templates during operation. This achieves reliability through learned patterns rather than explicit template storage.
3Loss of time
If traditional computational methods are used to determine eye type, then system constraints are met, but processing efficiency decreases
Solution Approach 1:
The patent performs preliminary detection of eye corners, iris center, and pupil center using optimized algorithms before the final eye type determination. This preliminary processing organizes the data in a way that makes the final classification computationally efficient, reducing the time required for the critical determination step.
Solution Approach 2:
The patent extracts only the essential features needed for eye type determination: the horizontal displacement between iris and pupil centers relative to eye corner positions. By extracting and using only these critical features rather than processing the entire image, the system achieves high processing efficiency.
Data Source
AI summary
A system and method for determining whether an eye image includes a left eye or a right eye image is described. After obtaining an eye image, the centers of the iris and pupil in the eye image are located. A horizontal displacement between the centers of the iris and pupil is determined. In addition, neural network classifiers are used to perform eye corner detection to identify inner and outer eye corners in the image. A determination is made as to whether the eye image includes a left eye image or a right eye image based on the detected eye corners and horizontal displacement.