Pupil Center Positioning via Histogram-Based Binarization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current eye-tracking technologies face low precision in positioning the center of the pupil due to inaccurate threshold definition methods, which are often based on experience rather than image characteristics.
Innovation Solution
An image processing method that acquires an initial image of a pupil, constructs an accumulative histogram set of grayscale values, determines a threshold for binarization based on the characteristics of the target region in the histogram, and performs binarization to enhance pupil positioning accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If threshold definition is based on experience rather than image characteristics, then the process is simple, but the precision of pupil center positioning deteriorates
Solution Approach 1:
The patent changes the parameter basis for threshold definition from empirical values to image-specific histogram characteristics. By analyzing the grayscale distribution in the image and identifying the target region through histogram peaks, the threshold adapts to each image's unique characteristics, thereby improving positioning precision while maintaining operational simplicity.
Solution Approach 2:
The system performs self-service by automatically determining the threshold based on the image's own histogram characteristics without requiring external empirical parameters. The algorithm extracts the target region's grayscale distribution from the image itself and uses this information to define the binarization threshold, eliminating the need for experience-based parameter setting.
2Measurement precision
If threshold definition uses image histogram characteristics, then the precision of pupil center positioning improves, but the computational complexity increases
Solution Approach 1:
The patent extracts only the essential information needed for threshold definition from the image - specifically the histogram of grayscale values and the identification of the target region's peak. By taking out only these critical features rather than processing the entire image data, the method achieves high positioning precision while keeping computational complexity manageable.
Solution Approach 2:
The patent segments the grayscale histogram into distinct regions, identifying the target region through peak detection. This segmentation approach divides the complex image data into manageable parts (grayscale ranges), allowing the algorithm to focus computation only on relevant segments and thereby reduce overall computational complexity while maintaining precision.
Data Source
AI summary
Embodiments of the present disclosure provides an image processing method, comprising: acquiring an initial image comprising an image of pupil; acquiring an accumulative histogram set of grayscale values of pixels in the initial image; determining a threshold for binarizing the initial image according to a grayscale value corresponding to a target region in respective accumulative histograms in the accumulative histogram set, wherein the target region corresponds to the image of pupil in the initial image; and binarizing the initial image according to the threshold. An image processing apparatus and a wearable device are further provided.