Pupil Position Acquisition Using Wavelet Transform
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional pupil tracking methods are time-consuming and prone to noise interference, resulting in inaccurate pupil position detection due to the use of binary image processing and Hough transform techniques.
Innovation Solution
A pupil position acquisition system and method utilizing a wavelet transform to analyze eyeball image signals, which includes a shooting module, scanning module, and signal analysis module to quickly and accurately determine the pupil position through high and low frequency signal analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If binary image processing and Hough transform technique are used to detect pupil position, then the method is conventional and widely applicable, but the processing is time-consuming and noise interference is easily generated resulting in low accuracy
Solution Approach 1:
The patent changes the fundamental parameter of image processing from binary threshold-based methods to wavelet transform-based continuous analysis. By transforming the eyeball image into wavelet coefficients and analyzing frequency components, the system achieves faster processing without sacrificing accuracy, directly resolving the time-accuracy tradeoff in pupil detection
Solution Approach 2:
The patent substitutes the mechanical Hough transform algorithm with a wavelet transform-based signal processing approach. This replacement eliminates the computational complexity of line-by-line Hough voting while maintaining edge detection capability, thereby reducing processing time and noise sensitivity simultaneously
2Measurement precision
If binary image processing and Hough transform technique are used to detect pupil position, then the method is conventional and widely applicable, but noise interference is easily generated resulting in low accuracy
Solution Approach 1:
The patent transitions from binary threshold processing to multi-scale wavelet analysis, changing how image data is represented and processed. This parameter change allows the system to distinguish signal from noise across different frequency scales, significantly reducing noise interference while improving measurement precision
Solution Approach 2:
The patent segments the image processing into different frequency bands through wavelet decomposition. By analyzing high-frequency and low-frequency components separately, the system can identify pupil boundaries more accurately while filtering out noise that appears as unwanted frequency components
Data Source
AI summary
A system and pupil position acquisition method and a device containing computer software for executing the same are provided. The system includes a shooting module, a scanning module, a signal transformation module, and a signal analysis module. The shooting module shoots an eyeball image using an image shooting device, such as a charge-coupled device (CCD) camera. The scanning module scans the eyeball image to acquire an eyeball signal. The signal transformation module performs a wavelet transform on the eyeball signal. The signal analysis module analyzes the eyeball signal after the wavelet transform to acquire a signal interval, and analyzes and acquires a position of a pupil with respect to the eyeball image according to the signal interval.