Fluorescence Image Evaluation Using Quotient Distribution Thresholding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for recognizing diseased tissues from fluorescence images are adversely affected by stray light, leading to unreliable detection due to noise and interference.
Innovation Solution
A method and device that create a quotient image from partial images recorded at different wavelengths, computing a distribution spectrum to determine a threshold value independent of perturbing light, which is used to generate a contrast image enhancing the visibility of diseased tissue regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If fluorescence images are used to recognize diseased tissues, then tissue differentiation is possible, but the results are influenced by stray light and noise
Solution Approach 1:
A distribution spectrum serves as an intermediary between the raw fluorescence images and the final tissue classification. The spectrum characterizes the quotient image data and enables robust threshold determination that is independent of stray light conditions, thereby mediating between noisy measurements and reliable diagnosis
Solution Approach 2:
The evaluation method uses the distribution spectrum to determine optimal threshold values based on the actual data characteristics. This feedback mechanism allows the system to adapt to varying stray light conditions and noise levels, improving both measurement precision and reliability dynamically
2Measurement precision
If a quotient image is created from partial images at different wavelengths, then discrimination between healthy and diseased tissue is enhanced, but noise is amplified in low intensity regions
Solution Approach 1:
The method performs preliminary evaluation by creating a distribution spectrum from the quotient image data before final threshold-based classification. This preliminary characterization of the data distribution allows for optimal threshold selection that accounts for noise characteristics, preventing noise amplification in the final diagnostic decision
Solution Approach 2:
The invention changes the parameter space by transforming pixel intensity values into a distribution spectrum representation. This parameter transformation allows the system to work with statistical characteristics of the data rather than individual noisy pixel values, thereby maintaining discrimination capability while reducing noise impact
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The method effectively reduces noise and enhances the discrimination between healthy and diseased tissue regions, providing a reliable representation of diseased areas that may be obscured in original images, while maintaining sensitivity and dynamic range.
Implementation Method 1
If dental enamel or similar tissue is irradiated with blue or ultraviolet light, the tissue fluoresces, whereby in the case of healthy and diseased tissues different spectral components in the fluorescent light are observed
Data Source
AI summary
The invention relates to a method and a device for evaluating fluorescence partial images representing the same object region, independently of stray light. Two partial images of the objects are produced in the red and in the green by a camera (10). A quotient image is produced in pixels from said two partial images, and the frequency of the occurrence of the image points having a pre-defined red/green ratio is determined for said quotient image. The mean values and the width are determined for the distribution curve obtained in this way. The two end variables of the distribution curve are used to calculate a threshold value. The quotient image is modified using said threshold value such that its contrast in relation to interesting details is increased.