Lateral Flow Assay Reading with Multi-Image Intensity Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for reading lateral flow tests (LFTs) in point-of-care settings face challenges due to variability in smartphone imaging hardware, requiring accessories or hardware modifications that are impractical during pandemic responses, and image processing techniques introduce errors and noise.
Innovation Solution
A method that captures multiple images of the LFT under varying conditions, groups intensity values into clusters, selects the cluster with the smallest variance, and calculates a mean intensity value to ensure accurate readings without additional hardware or modifications, using algorithms like local outlier factor and convolutional neural networks to filter out anomalies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple images are captured and processed using traditional image processing techniques, then measurement precision may be improved, but errors and noise are introduced
Solution Approach 1:
The patent extracts only the relevant intensity information from captured images by identifying and isolating the test line region, removing irrelevant background and artifacts before analysis. This extraction approach eliminates processing errors that would arise from analyzing entire images containing noise and irrelevant features.
Solution Approach 2:
The patent captures multiple copies (images) of the assay under varying conditions and uses clustering to identify the consistent signal across copies. By comparing multiple copies rather than processing a single image with potential artifacts, the method achieves accurate measurement without introducing processing errors.
2Measurement precision
If accessories or hardware modifications are used to control imaging conditions, then measurement precision is improved, but device complexity and ease of manufacture worsen
Solution Approach 1:
The patent employs dynamic capture of multiple images under varying imaging conditions (different lighting, angles, times) rather than requiring static controlled conditions. This dynamic approach allows the system to adapt to different hardware configurations without requiring complex hardware controls, achieving consistent measurements across diverse smartphone devices.
Solution Approach 2:
The patent changes imaging parameters (lighting conditions, capture angles, timing) across multiple images and uses computational clustering to identify consistent signals. This parameter variation strategy eliminates the need for fixed hardware controls while maintaining measurement precision across different device configurations.
3Productivity
If traditional image processing is used to analyze assay images, then productivity is maintained, but measurement precision deteriorates due to introduced errors
Solution Approach 1:
The patent segments the image analysis process into distinct stages: capturing multiple images, extracting intensity values from specific regions, clustering intensity values to identify consistent signals, and determining final results. This segmentation allows efficient processing while maintaining precision by focusing computational effort only on relevant data.
Solution Approach 2:
The patent replaces traditional mechanical/optical image processing methods with computational clustering algorithms that analyze intensity value distributions. This substitution eliminates processing errors inherent in traditional image filtering and enhancement techniques while maintaining high processing throughput through efficient computational methods.
Data Source
AI summary
A method for reading a test region of an assay includes: capturing a plurality of images of an assay with an imaging device; from each image of the plurality of images, extracting a region of interest comprising pixels of the image associated with a test region of the assay; from each extracted region of interest, estimating respective intensity values of at least a portion of the pixels; grouping the estimated intensity values into one or more clusters, said grouping comprising determining a total number of intensity values grouped into each cluster and a variance of each cluster; selecting the cluster having a total number of intensity values at or above a predetermined threshold and a smallest variance; calculating a mean intensity value of the selected cluster; and outputting the calculated mean intensity value as a result of the assay.