Platelet Counting via MSER Image Segmentation and Shape Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for accurately counting platelets in blood samples are hindered by high variability in platelet size, interference from erythrocytes and contaminants, and errors in automated analysis, leading to inaccurate results and the need for manual verification, especially in pathological cases.
Innovation Solution
The method involves preprocessing microscopic images of peripheral blood smears using the maximally stable external regions (MSER) algorithm for segmentation and analysis, including bright and dark region detection, convex hull calculation, filtering by shape, and classification based on circularity and mask area to differentiate platelets from other blood constituents.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated hematology analysers use impedance method for platelet counting, then productivity is improved, but measurement precision deteriorates due to differentiation errors between erythrocytes and thrombocytes
Solution Approach 1:
The patent applies fluorescence staining to platelets, causing them to emit light at specific wavelengths when illuminated. This optical property change allows automated analysers to differentiate platelets from erythrocytes and white blood cells based on their fluorescence characteristics, thereby maintaining high counting speed while significantly improving measurement precision.
Solution Approach 2:
The patent changes the detection parameter from simple impedance (electrical resistance) to fluorescence emission intensity and wavelength. By measuring the optical properties of stained platelets rather than their electrical impedance, the system achieves better differentiation between cell types and more accurate platelet counting without sacrificing productivity.
2Measurement precision
If manual counting method (FONIO) is used for platelet estimation, then measurement precision is improved, but productivity deteriorates due to time-consuming manual counting
Solution Approach 1:
The patent uses fluorescence microscopy to create optical copies (images) of the blood smear with stained platelets. Automated image processing algorithms then analyze these copies to count platelets, combining the precision of manual visual inspection with the speed of automated processing. The system effectively creates a digital replica that can be analyzed without manual intervention.
Solution Approach 2:
The patent replaces the mechanical manual counting process with an automated optical system. Instead of manually examining and counting platelets under a microscope, the system uses fluorescence illumination and automated image analysis to perform the counting, eliminating the time-consuming manual operation while preserving accuracy.
3Productivity
If low magnification is used in automated platelet counting, then productivity is improved by analyzing larger area, but measurement precision deteriorates due to cell clustering and shape similarity with white blood cells
Solution Approach 1:
The patent uses fluorescence staining that causes platelets to emit light at characteristic wavelengths distinct from white blood cells. This optical differentiation allows the system to reliably distinguish platelets from white blood cells even at low magnification where cells appear smaller and potentially clustered, maintaining measurement precision while preserving the productivity benefits of low-magnification wide-field imaging.
Data Source
AI summary
A method for analysing microscopic images of a blood smear allowing to determine the number of thrombocytes in a tested sample. The present disclosure uses an algorithm which allows for differentiating between platelets from other blood cells (including erythrocytes), and then counts the quantity of thrombocytes (number/μl) and determines their size (μm). The method includes providing a grayscale microscopic image of platelets, segmenting and analysing the image, wherein the step of segmentation and analysis of the image comprises analysing light regions of the image and analysing dark regions of the image, including detecting distinctive regions in the image using a maximally stable external regions algorithm; calculating for each light and dark region identified its convex hull and filtering the results obtained by shape; removing nesting regions; identifying aggregates; classifying cells into platelets and other blood constituents; and determining the number of platelets and masks thereof.