Specimen Classification Networks for Hemolysis, Icterus, and Lipemia

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing automated diagnostic systems face challenges in accurately and efficiently determining the presence and degree of hemolysis, icterus, and lipemia in serum or plasma portions of blood specimens, which can lead to erroneous test results due to subjective manual inspection and complex, computationally burdensome image processing methods.

Innovation Solution

A method and apparatus using a segmentation convolutional neural network (SCNN) with a deep semantic segmentation network (DSSN) to capture multi-spectral, multi-exposure images from multiple viewpoints, followed by verification networks to predict and verify the classification of hemolysis, icterus, and lipemia, enabling robust characterization of specimen integrity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If manual visual inspection is used to evaluate specimen integrity, then the process is simple and requires minimal equipment, but it is subjective, labor intensive, and prone to human error

Engineering Contradiction:
Improvesimplicity of inspection processVSAvoidobjectivity and accuracy of inspection
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The patent replaces manual visual inspection with an automated image capture device and machine learning classification system. The image capture device objectively captures specimen images, and the machine learning model automatically classifies specimen integrity (normal, hemolysis, icterus, lipemia), eliminating human subjectivity and error while maintaining operational simplicity through automation.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If automated machine-vision inspection is used to evaluate specimen integrity, then objectivity and automation are improved, but the system becomes more complex and computationally burdensome

Engineering Contradiction:
Improveobjectivity of inspectionVSAvoidcomplexity of automated inspection system
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the inspection process into distinct components: image capture, image processing, machine learning classification, and verification. This segmentation allows each component to be optimized independently and reduces overall system complexity by dividing the automated inspection into manageable functional blocks.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent uses machine learning models that process image parameters (color, texture, intensity) to classify specimen integrity. By transforming visual characteristics into quantifiable parameters and using trained neural networks to interpret them, the system achieves reliable automated inspection without requiring overly complex hardware or computational infrastructure.

Inventive Principle:
Principle #35Parameter changes

3Device complexity

If manual visual inspection is used, then equipment requirements are minimal, but time consumption and labor requirements increase

Engineering Contradiction:
Improvesimplicity of equipmentVSAvoidinspection speed and efficiency
Core Design Contradiction:
Device complexityVSProductivity

Solution Approach 1:

The patent replaces manual visual inspection with automated image capture and machine learning classification systems. This substitution dramatically increases inspection speed and productivity while reducing labor requirements, as the automated system can rapidly process multiple specimens without human intervention.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

4Measurement precision

If automated image processing is used to characterize specimen integrity, then inspection accuracy is improved, but processing time and computational resources increase

Engineering Contradiction:
Improveaccuracy of specimen characterizationVSAvoidprocessing time for image analysis
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary actions by capturing images at standardized viewpoints and pre-processing them through automated image enhancement and normalization techniques. This preparation before classification allows the machine learning model to quickly and accurately characterize specimen integrity, reducing overall processing time while maintaining high accuracy.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3853616B1Hypothesizing and verification networks and methods for specimen classification
Publication Date: 2025.07.23 SIEMENS HEALTHCARE DIAGNOSTICS INC
  • EP3853616B1 patent drawingFigure 1A~1B
  • EP3853616B1 patent drawingFigure 1C~2
  • EP3853616B1 patent drawingFigure 3A

AI summary

A method of characterizing a specimen as containing hemolysis, icterus, or lipemia is provided. The method includes capturing one or more images of the specimen, wherein the one or more images include a serum or plasma portion of the specimen. Pixel data is generated by capturing the image. The pixel data of the one or more images of the specimen is processed using a first network executing on a computer to predict a classification of the serum or plasma portion, wherein the classification comprises hemolysis, icterus, and lipemia. The predicted classification is verified using one or more verification networks. Quality check modules and specimen testing apparatus adapted to carry out the method are described, as are other aspects.