Specimen Analyzer Workflow for AI Load and Throughput Balance

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing specimen analysis methods using artificial intelligence algorithms face increased computational load due to large data volumes, particularly when analyzing specimens with multiple components and increased sample sizes, leading to inefficiencies and prolonged analysis times.

Innovation Solution

A specimen analyzer that apportions analysis processes between artificial intelligence algorithms and calculation processing operations, utilizing both AI analysis and non-AI calculation processing to reduce computational load, allowing for parallel sample preparation and improved throughput.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If all data corresponding to optical signals is analyzed using only artificial intelligence algorithms, then classification accuracy is improved, but computational load on the computer increases significantly

Engineering Contradiction:
Improveclassification accuracyVSAvoidcomputational load
Core Design Contradiction:
Measurement precisionVSPower

Solution Approach 1:

The patent segments the analysis process into two distinct pathways: AI-based analysis for specific measurement items requiring high classification accuracy, and calculation processing for other measurement items. This segmentation allows the system to apply computational resources selectively, reducing overall computational load while maintaining accuracy where needed.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different analysis methods to different measurement items based on their specific requirements. AI algorithms are applied locally to measurement items where high classification accuracy is critical, while calculation processing is used for items where it suffices, optimizing the balance between accuracy and computational efficiency.

Inventive Principle:
Principle #3Local quality

2Measurement precision

If the amount of information obtained from single components is increased to improve classification accuracy, then classification accuracy is improved, but data volume per specimen increases

Engineering Contradiction:
Improveclassification accuracyVSAvoiddata volume
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent segments information processing by applying AI algorithms only to specific measurement items rather than processing all obtained information uniformly. This selective approach extracts essential features for classification without unnecessarily increasing data volume from all components.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent extracts and processes only the necessary information for classification purposes using AI algorithms, rather than analyzing all available data. This extraction approach obtains sufficient information for accurate classification while avoiding the computational burden of processing excessive data.

Inventive Principle:
Principle #2Taking out (Extraction)

3Productivity

If the number of specimens to be tested is increased, then productivity is improved, but total data volume and computational load increase

Engineering Contradiction:
ImprovethroughputVSAvoidcomputational load
Core Design Contradiction:
ProductivityVSPower

Solution Approach 1:

The patent segments the analysis workflow into parallel processing streams: AI-based analysis for specific measurement items and calculation processing for others. This segmentation enables efficient handling of multiple specimens simultaneously, improving throughput while controlling computational load through method diversification.

Inventive Principle:
Principle #1Segmentation

4Measurement precision

If only artificial intelligence algorithms are used for analysis, then classification accuracy is improved, but analysis time is prolonged due to computational complexity

Engineering Contradiction:
Improveclassification accuracyVSAvoidanalysis time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent segments the analysis process into AI-based analysis and calculation processing, allowing computationally intensive AI operations to be applied only where necessary for high classification accuracy, while faster calculation methods handle other items, thereby reducing overall analysis time.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies AI algorithms partially only to specific measurement items where high classification accuracy is required, rather than applying them excessively to all measurement items. This partial application reduces computational time while maintaining necessary accuracy for critical classifications.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP4246123B1Specimen analyzer, specimen analysis method, and program
Publication Date: 2025.11.19 SYSMEX CORP
  • EP4246123B1 patent drawingFigure 1
  • EP4246123B1 patent drawingFigure 2
  • EP4246123B1 patent drawingFigure 3

AI summary

Disclosed is a specimen analyzer for analyzing an analyte in a specimen, the specimen analyzer including: a measurement unit including a plurality of first sample preparation parts each configured to prepare a first measurement sample on the basis of the specimen and a first reagent, a second sample preparation part configured to prepare a second measurement sample on the basis of the specimen and a second reagent, and an optical detection part configured to obtain a first optical signal from the first measurement sample and obtain a second optical signal from the second measurement sample; and an analysis unit configured to analyze first data that corresponds to the first optical signal and second data that corresponds to the second optical signal, wherein the analysis unit executes analysis of a first measurement item with respect to the first measurement sample, through a first analysis operation of processing the first data according to an artificial intelligence algorithm, executes analysis of a second measurement item with respect to the first measurement sample, through at least one of the first analysis operation and a second analysis operation of processing a first representative value, of the first data, that corresponds to a feature of the analyte, and executes analysis with respect to the second measurement sample, through a third analysis operation of processing a second representative value, of the second data, that corresponds to a feature of the analyte.