Lesion Detection Model Selection by Imaging Apparatus Type

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing lesion detection techniques in medical imaging are inaccurate due to variations in medical image quality, leading to reduced detection accuracy.

Innovation Solution

An image processing device that selects a lesion detection model based on apparatus identification information, utilizing specific models for different imaging apparatus types or manufacturers, and includes a general-purpose model for unidentified apparatuses, and retrains models using ground truth labels to enhance accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If a single general-purpose lesion detection model is used for all imaging apparatus, then the device complexity is reduced, but the detection accuracy deteriorates due to variations in medical image quality across different imaging devices

Engineering Contradiction:

Improvenumber of lesion detection modelsVSAvoidlesion detection accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The patent divides the single general-purpose model into multiple apparatus-specific models, each trained on data from a specific imaging apparatus type. This segmentation allows each model to specialize in detecting lesions from images with characteristic quality patterns of its corresponding device, thereby improving detection accuracy while managing complexity through organized model selection

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system changes the parameter of model selection based on apparatus identification information. By detecting which imaging apparatus captured the input image and selecting the corresponding trained model, the system adapts the detection algorithm to match the specific image quality characteristics of different devices, resolving the contradiction between using multiple models and maintaining simplicity

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If multiple apparatus-specific lesion detection models are used, then the lesion detection accuracy is improved, but the device complexity increases due to needing to manage and select from multiple models

Engineering Contradiction:

Improvelesion detection accuracyVSAvoidnumber of lesion detection models

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system implements a feedback mechanism where the apparatus identification information from the input image automatically triggers selection of the corresponding pre-trained model. This feedback loop ensures the correct specialized model is used without requiring complex manual configuration or user intervention, managing the complexity of multiple models through automated model selection

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent creates a universal framework that can handle multiple apparatus-specific models through a single unified system architecture. The model selection mechanism serves multiple functions: identifying the imaging apparatus, selecting the appropriate model, and ensuring optimal detection accuracy, thereby managing complexity through multi-functional design

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Measurement precision

If a lesion detection model is trained on data from a specific imaging apparatus, then the detection accuracy for that apparatus is improved, but the adaptability to other imaging apparatus deteriorates

Engineering Contradiction:

Improvelesion detection accuracy for specific apparatusVSAvoidapplicability across different imaging apparatus

Core Design Contradiction:

Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent segments the training data and models by imaging apparatus type, creating specialized models for each device category. This segmentation improves accuracy for each specific apparatus while the overall system maintains versatility through the collection of segmented models, each optimized for its target apparatus type

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system performs preliminary training actions by pre-training separate models on data from different imaging apparatus types before deployment. This preliminary specialization allows each model to be highly accurate for its target apparatus, while the system as a whole remains adaptable to various devices through the pre-prepared collection of apparatus-specific models

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12524874B2Image processing device, learning device, image processing method, learning method, image processing program, and learning program that can accurately detect a lesion from a medical image using a lesion detection model

Publication Date: 2026.01.13 FUJIFILM CORP

AI summary

A learning device includes at least one processor. The processor acquires a medical image to be detected, acquires apparatus identification information for identifying an imaging apparatus that has captured the medical image to be detected, selects any one of a plurality of lesion detection models, which detect a lesion from the medical image, on the basis of the apparatus identification information, and detects the lesion from the medical image to be detected, using the selected lesion detection model.