Pseudo 2D Breast Image Generation Using Compression-State Matched Training

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image generation models trained using a combination of tomosynthesis imaging and normal two-dimensional images captured in different compression states result in decreased accuracy of pseudo two-dimensional image generation.

Innovation Solution

An image generation apparatus and method that trains an image generation model using a combination of projection images or tomographic images and normal two-dimensional images captured in the same compression state, with categories such as person under examination, imaging settings, and image processing techniques, to improve generation accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If image generation models are trained using a combination of tomosynthesis imaging and normal two-dimensional images captured in different compression states, then training data availability is improved, but generation accuracy of pseudo two-dimensional image deteriorates

Engineering Contradiction:
Improvetraining data availabilityVSAvoidgeneration accuracy
Core Design Contradiction:
Quantity of substanceVSManufacturing precision

Solution Approach 1:

The patent applies local quality by categorizing training data according to compression states and matching them correspondingly. Instead of uniformly treating all training images, the system selectively pairs tomosynthesis images with normal two-dimensional images that share the same compression state, thereby improving generation accuracy while maintaining adequate training data availability through targeted data selection.

Inventive Principle:
Principle #3Local quality

2Ease of operation

If images in different compression states are used for training, then ease of data collection is improved, but consistency of breast state representation deteriorates

Engineering Contradiction:
Improveease of data collectionVSAvoidconsistency of breast state
Core Design Contradiction:
Ease of operationVSStability of the object's composition

Solution Approach 1:

The patent employs parameter changes by introducing compression state as a critical matching parameter. The system identifies and pairs images based on their compression state parameters, ensuring that tomosynthesis images are trained with normal two-dimensional images captured in the same compression state. This parameter-based matching maintains consistency of breast state representation while still allowing flexible data collection across various compression conditions.

Inventive Principle:
Principle #35Parameter changes

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

Enhances the accuracy of pseudo two-dimensional image generation by aligning training data conditions, allowing for more precise reproduction of breast lesion shapes.

Implementation Method 1

generates a pseudo two-dimensional image from a series of a plurality of projection images obtained by tomosynthesis imaging

Methodology Applied
Scientific EffectImage processing: Image Processing

Data Source

PatentEP4679375A1Image generation apparatus, image generation method, and program
Publication Date: 2026.01.14 FUJIFILM CORP
  • EP4679375A1 patent drawingFigure 1~2
  • EP4679375A1 patent drawingFigure 3
  • EP4679375A1 patent drawingFigure 4~5

AI summary

Provided are an image generation apparatus (16), an image generation method, and a program capable of improving generation accuracy of a pseudo two-dimensional image (102; 103), as compared with a case of using an image generation model that has been trained using a combination of a plurality of projection images or a plurality of tomographic images and a normal two-dimensional image captured separately from tomosynthesis imaging for obtaining the plurality of projection images or the plurality of tomographic images. An image processing apparatus as an image generation apparatus includes an image generation model (67) that has been trained in advance using a plurality of combinations of a plurality of projection images or a plurality of tomographic images (100; 101) and a normal two-dimensional image (111) captured by irradiating, with radiation (R), a breast (U) in a state of being compressed by a compression member (30) during tomosynthesis imaging for obtaining the plurality of projection images or the plurality of tomographic images.