Pseudo 2D Breast Image Generation With Matched Compression States
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image generation models trained with combinations of images from tomosynthesis and normal two-dimensional imaging in different compression states result in decreased accuracy of pseudo two-dimensional image generation due to varying breast states, and the correlation between input and output information is not always high.
Innovation Solution
An image generation apparatus and method that trains an image generation model using paired images captured in the same compression state, combining composite two-dimensional images from tomosynthesis with normal two-dimensional images, and adjusts radiation parameters to improve accuracy, including classification and training for specific categories like breast thickness and imaging conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image generation models are trained with combinations of images from tomosynthesis and normal two-dimensional imaging in different compression states, then the training can be performed with available images, but the generation accuracy of pseudo two-dimensional images decreases due to varying breast states
Solution Approach 1:
The patent applies parameter changes by controlling the compression state parameter during image acquisition. Specifically, the breast compression amount is adjusted to match between tomosynthesis imaging and normal two-dimensional imaging, ensuring that images are captured under consistent compression conditions. This parameter control ensures high correlation between input (tomosynthesis images) and output (normal two-dimensional images) during model training, thereby improving the generation accuracy of pseudo two-dimensional images.
2Loss of information
If multiple images are input to the image generation model, then more information is provided for processing, but the correlation between input information and output information used for training is not always high
Solution Approach 1:
The patent extracts and selects specific images with high correlation from the multiple images obtained during tomosynthesis imaging. Instead of using all projection images, the system identifies and selects images captured at compression states that match the normal two-dimensional imaging conditions. This extraction of high-correlation images ensures that the training data maintains strong input-output relationships, improving generation accuracy while avoiding the use of low-correlation images.
3Ease of manufacture
If images captured in different compression states are used for training, then the model can be trained with readily available images, but the breast state such as size, thickness, shape, and position differs between images
Solution Approach 1:
The patent applies preliminary action by controlling the compression state before image acquisition. The compression amount is adjusted in advance to match between tomosynthesis imaging and normal two-dimensional imaging. By establishing consistent compression conditions before capturing the images, the patent ensures that breast state parameters (size, thickness, shape, position) remain consistent across different imaging modalities, thereby improving training effectiveness and generation accuracy.
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 generation accuracy of pseudo two-dimensional images by aligning input and output correlations, improving the model's ability to reproduce breast lesion shapes accurately.
Implementation Method 1
an image generation model that has been trained in advance using a plurality of combinations of the composite two-dimensional image and a normal two-dimensional image captured by irradiating, with radiation, the breast in a state of being compressed by the compression member during the tomosynthesis imaging for obtaining the composite two-dimensional image
Data Source
AI summary
An image processing apparatus as an image generation apparatus includes an image generation model that has been trained in advance using a plurality of combinations of a composite two-dimensional image and a normal two-dimensional image captured by irradiating, with radiation, a breast in a state of being compressed by a compression member during tomosynthesis imaging for obtaining the composite two-dimensional image.