Tomographic Spicula Detection via Center Position Variation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods struggle to accurately detect spicula candidates for lesions in breast cancer diagnosis using tomosynthesis imaging due to the difficulty in distinguishing them from normal mammary glands, and three-dimensional analysis is time-consuming.
Innovation Solution
An image processing device and method that utilizes a processor to analyze tomographic images for spicula candidates by comparing the center position variation of radial structures across multiple images, and generates composite two-dimensional images to enhance detection accuracy and reduce processing time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If three-dimensional analysis is performed to detect spicula candidates from volume data, then detection accuracy is improved, but processing time increases significantly
Solution Approach 1:
The patent segments the three-dimensional volume data into multiple two-dimensional tomographic images at different depth positions. By analyzing these 2D slices sequentially and comparing spicula candidate positions across slices, the system achieves accurate spicula detection without the computational burden of full 3D analysis, thus reducing processing time while maintaining detection accuracy.
Solution Approach 2:
The patent creates composite two-dimensional images that copy and integrate information from multiple tomographic slices. These composite images represent the spatial distribution of spicula candidates across different depths, enabling accurate detection through 2D comparison rather than computationally intensive 3D processing.
2Measurement precision
If multiple tomographic images are analyzed to distinguish spicula from normal mammary glands, then discrimination accuracy is improved, but device complexity increases
Solution Approach 1:
The patent utilizes the depth dimension by analyzing spicula candidate positions across multiple tomographic slices at different z-positions. Normal mammary glands exhibit position shifts across slices, while true spicula maintain consistent positions. This dimensional approach enables discrimination without requiring complex device modifications.
Solution Approach 2:
The patent performs preliminary detection of spicula candidates in each individual tomographic slice before conducting the comparative analysis across slices. This preliminary action simplifies the subsequent discrimination process by pre-identifying potential candidates, reducing the complexity of the overall analysis.
3Productivity
If composite two-dimensional images are generated from multiple tomographic images, then interpretation efficiency is improved, but additional processing steps are required
Solution Approach 1:
The patent merges multiple tomographic images into a single composite two-dimensional image that integrates spicula candidate information from all slices. This merging process uses maximum intensity projection or similar techniques to create a unified view, improving interpretation efficiency by radiologists while automating the processing steps to minimize additional complexity.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
Provided are an image processing device, an image processing method, and an image processing program that can accurately detect a spicula in a short processing time. An image processing device (4) detects a spicula candidate region (K21, K31, K32, K41, K42, K51) having a radial line structure from each of a plurality of tomographic images indicating a plurality of tomographic planes of an object and determines whether or not the spicula candidate region is a spicula on the basis of an amount of change of a center position of the line structure included in the spicula candidate region between the tomographic planes.