Multi-Directional 3D Colon Imaging for Hidden Polyp Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for assisting image diagnosis using 3D data, such as 3DCT images, struggle to adequately observe complex structures like the inner walls of the large intestine, particularly when diseased portions are hidden by folds or when fine details are difficult to discern from a single direction, leading to inaccurate shape and size assessments of polyps and other features.
Innovation Solution
A diagnosis assisting apparatus and method that includes volume data storage, overview image display, target point setting, target volume setting, line of sight adjustment, and detailed image projection, allowing for the generation of detailed images from various angles by setting target volumes and adjusting line of sight vectors to provide accurate and comprehensive views of structures, thereby overcoming the limitations of single-direction observation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If single-direction observation is used to generate images, then image generation time is reduced and simplicity is maintained, but hidden portions of structures cannot be observed and diagnostic accuracy deteriorates
Solution Approach 1:
The patent transitions from single-direction 2D observation to multi-directional 3D observation by setting multiple lines of sight from different directions and positions. This allows structures to be observed from multiple angles simultaneously, revealing hidden portions that would be invisible in single-direction views without significantly increasing processing time.
Solution Approach 2:
The patent divides the observation task into multiple lines of sight, each targeting specific regions or angles. By segmenting the viewing directions and positions, the system comprehensively covers structures that would be hidden in any single view, improving diagnostic accuracy while maintaining efficient image generation.
2Measurement precision
If multiple directions are used to observe structures, then diagnostic accuracy is improved, but image generation complexity and time increase
Solution Approach 1:
The patent efficiently implements multi-directional observation by utilizing the third dimension (depth/Z-axis) in addition to the standard two viewing directions. This allows structures to be observed from multiple angles without proportionally increasing complexity, as the additional spatial dimension provides natural viewing variations.
Solution Approach 2:
The patent creates images that serve multiple diagnostic functions simultaneously. Each generated image incorporates information from multiple lines of sight, allowing a single image to provide both overall structure visualization and detailed feature observation from different angles, reducing the need for multiple separate images.
3Ease of manufacture
If structures are observed from a single direction, then image processing is simplified, but fine details and hidden features cannot be discerned
Solution Approach 1:
The patent adds viewing directions from the depth dimension (Z-axis) to the conventional two-directional viewing system. This enables fine details and hidden features to be observed from multiple angles, improving detail discrimination capability while extending the processing framework rather than fundamentally complicating it.
Solution Approach 2:
The patent segments the observation into multiple lines of sight, each optimized for detecting specific features or details. This segmentation allows fine details to be captured from optimal angles while maintaining systematic and manageable image processing through structured multi-directional acquisition.
Data Source
Figure 1
Figure 2~3
Figure 4A~4D
AI summary
Structures in the vicinity of a target point (point of interest) within a diagnostic image are enabled to be observed in detail from multiple directions. Overview images (8, 9) that represent the overview of a structure (e.g., the large intestine) are generated based on volume data, and displayed on the screen (5) of a display. Points within the overview images and points corresponding thereto in the volume data are set as target points. Target volumes that include the target points and line of sight vectors within the volume data having the target points as endpoints and a movable viewpoint as a starting point are set within the volume data. The directions of the line of sight vectors are changed, and the target volumes are projected onto projection planes perpendicular to the directions of the line of sight vectors to generate detailed images (11) that represent details of the structure in the vicinity of the target points. The detailed images (11) which are generated accompanying changes in the line of sight direction are displayed sequentially or alongside each other on the screen.