Unfolded Rib Cage Manifold Imaging for Automatic Fracture Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for automatic rib fracture detection in medical imaging, such as CT scans, face challenges due to the trade-off between receptive field size and resolution, leading to issues like imaging artifacts, incorrect rib lengths, and difficulty in detecting subtle fractures, especially when using deep-learning-based approaches.
Innovation Solution
A new manifold view is introduced for rib fracture detection, where the rib cage is reformatted into a standardized two-dimensional manifold slice, allowing for consistent rib lengths and locations, and a trained fracture detection model is applied to this view to predict fractures accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deep-learning-based approaches are used with high image resolution to identify subtle rib fractures, then measurement precision is improved, but device complexity and computational resources increase due to the trade-off between input resolution and receptive field size
Solution Approach 1:
The patent transforms the 3D rib cage volume into a 2D unfolded manifold view, changing the dimensional representation to enable simultaneous high resolution and large receptive field. This dimensional transformation allows the model to process the entire rib cage in a single shot without compromising detection precision or increasing complexity
Solution Approach 2:
The patent segments the rib cage into individual rib structures and unfolds them onto a 2D manifold, allowing each rib to be processed independently while maintaining contextual relationships. This segmentation enables efficient processing without requiring excessive computational resources
2Loss of information
If the rib cage is visualized using traditional 3D volume rendering, then complete anatomical context is provided, but detection of subtle fractures becomes difficult due to limited resolution in specific regions
Solution Approach 1:
The patent creates a 2D unfolded manifold representation from the 3D volume, preserving all anatomical context while enabling high-resolution analysis of individual ribs. This dimensional transformation eliminates the need to choose between context and precision
3Measurement precision
If manual inspection of ribs from multiple perspectives is performed to ensure accurate fracture detection, then measurement precision is improved, but inspection time increases significantly
Solution Approach 1:
The patent performs preliminary unfolding and straightening of the rib cage into a standardized 2D manifold view, organizing all ribs in a consistent orientation before fracture detection. This preliminary action eliminates the need for manual multi-perspective inspection while maintaining high detection accuracy
Solution Approach 2:
The unfolded manifold view automatically presents all ribs in a standardized, easily inspectable format, allowing the detection system to self-evaluate without requiring manual reorientation or multi-perspective analysis by radiologists
4Productivity
If automated fracture detection is implemented using current methods, then productivity is improved, but reliability decreases due to imaging artifacts and false positives from adjacent ribs
Solution Approach 1:
The patent extracts each rib from the complex 3D volume and unfolds it onto a 2D manifold, isolating individual rib structures from adjacent ribs that cause artifacts. This extraction eliminates false positives while maintaining automated detection speed
Solution Approach 2:
By transforming to 2D unfolded views, the patent separates overlapping rib structures that cause artifacts in 3D rendering, improving reliability while maintaining automated processing efficiency
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Method and apparatus of automatic fracture detection from imaging scans are disclosed. A innovative manifold view combing advantages of earlier approaches while avoiding the limitations of these approaches is coupled with a deep learning based fracture detection model. Ribs and spine from a three-dimensional scan are segmented, rib centerlines and vertebra center landmarks are detected and labeled. Three-dimensional position coordinates corresponding to the received data are mapped to a defined two-dimensional manifold plane followed by interpolation and sampling techniques to generate a two- dimensional manifold slice and corresponding mapping function. A three-dimensional visualization of the rib cage is generated from a plurality of manifold slices. A trained fracture prediction model receives the plurality of manifold slices and generates a revised stack of two- dimensional manifold slices showing predicted fractures. The predicted fractures may be shown in the two-dimensional manifold slice or mapped back to the image space of the original three-dimensional scan.