Semi-spherical and Cylindrical Projection Model for Virtual Endoscopy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for visualizing hollow organs through virtual endoscopy, such as CT colonoscopy, face limitations including visual distortion, increased reviewing time, and reduced exposure to lesions due to the transition from perspective 3D projections to 2D projections, which can lead to inefficient clinical readings and missed information, especially in areas with compact folds.
Innovation Solution
A method utilizing a projection model with a semispherical projection for far views and a cylindrical projection for near views, characterized by similar transition vectors, to render areas blinded by traditional perspective rendering, minimizing distortion and maintaining linear radial perception, thereby increasing the inspected colonic mucosa without altering original information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a transition from perspective 3D projections to 2D projections is made to increase the amount of visualized colonic mucosa, then the quantity of visualized area is improved, but visual distortion increases and requires observers to re-learn interpretation
Solution Approach 1:
The patent transitions from traditional 2D projections to a 3D virtual environment that maintains spatial relationships while expanding the visualized area. The virtual camera moves through a reconstructed 3D model of the colon, allowing observers to view mucosal surfaces from multiple perspectives without the severe distortion inherent in flat 2D projections.
Solution Approach 2:
The system employs a dynamic virtual camera that can move freely through the 3D reconstructed colon model, allowing interactive exploration of different regions. This dynamic viewing capability enables comprehensive visualization of colonic mucosa while maintaining accurate spatial representation, avoiding the static distortion problems of traditional 2D projections.
2Area of stationary object
If cube views are used to represent a complete 360 environment, then the visual exposure to lesions is improved, but objects can be partially present in multiple edges creating redundancy and confusion
Solution Approach 1:
The patent replaces the static cube view structure with a dynamic virtual camera that can continuously move through the 3D space. This allows the observer to view the entire 360-degree environment without the artificial segmentation into cube faces, eliminating object duplication while maintaining complete visual exposure.
Solution Approach 2:
The patent segments the viewing experience into manageable portions through controlled camera movements and fly-through sequences, allowing systematic examination of different colon regions without the confusing overlap inherent in cube view representations.
3Device complexity
If cube view distortion is applied to assemble views around the Front-view, then information overflow is reduced, but non-homogeneous distortion creates additional artifacts depending on object position
Solution Approach 1:
The patent moves the assembly process from 2D cube face distortion to 3D spatial arrangement, where multiple views are integrated into a coherent three-dimensional virtual environment. This eliminates position-dependent artifacts by maintaining consistent spatial relationships across all viewing angles.
4Ease of operation
If mapping processes are used to fit data into the final assembled space, then the visualization layout is improved, but original information is deleted, interpolated or altered increasing processing time
Solution Approach 1:
The patent performs preliminary 3D reconstruction of the colon model from medical images before the viewing process. This pre-processing creates an accurate three-dimensional representation that can be directly viewed and explored, eliminating the need for complex real-time mapping and interpolation operations during visualization.
Solution Approach 2:
The system creates a virtual copy of the colon's three-dimensional structure from medical imaging data. This virtual model can be explored interactively without repeatedly processing or mapping the original medical images, significantly reducing processing time while preserving all original information.
Data Source
AI summary
A method and system for performing rendering at a given viewpoint of at least one portion of visual areas blinded to a perspective rendering at the given viewpoint are disclosed, the method comprising providing a projection model comprising a first part used for rendering far view and a second part for rendering near view such that at least one portion of the visual areas is rendered using the second part of the projection model and further wherein a transition between the first part of the projection model and the second part of the projection model is characterized by projection vectors substantially similar, obtaining image data, performing a first projection of at least one part of the image data according to the first part of the projection model to generate first projected data, performing a second projection of another part of the image data according to the second part of the projection model to generate second projected data and displaying projected data using the first projected data and the second projected data.