Medical Image Rendering via Shifted Voxel Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical image generating methods using ray casting struggle to clearly display invasive abnormal regions within the body's tissue, as they either obscure surface features or require complex and time-consuming adjustments in opacity values to visualize both surface and subsurface abnormalities simultaneously.
Innovation Solution
A medical image generating apparatus and method that employs a shading section with a shift position specifying section to obtain gradation information from a position shifted from the virtual light ray's arrival point, allowing for the generation of three-dimensional images that distinctly display invasive abnormal regions by adjusting shading and color information based on voxel values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Shape
If ray casting with gradient shading is used to clearly display surface features, then surface shape rendering is improved, but invasive abnormal regions within tissue cannot be displayed
Solution Approach 1:
The patent introduces a new dimension by sampling voxel values at multiple depths along the light ray path, not just at the surface arrival point. This allows the system to capture subsurface information (invasive abnormal regions) while maintaining surface rendering quality, effectively adding a depth dimension to the rendering process.
Solution Approach 2:
The patent segments the light ray path into multiple sampling points at different depths. By dividing the volumetric data along the ray path into surface level and subsurface levels, the system can independently process and render surface features and invasive abnormalities separately, then combine them for comprehensive visualization.
2Loss of information
If opacity value is reduced to display invasive abnormal regions, then subsurface abnormalities are clearly displayed, but surface portion becomes unclear and shape is not clearly displayed
Solution Approach 1:
The patent applies different rendering strategies to different spatial locations along the light ray path. Surface voxels receive gradient-based shading for clear shape rendering, while subsurface voxels at sampling points are evaluated for invasive abnormality detection with appropriate opacity handling. This local differentiation resolves the contradiction between surface clarity and subsurface visibility.
3Loss of information
If multiple images with different opacity values are generated to display both surface and subsurface abnormalities, then comprehensive diagnostic information is obtained, but complicated working and much time are required, preventing rapid diagnosis
Solution Approach 1:
The patent merges surface rendering and subsurface abnormality detection into a single integrated rendering process. By combining gradient shading for surface features with multi-depth voxel sampling for invasive abnormalities in one pass, the system eliminates the need to generate and compare multiple separate images, thereby maintaining comprehensive diagnostic information while significantly improving diagnosis speed.
Solution Approach 2:
The patent performs preliminary sampling of voxel values at multiple predetermined depths along the light ray path before final image generation. This preliminary action captures both surface and subsurface information in advance, allowing the rendering process to efficiently combine this pre-sampled data without requiring multiple separate rendering passes or extensive post-processing adjustments.
Data Source
AI summary
An image processing apparatus acquires voxel data obtained by imaging an interior of a living body by a modality and generates a three-dimensional medical image by volume rendering using a ray casting method. The image processing apparatus generates the three-dimensional medical image wherein shade of a surface of an inner wall of an intestinal tract is clearly displayed and an abnormal region invasively developed in the interior of the inner wall is distinguishably displayed based on color information corresponding to voxel data placed at a position shifted from the surface by a predetermined distance.