Octree-Based 3D Image Rendering for Target Marking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for 3D image rendering, such as Ray Casting, face inefficiencies and inaccuracies in removing transparent regions, particularly in multi-energy X-ray security inspection systems, requiring operators to manually navigate through extensive blank areas to mark targets, which is time-consuming and error-prone.
Innovation Solution
The method involves building an octree data structure to identify and remove transparent regions using a fixed 2D transfer function, generating a lookup table for transparency maxima, and utilizing this information to accelerate the process, allowing for rapid and accurate marking of targets in 3D images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If conventional transparent region skipping methods are used in Ray Casting, then rendering speed is improved, but accuracy is poor and the methods are limited to one-dimensional transfer functions
Solution Approach 1:
The patent divides the 3D volume data into an octree structure, segmenting the space into hierarchical regions. This allows the system to process and identify transparent regions at multiple levels of detail, improving both speed through coarse-level filtering and accuracy through fine-level verification.
Solution Approach 2:
The patent transitions from one-dimensional transfer functions to two-dimensional transfer functions, adding an additional dimension for material classification. This enables more accurate differentiation of transparent regions by considering both density and material composition simultaneously.
2Ease of operation
If operators manually navigate through blank transparent areas to mark targets, then marking capability is maintained, but time consumption increases significantly
Solution Approach 1:
The patent performs preliminary removal of transparent regions from the 3D volume data before the operator needs to mark targets. By pre-processing the data to eliminate blank areas, the system prepares an optimized view that allows operators to immediately focus on relevant regions without wasting time navigating through transparent spaces.
Solution Approach 2:
The patent extracts and removes transparent regions from the 3D image data, separating them from the meaningful content. This extraction process eliminates unnecessary blank areas that would otherwise require operator attention, leaving only the relevant opaque regions for target marking.
3Productivity
If existing transparent region skipping methods are used, then some acceleration is achieved, but the methods have low speed and poor accuracy for multi-dimensional transfer functions
Solution Approach 1:
The octree structure segments the volume data hierarchically, allowing the system to quickly identify and skip large transparent regions at upper levels while maintaining the ability to accurately process smaller regions at lower levels. This segmentation enables efficient processing that scales with data size.
Solution Approach 2:
The patent generates a lookup table of transparency maxima in advance, before the actual rendering process. This preliminary computation stores pre-calculated transparency information that can be quickly referenced during rendering, avoiding repeated calculations and improving both speed and accuracy.
Data Source
AI summary
The present disclosure provides method and apparatus for marking a target in a 3D image. The method include steps of: acquiring Computed Tomography (CT) image data of a scene; rendering a 3D image of the scene using ray casting based on the CT image data; removing a transparent region from the 3D image based on a fixed 2D transfer function; and marking the target in the 3D image.