A Contour Tree Construction Method Based on Multithreading

Abstract

The invention discloses a method for constructing a contour tree based on multithreading. Firstly, a slice image set generated by an imaging device is obtained, and a three-dimensional data set is formed in a sequence from top to bottom or bottom to top; Sub-operation, divide it into multiple sub-datasets according to the range interval; then operate each sub-dataset separately, using multi-threading technology, assign 2 threads to each sub-dataset to calculate the split tree and connection tree of the contour tree, and then merge Become a sub-contour tree, and then merge the sub-contour trees formed by each sub-dataset to form a complete contour tree, and finally simplify the topology of the formed contour tree to obtain the final required contour tree. The present invention can not only obtain clear display effect, but also can perform interaction flexibly.

Description

technical field [0001] The present invention relates to the technical field of medical image processing and application, in particular to a method for constructing contour trees based on multithreading. Background technique [0002] Volume rendering based on optical properties is a mainstream technique for medical data visualization. Volume rendering simulates the transmission process of light in volume data, and then calculates the contribution of each point on the transmission path. By assigning color and transparency attributes to the sampling points of the volume data through the transfer function, combined with the GPU-based ray-casting algorithm, a vivid 3D image can be reconstructed, such as the document "Efficient Empty Space Skipping for Large-Scale Volume Rendering" (IEEEtransactions on visualization and computer graphics, 2018, 24(1):974-983). Therefore, the quality of transfer function design determines the quality of volume rendering. [0003] Effective volum...

AI Technical Summary

Problems solved by technology

Contour tree construction algorithms based on distributed environments, such as the literature "Distributed ContourTrees" (Topological Methods in Data Analysis and Visualization III. Springer, Cham, 2014: 89-102), however, the overall process requires multiple serial calculations, reducing computing performance

[0005] The above algorithms all use parallel calculations in some stages, and the construction speed of the contour tree is slow, and because the rendering of the contour tree requires complex preprocessing, the interactive operation is not ideal

Images