Volume Data Recognition Method Based on Two-Dimensional Transfer Function Based on Watershed Algorithm

Abstract

The invention discloses a volume data identification method based on a watershed algorithm for a two-dimensional transfer function. At first, voxels of the transfer function are set through traversal belonging, the voxels are classified by means of the watershed algorithm according to connectivity of the voxels on space, interesting structures are recognized, then connected interesting structures are divided again by reutilization of the watershed algorithm according to grey levels of the structures, and set operation is adopted in wrongly-divided structures caused by over-segmentation through the watershed algorithm in volume data to obtain the whole single structure. As for different interesting structures overlaid in space of the transfer function, various structures are respectively recognized through the watershed algorithm on the premise that new dimensionality is not added, and especially when the structures are mutually connected on the space, a satisfying recognition effect can also be obtained; the wrongly-divided structures caused by over-segmentation through the watershed algorithm are collected, so the whole interesting structures are obtained; the principle is simple and beneficial to understanding and implementation.

Description

technical field [0001] The invention belongs to the field of three-dimensional visualization in computer graphics, and in particular relates to a new method for identifying structures of interest in volume data, that is, a method for identifying different structures of interest overlapping in a two-dimensional transfer function space using a watershed algorithm . Background technique [0002] With the application and continuous progress of new image acquisition technologies (such as CT, MRI, etc.) in medicine, not only the ability of people to detect diseases has been improved, but also a large amount of medical body data has been generated. The direct volume rendering (DirectVolumeRendering) method is a very practical method to understand and visualize these medical volume data. Different from the surface rendering (SurfaceRendering) method, which only displays the surface of the object by extracting geometric primitives, it directly generates the final rendering result fr...

AI Technical Summary

Problems solved by technology

[0006] Aiming at the problem that the feature information of different structures of interest in the existing two-dimensional transfer function space overlaps, especially when these structures are connected in space, they cannot be identified by setting the transfer function. The present invention proposes a method based on watershed The volume data recognition method of the two-dimensional transfer function of the algorithm uses the watershed algorithm to be sensitive to the gray level change of the image, can accurately determine the edge, and can segment the nature of the connected tissue. The structure of interest combined with the set operation can achieve better recognition results even when different tissues of interest overlap in the transfer function space and connect in the volume data space

Images