Method for automatically extracting bed plate

Abstract

The invention discloses a method for automatically extracting a bed plate. The method for automatically extracting a bed plate comprises the steps of (a) providing reconstructed CT data, carrying out sampling to obtain a CT image, and carrying out connected domain calculation on a CT value in the CT image to obtain at least one connected domain, (b) extracting a connected domain to be processed, calculating the pixel number of the connected domain, and carrying out bed plate feature detection of the connected domain in a pixel number range, (c) repeating the step (b) until the bed plate feature detection is completed, (d) carrying out three-dimensional region growing on the connected domain after the bed plate feature is detected, and extracting the bed plate. The method for automatically extracting the bed plate provided by the invention has the advantages of short operation time and high accuracy, various shapes of the bed plates can be extracted, and the method is suitable for all kinds of CT image data.

Description

technical field [0001] The invention relates to a medical image processing method, in particular to a method for automatically extracting a bed board. Background technique [0002] In CT imaging technology, the bed board often exists in the reconstructed image, and the density of the bed board is similar to that of human body parts, such as human tissues such as blood vessels and bones injected with contrast agent, but the bed board not only does not carry usable information, but also processes data It will increase the amount of calculation, and even in some applications such as 3D volume rendering and maximum intensity projection (MIP), it will block effective anatomical information and affect the accuracy of imaging results. Therefore, it is necessary to automatically extract the bed board from the CT image. [0003] In the prior art, the bed board extraction method is mainly realized through various algorithms. "A Fully Automatic Bed / Linen Segmentation for Fused PET / CT...

AI Technical Summary

Problems solved by technology

[0004]body PET-CT Studies” (the 5th International Conference on Information Technology and Application in Biomedicine, Shenzhen, China, May 30-31, 2008) proposed an extraction The method of human body outline, this method firstly sets the standard uptake value (SUV, standard uptake value) through the PET image to obtain the outline of the human body, and then obtains the human body through the CT image, and indirectly removes the bed board of the CT image, but this method is only applicable to PET -CT image

"Automated Segmentation of Soft tissue in abdominal CT scans" proposes a median filtering method, which uses the median filtering method to remove linear structures, but the bed board does not only have linear structures, such as Figure 1The bed board in the schematic diagram of the leg CT image is a non-linear structure, and the median filter cannot be used bed board

In addition, in the prior art, the bed board is removed during CT data reconstruction. This method cannot guarantee that each set of CT data can effectively remove the bed board. For example, some bed boards include two parts, the headrest and the bed body. The bed board is not obvious or the bed board is broken

There is also the use of points above a certain threshold as seed points for region growth to extract human body images and indirectly remove the bed board, but this method is time-consuming

[0005] In order to solve the above problems, the present invention proposes a method for automatically extracting bed boards, which is applicable to various CT images, and has short running time and high accuracy , suitable for bed boards of various shapes, and will not cause mis-lifting or missing-lifting

Images