CN-114998297-B - Vertebrae numbering method, device, electronic device and storage medium

CN114998297BCN 114998297 BCN114998297 BCN 114998297BCN-114998297-B

Abstract

The application relates to a vertebra numbering method, a vertebra numbering device, an electronic device and a storage medium, wherein the vertebra numbering method comprises the steps of obtaining a first detection image and a second detection image of a preset scanning area of a target object, inputting the first detection image into an image segmentation model, determining a plurality of target part areas, each target part area comprises a plurality of vertebrae, determining first position information of a target part on the first detection image according to the target part areas, determining second position information on the second detection image according to the first position information, determining the target part on the second detection image according to the second position information, and carrying out vertebra numbering on the second detection image along a preset direction according to the second position information. The application solves the problem of lower accuracy of the vertebrae number and realizes the technical effect of accurately carrying out the vertebrae number.

Inventors

MA RUNXIA
JI ZIJUN

Assignees

上海联影医疗科技股份有限公司

Dates

Publication Date: 20260505
Application Date: 20220624

Claims (7)

1. A method of numbering vertebrae, comprising: Acquiring a first detection image and a second detection image of a preset scanning area of a target object, wherein the first detection image is a CT image, and the second detection image is a PET image; Inputting the first detection image into an image segmentation model, and determining a plurality of target part areas, wherein each target part area comprises a plurality of vertebrae; Determining first position information of a target part on the first detection image according to the target part area; determining second position information on the second detection image according to the first position information, wherein the second position information is used for determining the target part on the second detection image; Performing vertebrae numbering on the second detection image along a preset direction according to the second position information; The determining, according to the target region, first position information of the target region on the first detection image, and determining, according to the first position information, second position information on the second detection image includes: adding a first position mark on the first detection image according to the target part area; registering the first detection image and the second detection image, and determining the mapping relation between the first detection image and the second detection image; Adding a second position mark on the second detection image according to the first position mark and the mapping relation, wherein the second position information comprises a second neck position mark, a second chest position mark and a second abdomen position mark; The adding of the first position mark on the first detection image according to the target position area comprises the steps of carrying out position calibration on the first detection image according to the cervical vertebra initial position or the cervical vertebra end position of the neck area, determining the first neck position mark, carrying out position calibration on the first detection image according to the thoracic vertebra initial position or the thoracic vertebra end position of the chest area, determining the first chest position mark, carrying out position calibration on the first detection image according to the lumbar vertebra initial position or the lumbar vertebra end position of the abdomen area, determining the first abdomen position mark, and taking the first neck position mark, the first chest position mark and the first abdomen position mark as the first position mark.
2. The method of claim 1, wherein the inputting the first detection image into the image segmentation model is preceded by: acquiring a plurality of groups of training images with target position marks, and forming a training set from the training images; and training a deep learning model based on the training set to obtain the image segmentation model.
3. The method of claim 1, wherein the numbering the second detected image along a predetermined direction according to the second position information comprises: determining the number of vertebrae of the target site in the second detection image according to the second position information; clustering the target parts in the second detection image according to the vertebrae number to determine vertebrae contours; Numbering the vertebrae contours along the preset direction according to the second position information.
4. The method of claim 1, wherein the performing the vertebra numbering on the second detection image along the preset direction according to the second position information comprises: Numbering a first preset number of vertebrae in the second detection image along a preset direction according to the second neck position mark; Numbering a second preset number of vertebrae in the second detection image along a preset direction according to the second chest position identification; and numbering a third preset number of vertebrae in the second detection image along a preset direction according to the second abdomen position mark.
5. A vertebral numbering device, comprising: the acquisition module is used for acquiring a first detection image and a second detection image of a preset scanning area of a target object, wherein the first detection image is a CT image, and the second detection image is a PET image; the segmentation module is used for inputting the first detection image into an image segmentation model to obtain a plurality of target part areas, wherein each target part area comprises a plurality of vertebrae; the first calculation module is used for determining first position information of the target part on the first detection image according to the target part area; The second calculation module is used for determining second position information on the second detection image according to the first position information, and the second position information is used for determining the target part on the second detection image; the numbering module is used for numbering vertebrae of the second detection image along a preset direction according to the second position information; The determining, according to the target region, first position information of the target region on the first detection image, and determining, according to the first position information, second position information on the second detection image includes: adding a first position mark on the first detection image according to the target part area; registering the first detection image and the second detection image, and determining the mapping relation between the first detection image and the second detection image; Adding a second position mark on the second detection image according to the first position mark and the mapping relation, wherein the second position information comprises a second neck position mark, a second chest position mark and a second abdomen position mark; The adding of the first position mark on the first detection image according to the target position area comprises the steps of carrying out position calibration on the first detection image according to the cervical vertebra initial position or the cervical vertebra end position of the neck area, determining the first neck position mark, carrying out position calibration on the first detection image according to the thoracic vertebra initial position or the thoracic vertebra end position of the chest area, determining the first chest position mark, carrying out position calibration on the first detection image according to the lumbar vertebra initial position or the lumbar vertebra end position of the abdomen area, determining the first abdomen position mark, and taking the first neck position mark, the first chest position mark and the first abdomen position mark as the first position mark.
6. An electronic device comprising a memory and a processor, wherein the memory has stored therein a computer program, the processor being arranged to run the computer program to perform the method of any of claims 1 to 4.
7. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, carries out the steps of the vertebra numbering method according to any one of claims 1 to 4.

Description

Vertebrae numbering method, device, electronic device and storage medium Technical Field The present application relates to the field of medical imaging technology, and in particular, to a method, apparatus, electronic device and storage medium for numbering vertebrae. Background Bone metastasis refers to the symptoms that when a primary tumor exists, the tumor migrates to bone tissue through blood, lymph or local migration to cause bone tissue destruction, and pain often occurs in patients. In medical diagnosis, imaging examination procedures such as Positron Emission Tomography (PET) and Computed Tomography (CT) are commonly used for diagnosis. In connection with diagnosis of scanned images, a physician needs to determine that the transition site is present at a vertebral location of the patient's spine. The existing vertebra number requires a doctor to manually count and number, and the accuracy and efficiency of the vertebra number are low. With the development of machine vision technology, a part of medical imaging instruments perform image processing through scanned images of a patient, identify each vertebra of the patient and number the vertebrae. However, because the human spine is long and there are more vertebrae throughout the entire complete spine, there are situations where the number of vertebrae is inaccurate when the imaging accuracy is poor or the machine identification accuracy is low. Aiming at the problem of low accuracy of the vertebrae number in the related art, no effective solution is proposed at present. Disclosure of Invention In this embodiment, a method, an apparatus, an electronic apparatus, and a storage medium for numbering vertebrae are provided to solve the problem of low accuracy of numbering vertebrae in the related art. In a first aspect, in this embodiment, there is provided a vertebral numbering method including acquiring a first detection image and a second detection image of a target object, inputting the first detection image into an image segmentation model, determining a plurality of target site areas, each including a plurality of vertebrae, determining first position information of a target site on the first detection image according to the target site areas, determining second position information on the second detection image according to the first position information, the second position information being used for determining the target site on the second detection image, and numbering the second detection image along a preset direction according to the second position information. In one embodiment, before the first detection image is input into the image segmentation model, acquiring a plurality of groups of training images with target part labels, forming a training set by the training images, and training a deep learning model based on the training set to obtain the image segmentation model. In one embodiment, the step of numbering the second detection image along the preset direction according to the second position information comprises the steps of determining the number of vertebrae of the target part in the second detection image according to the second position information, carrying out clustering processing on the target part in the second detection image according to the number of vertebrae to determine a vertebrae outline, and numbering the vertebrae outline along the preset direction according to the second position information. In one embodiment, the determining the first position information of the target part on the first detection image according to the target part area and the second position information on the second detection image according to the first position information include adding a first position identifier on the first detection image according to the target part area, registering the first detection image with the second detection image, determining a mapping relation between the first detection image and the second detection image, and adding a second position identifier on the second detection image according to the first position identifier and the mapping relation. In one embodiment, the target portion area comprises a neck area, a chest area and an abdomen area, the adding of the first position mark on the first detection image according to the target portion area comprises the steps of performing position calibration on the first detection image according to the cervical vertebra starting position or the cervical vertebra end position of the neck area, determining the first neck position mark, performing position calibration on the first detection image according to the thoracic vertebra starting position or the thoracic vertebra end position of the chest area, determining the first chest position mark, performing position calibration on the first detection image according to the lumbar vertebra starting position or the lumbar vertebra end position of the abdomen area, determining the first abdomen position mark, and ta