An image processing method for separating the head and neck regions

Abstract

The present invention provides an image processing method for separating the head and neck regions, which is characterized by including the following specific steps: S1, segmentation of human tissues and air; S2, positioning of the reference cross-section; S3, based on the reference cross-section, positioning of the head and neck separation layer. The present invention finds the reference cross-section and the head and neck separation layer based on the judgment of image features, and can process both CTA images and MRA images; there is no requirement for the scanning range of the head and neck, and it can also process images that only include part of the head and part of the neck; moreover, the present invention can operate automatically, improving work efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and specifically relates to an image processing method for separating the head and neck regions. Background Art

[0002] The carotid arteries of the head and neck are the main vascular pathways for transporting blood from the heart to the brain. Clinically, computed tomography angiography (CTA) and magnetic resonance angiography (MRA) are usually used to image the carotid arteries of the head and neck, and morphological examinations of the blood vessels are carried out to observe problems such as arteriosclerosis of the head and neck, stenosis of the carotid arteries of the head and neck, plaques of the carotid arteries of the head and neck, and blood rheology. In order to better analyze the characteristics such as the morphology and course of blood vessels, it is usually necessary to separate the carotid arteries of the head and neck in the CTA or MRA head and neck artery images from other human tissues. However, due to the differences in anatomical structure and tissue composition between the head and neck, directly performing vascular segmentation based on the head and neck artery images results in poor segmentation effects. Therefore, during the process of head and neck vascular segmentation, it is often necessary to separate the head and neck and perform separate vascular extraction respectively.

[0003] The existing methods for separating the head and neck regions are mainly divided into two types. One is manual selection and separation, that is, during the process of extracting the head and neck blood vessels by a doctor, the doctor manually selects the separation layer between the head and neck, and then continues to extract the head and neck blood vessels; the other is to generate a sagittal image of the head and neck region using a head and neck CTA image, separate the human tissues from the background in the sagittal image, and then calculate the distance from the front surface to the back surface of the human tissues. Since the distance between the front and back surfaces of the head and neck is quite different, the layer with the smallest front-back distance is used as the separation layer. However, it cannot process MRA, and requires the image to completely contain the head and neck. It is difficult to separate images that only contain part of the head or part of the neck region. Summary of the Invention

[0004] The purpose of the present invention is to propose an image processing method for separating the head and neck regions to solve the problems raised in the above background art.

[0005] To achieve the above purpose, the present invention provides the following technical solutions:

[0006] An image processing method for separating the head and neck regions includes the following specific steps:

[0007] S1. Separation of human tissues and air;

[0008] S2. Positioning the reference cross-section;

[0009] S3. Based on the reference cross-section, positioning the head and neck separation layer.

[0010] The specific content of the step S1 includes:

[0011] Calculate the grayscale statistical histogram hist of the image. Use the grayscale value corresponding to the a-th percentile of the grayscale statistical histogram hist of the image as the grayscale threshold Ta between air and human tissues, where the optimal value of a is 2. Use the grayscale threshold Ta between air and human tissues as the threshold to separate human tissues from air.

[0012] Step S2 is divided into four schemes, and the specific contents are as follows:

[0013] Scheme 1: Calculate the area characteristics of air and tissues in the body layer by layer using cross-sections. This characteristic is: the ratio of the area of air to tissues in the body, or the ratio of the area of air in the body to the total area of air and tissues in the body, or the ratio of the area of tissues in the body to the total area of air and tissues in the body. Select the layer with the area characteristic of air and tissues in the body closest to the reference value as the reference cross-section.

[0014] Specifically, by calculating the internal area of the human body occupied by air and tissues, obtain the area characteristics Ra of air and tissues in the body for each layer. This characteristic is: the ratio of the area of air to tissues in the body, or the ratio of the area of air in the body to the total area of air and tissues in the body, or the ratio of the area of tissues in the body to the total area of air and tissues in the body. Select the layer with the area characteristic Ra of air and tissues in the body closest to the reference value Pa of the area characteristic of air and tissues in the body as the reference cross-section Sr. Among them, when the area characteristic Ra of air and tissues in the body is selected as the ratio of the area of air to tissues in the body, the optimal value of the reference value Pa of the area characteristic of air and tissues in the body is 1:9. When it is selected as the ratio of the area of air in the body to the total area of air and tissues in the body, the optimal value of the reference value Pa of the area characteristic of air and tissues in the body is 1:10. When it is selected as the ratio of the area of tissues in the body to the total area of air and tissues in the body, the optimal value of the reference value Pa of the area characteristic of air and tissues in the body is 9:10.

[0015] .

[0016] Scheme 2: Obtain the average value of the area characteristics of air and tissues in the reference cross-section of the body through a reference template. This characteristic is: the ratio of the area of air to tissues in the body, or the ratio of the area of air in the body to the total area of air and tissues in the body, or the ratio of the area of tissues in the body to the total area of air and tissues in the body. Obtain the area characteristics of air and tissues in the reference cross-section of each template, and then calculate their average value to obtain the average value of the area characteristics of air and tissues in the reference cross-section of the template. Then calculate the area characteristics of air and tissues in the body layer by layer, and select the layer with the area characteristic of air and tissues in the body closest to the average value of the area characteristics of air and tissues in the reference cross-section of the body as the reference cross-section.

[0017] Specifically, select one or more standard head and neck images as reference templates, manually select the reference cross-sections of each template, calculate the human tissues obtained by the method in S1, perform morphological hole filling processing, and the filled part is the in-vivo air; by calculating the areas of the in-vivo air and tissues in the human body, and selecting the same calculation method as the area feature Ra of the in-vivo air and tissues, obtain the area features of the in-vivo air and tissues of the reference cross-sections of each template, and then calculate their average value to obtain the average value Ma of the area features of the in-vivo air and tissues of the reference cross-section of the template. Then, calculate the area features Ra of the in-vivo air and tissues of the data to be measured layer by layer, and take the layer with the area feature Ra of the in-vivo air and tissues closest to the average value Ma of the area features of the in-vivo air and tissues of the reference cross-section as the reference cross-section Sr;

[0018] 。

[0019] Scheme 3: Obtain the candidate layer range by calculating the distance ratio between the first layer and the last layer close to the cranial vertex side, calculate the area ratio of the in-vivo air and tissues layer by layer using the cross-section, and take the layer with the largest area ratio of the in-vivo air and tissues in the candidate layer range as the reference cross-section;

[0020] Specifically, determine the candidate range of the layer where the reference cross-section is located. This range is from the candidate starting layer S0 to the candidate ending layer S1. Take the layer with the distance ratio to the first layer and the last layer close to the cranial vertex side closest to the starting layer ratio b as the candidate starting layer S0. The total number of image layers is K, where the optimal value of the starting layer ratio b is 0:(K - 1); take the layer with the distance ratio to the first layer and the last layer close to the cranial vertex side closest to the ending layer ratio c as the candidate ending layer S1, where the optimal value of the ending layer ratio c is 7:3; then calculate the area ratio of the in-vivo air and human tissues within the range from the candidate starting layer S0 to the candidate ending layer S1 to obtain the in-vivo air tissue ratio Rb of each layer, and select the layer where the in-vivo air tissue ratio Rb is the largest as the reference cross-section Sr.

[0021] Scheme 4: Extract image features layer by layer and match them with the predefined reference cross-section features, and set the most matching image layer as the reference cross-section;

[0022] Specifically, select one or more standard head and neck images as reference templates, manually select the reference cross-sections of each template, and then extract the reference cross-section features F of each template. The feature extraction method is: neural network feature extraction or radiomics feature extraction. Use the reference cross-section features F and the data of each template and each reference cross-section to train to obtain the classifier C. The training method is: neural network or support vector machine. Finally, calculate the classification probability P of the cross-section in the classifier C layer by layer, and select the layer where the classification probability P is the largest as the reference cross-section Sr.

[0023] The specific content of step S3 includes: calculating the position of the head and neck separation layer based on the pre-defined relationship between the reference cross-section and the distance of the head and neck separation layer.

[0024] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0025] 1. The present invention finds the reference cross-section and the head and neck separation layer based on the judgment of image features, and can process both CTA images and MRA images;

[0026] 2. The present invention has no requirements for the scanning range of the head and neck, and can also process images that only contain part of the head and part of the neck;

[0027] 3. The present invention automatically processes, improving efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is the gray-scale statistical histogram hist of the present invention;

[0029] Figure 2 It is a schematic diagram of the relationship between the head and neck separation layer Ss, the reference cross-section Sr and the plane position difference D of the present invention;

[0030] Figure 3 It is a schematic diagram of the standard head and neck image reference template of the present invention;

[0031] Figure 4 It is a schematic diagram of the CTA standard head and neck image reference template of the present invention;

[0032] Figure 5 It is a schematic diagram of the MRA standard head and neck image reference template of the present invention;

[0033] Figure 6 It is a schematic diagram of the relationship between the candidate starting layer S0, the candidate ending layer S1, the first layer near the cranial vertex side, the last layer near the cranial vertex side, the starting layer ratio b and the ending layer ratio c of the present invention;

[0034] Figure 7 It is a schematic diagram of the reference cross-section feature F of the present invention and the training and calculation process of obtaining the classifier C;

[0035] Figure 8 It is the overall flow chart of the method of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0036] To clarify the technical problems, technical solutions, implementation processes, and performance demonstrations, the present invention will be further described in detail below in conjunction with embodiments. It should be understood that the specific embodiments described herein are merely used to explain the present invention and are not used to limit the present invention. Various exemplary embodiments, features, and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numerals in the drawings denote elements having the same or similar functions. Although various aspects of the embodiments are shown in the drawings, the drawings do not have to be drawn to scale unless otherwise specified.

[0037] The specifically used term "exemplary" herein means "serving as an example, embodiment, or illustration". Any embodiment described herein as "exemplary" does not have to be construed as being superior to or better than other embodiments.

[0038] In addition, to better illustrate the present disclosure, numerous specific details are given in the following specific implementation manners. Those skilled in the art should understand that the present disclosure can also be implemented without some specific details. In some instances, methods, means, elements, and circuits well-known to those skilled in the art are not described in detail in order to highlight the gist of the present disclosure.

[0039] Embodiment 1

[0040] As Figure 1 and Figure 2 shown, an image processing method for separating the head and neck regions specifically includes the following steps:

[0041] S1. Segmentation of human tissues and air; calculate the gray-scale statistical histogram hist of the image, count the number of pixels of each gray level in the image, and obtain the frequency of occurrence of each gray level in the image. The abscissa is the gray value, and the ordinate is the frequency of occurrence of the gray value. Take the gray value corresponding to the a-th percentile of the gray-scale statistical histogram hist of the image as the gray-scale threshold Ta between air and human tissues, where the optimal value of a is 2. Using the gray-scale threshold Ta between air and human tissues as the threshold, separate human tissues and air. Pixels greater than or equal to the gray-scale threshold Ta are classified as human tissues, and pixels lower than the gray-scale threshold Ta are classified as air;

[0042] S2. Locate the reference cross-section; perform morphological hole filling on the human tissues obtained in step S1, and the filled part is the in-vivo air. By calculating the areas of in-vivo air and tissues occupying the internal area of the human body, obtain the area characteristics Ra of in-vivo air and tissues for each layer. This characteristic is selected as the ratio of in-vivo air to tissue area, and the corresponding reference value Pa of the area characteristic of in-vivo air to tissue is 1:9. Select the layer where the area characteristic Ra of in-vivo air and tissues is closest to the reference value Pa of the area characteristic of in-vivo air to tissue as the reference cross-section Sr;

[0043]

[0044] S3. Based on the reference cross-section, locate the head-neck separation layer; the plane position difference between the head-neck separation layer Ss and the reference cross-section Sr is D. In this embodiment, the optimal value of D is 40 mm; with the direction from the top of the head to the neck as the positive direction, the head-neck separation layer Ss is located at a distance D from the reference cross-section Sr along the positive direction. When the distance exceeds the image range, the head-neck separation layer Ss is selected as the layer with the maximum distance in the positive direction.

[0045] Example 2

[0046] As Figure 1 and Figure 2 shown, an image processing method for separating the head and neck regions specifically includes the following steps:

[0047] S1. Separate human tissue from air; calculate the image gray-scale statistical histogram hist, count the number of pixels of each gray level in the image, and obtain the frequency of occurrence of each gray level in the image. The abscissa is the gray value, and the ordinate is the frequency of occurrence of the gray value. Take the gray value corresponding to the a-th percentile of the image gray-scale statistical histogram hist as the gray threshold Ta between air and human tissue, where the optimal value of a is 2. Using the gray threshold Ta between air and human tissue as the threshold, separate human tissue from air. Pixels greater than or equal to the gray threshold Ta are classified as human tissue, and pixels lower than the gray threshold Ta are classified as air.

[0048] S2. Locate the reference cross-section; perform morphological hole filling on the human tissue obtained in step S1. The filled part is the internal air in the body. By calculating the areas of the internal air and tissue in each layer of the human body, obtain the area characteristics Ra of the internal air and tissue in each layer. This characteristic is selected as the area ratio of the internal air to the total area of the internal air and tissue, and the corresponding reference value Pa of the area characteristics of the internal air and tissue is 1:10. Select the layer with the area characteristic Ra of the internal air and tissue closest to the reference value Pa of the area characteristics of the internal air and tissue as the reference cross-section Sr;

[0049]

[0050] S3. Based on the reference cross-section, locate the head-neck separation layer; the plane position difference between the head-neck separation layer Ss and the reference cross-section Sr is D. In this embodiment, the optimal value of D is 40 mm; with the direction from the top of the head to the neck as the positive direction, the head-neck separation layer Ss is located at a distance D from the reference cross-section Sr along the positive direction. When the distance exceeds the image range, the head-neck separation layer Ss is selected as the layer with the maximum distance in the positive direction.

[0051] Example 3

[0052] As Figure 1 and Figure 2As shown in the figure, an image processing method for separating the head and neck regions specifically includes the following steps:

[0053] S1. Segmentation of human tissue and air: Calculate the gray-scale statistical histogram hist of the image, count the number of pixels of each gray level in the image, and obtain the frequency of occurrence of each gray level in the image. The abscissa is the gray value, and the ordinate is the frequency of occurrence of the gray value. Use the gray value corresponding to the a-th percentile of the gray-scale statistical histogram hist of the image as the gray threshold Ta between air and human tissue, where the optimal value of a is 2. Use the gray threshold Ta between air and human tissue as the threshold to separate human tissue and air. Pixels greater than or equal to the gray threshold Ta are classified as human tissue, and pixels lower than the gray threshold Ta are classified as air.

[0054] S2. Locate the reference cross-section: Perform morphological hole filling on the human tissue obtained in step S1. The filled part is the internal air in the body. Calculate the area characteristics Ra of the internal air and tissue in each layer by calculating the internal area occupied by the internal air and tissue in the body. This characteristic is selected as the area ratio of the internal tissue to the total area of the internal air and tissue. The corresponding reference value Pa of the area characteristic of the internal air and tissue is 9:10. Select the layer with the area characteristic Ra of the internal air and tissue closest to the reference value Pa of the area characteristic of the internal air and tissue as the reference cross-section Sr.

[0055]

[0056] S3. Based on the reference cross-section, locate the head and neck separation layer: The plane position difference between the head and neck separation layer Ss and the reference cross-section Sr is D. In this embodiment, the optimal value of D is 40 mm. Taking the direction from the top of the head to the neck as the positive direction, the head and neck separation layer Ss is located at a distance D from the reference cross-section Sr along the positive direction. When the distance exceeds the image range, the head and neck separation layer Ss is selected as the layer with the maximum distance in the positive direction.

[0057] Example 4

[0058] As Figure 1 and Figures 3 - 5 shown in the figure, an image processing method for separating the head and neck regions specifically includes the following steps:

[0059] S1. Divide the human tissue from the air; calculate the grayscale statistical histogram hist of the image, count the number of pixels of each grayscale level in the image, and obtain the frequency of each grayscale appearance in the image. The abscissa is the grayscale value, and the ordinate is the frequency of the grayscale value appearance. Use the grayscale value corresponding to the a-th percentile of the grayscale statistical histogram hist of the image as the grayscale threshold Ta between the air and the human tissue, where the optimal value of a is 2. Use the grayscale threshold Ta between the air and the human tissue as the threshold to separate the human tissue from the air. Pixels greater than or equal to the grayscale threshold Ta are classified as human tissue, and pixels lower than the grayscale threshold Ta are classified as air.

[0060] S2. Locate the reference cross-section; select one or more standard head and neck images as reference templates, manually select the reference cross-sections of each template, perform morphological hole filling on the human tissue obtained by using the method in S1, and the filled part is the internal air in the body; by calculating the internal area of the body occupied by the internal air and tissue, and selecting the same calculation method as the area feature Ra of the internal air and tissue in the body, obtain the area features of the internal air and tissue of the reference cross-sections of each template, and then calculate their average value to obtain the average value Ma of the area features of the internal air and tissue of the reference cross-section of the template. Then, layer by layer, calculate the area features Ra of the internal air and tissue in the data to be measured. The layer with the area feature Ra of the internal air and tissue closest to the average value Ma of the area features of the internal air and tissue of the reference cross-section is used as the reference cross-section Sr.

[0061]

[0062] S3. Based on the reference cross-section, locate the head and neck separation layer; the planar position difference between the head and neck separation layer Ss and the reference cross-section Sr is D. In this embodiment, the optimal value of D is 40 mm; with the direction from the top of the head to the neck as the positive direction, the head and neck separation layer Ss is located at a distance of D from the reference cross-section Sr along the positive direction. When the distance exceeds the image range, the head and neck separation layer Ss is selected as the layer with the maximum distance in the positive direction.

[0063] Example 5

[0064] As Figure 1 and Figure 6 shown, an image processing method for separating the head and neck parts specifically includes the following steps:

[0065] S1. Segmentation of human tissue and air; calculate the gray-scale statistical histogram hist of the image, count the number of pixels of each gray level in the image, and obtain the frequency of each gray level appearing in the image. The abscissa is the gray value, and the ordinate is the frequency of the gray value appearing. Use the gray value corresponding to the a-th percentile of the gray-scale statistical histogram hist of the image as the gray-scale threshold Ta between air and human tissue, where the optimal value of a is 2. Using the gray-scale threshold Ta between air and human tissue as the threshold, separate human tissue and air. Pixels greater than or equal to the gray-scale threshold Ta are classified as human tissue, and pixels lower than the gray-scale threshold Ta are classified as air.

[0066] S2. Locate the reference cross-section; determine the candidate range of the layer where the reference cross-section is located. This range is from the candidate starting layer S0 to the candidate ending layer S1. Use the layer with the distance ratio to the first layer and the last layer near the cranial apex side closest to the starting layer ratio b as the candidate starting layer S0. The total number of image layers is K, where the optimal value of the starting layer ratio b is 0:(K - 1); use the layer with the distance ratio to the first layer and the last layer near the cranial apex side closest to the ending layer ratio c as the candidate ending layer S1, where the optimal value of the ending layer ratio c is 7:3. Then, by calculating the ratio of the area of in-vivo air to human tissue within the range from the candidate starting layer S0 to the candidate ending layer S1, obtain the in-vivo air-tissue ratio Rb of each layer, and select the layer with the maximum in-vivo air-tissue ratio Rb as the reference cross-section Sr.

[0067] S3. Based on the reference cross-section, locate the head-neck separation layer; the plane position difference between the head-neck separation layer Ss and the reference cross-section Sr is D. In this embodiment, the optimal value of D is 40 mm; with the direction from the top of the head to the neck as the positive direction, the head-neck separation layer Ss is located at a distance D from the reference cross-section Sr along the positive direction. When the distance exceeds the image range, the head-neck separation layer Ss is selected as the layer with the maximum distance in the positive direction.

[0068] Example 6

[0069] As Figure 1 and Figure 7 shown, an image processing method for separating the head and neck regions specifically includes the following steps:

[0070] S1. Segmentation of human tissue and air; calculate the gray-scale statistical histogram hist of the image, count the number of pixels of each gray level in the image, and obtain the frequency of each gray level appearing in the image. The abscissa is the gray value, and the ordinate is the frequency of the gray value appearing. Use the gray value corresponding to the a-th percentile of the gray-scale statistical histogram hist of the image as the gray-scale threshold Ta between air and human tissue, where the optimal value of a is 2. Using the gray-scale threshold Ta between air and human tissue as the threshold, separate human tissue and air. Pixels greater than or equal to the gray-scale threshold Ta are classified as human tissue, and pixels lower than the gray-scale threshold Ta are classified as air.

[0071] S2. Positioning reference cross-section: Select one or more standard head and neck images as reference templates, manually select the reference cross-sections of each template, and then extract the reference cross-section features F of each template. The feature extraction method is: neural network feature extraction or radiomics feature extraction. Using the reference cross-section features F and the data of each template as input data, and using whether each layer of the template is a reference cross-section as a label to train to obtain a classifier C. The training method is: neural network or support vector machine. Finally, by inputting the data to be measured for each layer into the classifier C, calculate the classification probability P of the cross-section in the classifier C layer by layer, and select the layer where the maximum value of the classification probability P is located as the reference cross-section Sr.

[0072] S3. Based on the reference cross-section, position the head and neck separation layer: The plane position difference between the head and neck separation layer Ss and the reference cross-section Sr is D. In this embodiment, the optimal value of D is 40 mm. Taking the direction from the top of the head to the neck as the positive direction, the head and neck separation layer Ss is located at a distance of D from the reference cross-section Sr along the positive direction. When the distance exceeds the image range, the head and neck separation layer Ss is selected as the layer with the largest distance in the positive direction.

[0073] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. An image processing method for separating the head and neck regions, characterized in that, It includes the following specific steps: S1. Divide human tissues from air; Calculate the gray - level statistical histogram hist of the image. Take the gray - level value corresponding to the a - percentile of the gray - level statistical histogram hist of the image as the gray - level threshold Ta between air and human tissues. Using the gray - level threshold Ta between air and human tissues as the threshold, separate human tissues from air; S2. Locate the reference cross - section; S3. Based on the reference cross - section, locate the head - neck separation layer; The first method that can be adopted for the step S2 to locate the reference cross - section is: Use cross - sections to calculate the area characteristics of air and tissues in the body layer by layer. This characteristic is: the ratio of the area of air to tissues in the body, or the ratio of the area of air to the total area of air and tissues in the body, or the ratio of the area of tissues to the total area of air and tissues in the body. Take the layer with the area characteristic of air and tissues in the body closest to the reference value as the reference cross - section; The second method that can be adopted for the step S2 to locate the reference cross - section is: Obtain the mean value of the area characteristics of air and tissues in the reference cross - section of the body through a reference template. This characteristic is: the ratio of the area of air to tissues in the body, or the ratio of the area of air to the total area of air and tissues in the body, or the ratio of the area of tissues to the total area of air and tissues in the body. Then calculate the area characteristics of air and tissues in the body layer by layer. Take the layer with the area characteristic of air and tissues in the body closest to the mean value of the area characteristics of air and tissues in the reference cross - section of the body as the reference cross - section; The third method that can be adopted for the step S2 to locate the reference cross - section is: Obtain the candidate layer range by calculating the distance ratio between the first layer and the last layer near the cranial vertex side. Use cross - sections to calculate the ratio of the area of air to tissues in the body layer by layer. Take the layer with the largest ratio of the area of air to tissues in the candidate layer range as the reference cross - section.

2. The image processing method for separating the head and neck part according to claim 1, wherein, The third solution includes determining the candidate range of the layer where the reference cross-section is located. This range is from the candidate starting layer S0 to the candidate ending layer S1, and the layer with the distance ratio to the first and last layers closest to the cranial apex side being closest to the starting layer ratio is used as the candidate starting layer S0. The total number of image layers is K, where the starting layer ratio has an optimal value of ; The ratio of the distances to the first layer and the last layer close to the cranial vertex side is the ratio closest to the termination layer ratio of the layer, which is used as the candidate termination layer S1, where the termination layer ratio has an optimal value of 7:3; then, by calculating the ratio of the area of in-vivo air to human tissue within the range from the candidate starting layer S0 to the candidate termination layer S1, the in-vivo air-tissue ratio Rb of each layer is obtained, and the layer where the maximum value of the in-vivo air-tissue ratio Rb is located is selected as the reference cross-section Sr.

3. The image processing method for separating the head and neck part according to claim 1, wherein, The specific content of the step S3 includes: The planar position difference between the head - neck separation layer Ss and the reference cross - section Sr is D. Taking the direction from the top of the head to the neck as the positive direction, the head - neck separation layer Ss is located at a distance of D from the reference cross - section Sr along the positive direction. When the distance exceeds the image range, the head - neck separation layer Ss is selected as the layer with the maximum distance in the positive direction.

Images