Chest CT image processing method and system

Abstract

The invention discloses a chest CT (Computed Tomography) image processing method and system, and relates to the technical field of medical image processing, and the main scheme is as follows: acquiring image data of two groups of X-rays with different energies and attenuation measurement values of pixel points by adopting dual-energy CT scanning, respectively calculating concentration values of soft tissues and blood vessels, and generating concentration maps; generating an initial image based on the soft tissue and blood vessel concentration map, calculating a pixel value, obtaining a target image pixel value through contrast stretching, segmenting the target image pixel value to obtain a binary image, and distinguishing a blood vessel region and a soft tissue region in the binary image by identifying and analyzing a connected region in the binary image; carrying out center line extraction on the blood vessel region, calculating a local diameter, and respectively comparing the local diameter with a blood vessel dilation threshold value and a blood vessel constriction threshold value to judge abnormal blood vessel dilation or constriction; extracting features from the soft tissue region to form a first feature vector, and comparing the first feature vector with a preset second feature vector to judge abnormality; abnormal conditions are automatically labeled or highlighted, so that the diagnosis accuracy and efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to a chest CT image processing method and system. Background Art

[0002] With the continuous development of medical imaging technology, chest CT examination plays an increasingly important role in clinical applications. The chest contains a wealth of soft tissue and vascular structures, and it is crucial to accurately analyze and identify whether these structures are abnormal.

[0003] Traditional chest CT image processing methods often have difficulty in clearly distinguishing soft tissue and blood vessels, which brings certain difficulties to doctors' diagnosis. In clinical practice, doctors need to accurately judge whether there are abnormalities such as dilation and contraction of blood vessels, as well as whether there are abnormalities in soft tissue. However, due to the complexity of chest structure and the limitations of traditional methods, accurate analysis of blood vessels and soft tissue has always been a challenge.

[0004] At present, although some image processing methods have been tried to process chest CT images, they are still insufficient in distinguishing blood vessels and soft tissue areas and judging abnormal conditions of blood vessels and soft tissues. Summary of the invention

[0005] 1. Technical issues to be resolved In view of the deficiencies in the prior art, the present invention provides a chest CT image processing method and system, which obtains X-ray image data of different energies through dual-energy CT scanning, calculates attenuation measurement values ​​and concentration values ​​of soft tissues and blood vessels and converts them into concentration maps, generates a contrast initial image based on this and stretches it to obtain a target image; after segmenting the target image, the connected areas of the binary image are marked and analyzed to distinguish between blood vessels and soft tissues, the center line of the blood vessel area is extracted to calculate the local diameter and compare it with a preset threshold to determine blood vessel abnormalities, the feature component vector is extracted from the soft tissue area and compared with the threshold to determine soft tissue abnormalities, and the identified abnormalities are automatically marked or highlighted to improve the accuracy and efficiency of doctors' diagnosis.

[0006] (II) Technical solution To achieve the above objectives, the present invention is implemented through the following technical solutions: A chest CT image processing method, comprising: Step 1: using dual-energy CT scanning technology to respectively obtain two sets of image data 1 and image data 2 of X-rays with different energies; obtaining first attenuation measurement values ​​and second attenuation measurement values ​​corresponding to each pixel point under the two sets of X-rays with different energies through image data 1 and image data 2; Step 2: Calculate the concentration value of the soft tissue at each pixel point according to the first attenuation measurement value corresponding to each pixel point Calculate the concentration value of the blood vessel at each pixel point according to the second attenuation measurement value corresponding to each pixel point Generate a soft tissue concentration map based on the concentration values ​​corresponding to all soft tissues at each pixel point; generate a blood vessel concentration map based on the concentration values ​​of all blood vessels at each pixel point; Step 3: Generate a soft tissue-vessel contrast initial image based on the soft tissue concentration map and the blood vessel concentration map, and calculate the initial image pixel value corresponding to each pixel point on the soft tissue-vessel contrast initial image according to the concentration value corresponding to each pixel point of the soft tissue and the concentration value corresponding to each pixel point of the blood vessel; Step 4: performing contrast stretching on the initial image pixel value corresponding to each pixel point on the soft tissue-vessel contrast initial image, and calculating the target image pixel value corresponding to each pixel point in the soft tissue-vessel contrast target image; Step 5: Perform image segmentation on the target image pixel values ​​corresponding to each pixel point in the soft tissue-vessel contrast target image to obtain each pixel value in the binary image, and generate a binary image based on all the pixel values; distinguish the vascular area and the soft tissue area in the binary image by identifying and analyzing the connected areas in the binary image; Step 6: Extract the centerline of the blood vessel in the connected area and calculate the local diameter of the blood vessel Preset a vascular expansion threshold and a vascular contraction threshold, compare the local diameter of the blood vessel with the vascular expansion threshold to determine whether the blood vessel is abnormally expanded, and compare the local diameter of the blood vessel with the vascular contraction threshold to determine whether the blood vessel is abnormally contracted; Step 7: extract soft tissue features in the connected area belonging to the soft tissue area, and combine the extracted features into a first feature vector; preset a second feature vector, compare the first feature vector with the second feature vector, and determine whether the soft tissue has abnormalities; Step 8: Automatically mark or highlight the identified abnormal situations.

[0007] In the preferred embodiment of the above-mentioned chest CT image processing method, the concentration value of the soft tissue at each pixel point and the concentration value of the blood vessel at each pixel point are calculated based on the following formula:

[0008] in, Indicates energy In pixels The first attenuation measurement of Indicates energy In pixels The second attenuation measurement value, Is soft tissue A in energy The attenuation coefficient under , the value range is non-negative real number; It is the energy of blood vessel B The attenuation coefficient under , the value range is non-negative real number; Is soft tissue A in energy The attenuation coefficient under , the value range is non-negative real number; It is the energy of blood vessel B The attenuation coefficient under , the value range is non-negative real number; is soft tissue A at pixel point The corresponding concentration value is It is the blood vessel B at the pixel point The corresponding concentration value is and The value range is non-negative real numbers, indicating the distribution of matter in the image.

[0009] In the preferred embodiment of the above-mentioned chest CT image processing method, the method of forming the soft tissue concentration map and the blood vessel concentration map is: Two two-dimensional arrays are created for soft tissue and blood vessels respectively, and their sizes are the same as image data one and image data two; the concentration value corresponding to the soft tissue at each pixel point is stored in the corresponding position in the two-dimensional array of the soft tissue to form a soft tissue concentration map; the concentration value corresponding to the blood vessel at each pixel point is stored in the corresponding position in the two-dimensional array of the blood vessel to form a blood vessel concentration map; each pixel value in each concentration map represents the concentration value of the soft tissue or blood vessel at the pixel point at the corresponding position.

[0010] In the preferred embodiment of the above-mentioned chest CT image processing method, the initial image pixel value corresponding to each pixel point on the soft tissue-vessel contrast initial image is calculated based on the formula:

[0011] in, Represents the pixel point on the initial image of soft tissue-vessel contrast The corresponding initial image pixel value is yes The weight coefficient ranges from 0 to 1. is soft tissue A at pixel point The corresponding concentration value is It is the blood vessel B at the pixel point The corresponding concentration value.

[0012] In the preferred embodiment of the above-mentioned chest CT image processing method, the target image pixel value corresponding to each pixel point in the soft tissue-vessel contrast target image is calculated based on the formula:

[0013] in, Represents the pixel point in the target image of soft tissue-vessel contrast The corresponding target image pixel value is is the minimum pixel value of the initial image of soft tissue-vessel contrast, is the maximum pixel value of the initial soft tissue-vessel contrast image, ranging from ; M is the upper limit of the target pixel value range after stretching, m is the lower limit of the target pixel value range, ; Used to prevent the denominator from being zero. The value range is .

[0014] In the preferred embodiment of the above-mentioned chest CT image processing method: calculating each pixel value in the binary image The formula is:

[0015] in, Represents the value of each pixel in the binary image. When the value is 1, it represents the foreground area, which represents the blood vessels and soft tissue area. When the value is 0, it represents the background area, which represents the removed area. It is the segmentation threshold, which is used to separate the foreground area from the background area.

[0016] In the preferred embodiment of the above-mentioned chest CT image processing method, the specific method for distinguishing different blood vessels and soft tissue areas is: Identify and label binary images Every connected region in ; For each marked connected region , extract connected regions All pixel coordinates that satisfy =1; is the index of the region; By analyzing each connected region , obtain the dimension data, the dimension data at least includes the area , circularity , rectangularity And aspect ratio ; Set the blood vessel area range [ and soft tissue area , Vessel circularity threshold T and soft tissue circularity threshold Y, blood vessel rectangularity range [ and soft tissue rectangular range [ and the range of vessel aspect ratios[ and soft tissue aspect ratio range [ ; The area Respectively with the range of blood vessel area [ and soft tissue area Compare and determine the connected areas belongs to the vascular area or soft tissue area; The circularity The vessel circularity threshold T and soft tissue circularity threshold Y is compared to determine the connected area belongs to the vascular area or soft tissue area; The rectangular Respectively with the range of blood vessel rectangularity [ and soft tissue rectangular range [ Compare and determine the connected areas belongs to the vascular area or soft tissue area; The aspect ratio Respectively with the range of vessel length-width ratio [ and soft tissue aspect ratio range [ Compare and determine the connected areas belongs to the vascular area or soft tissue area; When the area of ​​the connected region , circularity , rectangularity and aspect ratio When all of them meet the characteristic range of the blood vessel region, it is determined that the connected region belongs to the blood vessel region; When the area of ​​the connected region , circularity , rectangularity and aspect ratio When all of them meet the characteristic range of the soft tissue area, it is determined that the connected area belongs to the soft tissue area; When the area of ​​the connected region appears , circularity , rectangularity and aspect ratio When one or several of the items do not belong to the characteristic range of the soft tissue area or the vascular area at the same time, a secondary judgment is required based on the performance of the connected area at different angles and positions.

[0017] In the preferred embodiment of the above-mentioned chest CT image processing method, the specific method for determining whether there is an abnormality in a blood vessel is: For connected regions belonging to vascular regions Extract the centerline of the blood vessel and estimate the local diameter of the blood vessel by calculating the distance between adjacent points on the centerline ; Compare the local diameter of the blood vessel with the set blood vessel dilation threshold Te and blood vessel constriction threshold Tc respectively: When > Te, it is determined that the blood vessel is dilated; When < Tc, it is determined that the blood vessel is constricted; When Tc ≤ ≤ Te, it is determined to be normal; The specific method for determining whether there is an abnormality in the soft tissue is: Perform various feature extractions on the connected region belonging to the soft tissue region to obtain the area , circularity , the contrast Con and energy Asm of the gray-level co-occurrence matrix of this connected region Combine the area , circularity , the contrast Con and energy Asm of the gray-level co-occurrence matrix into a first feature vector f(x, y); the formula is:

[0018] Preset a second feature vector F(x, y); When , it indicates that the region where the point is located is abnormal; When , it indicates that the region where the point is located is normal.

[0019] In the preferred solution of the above-mentioned chest CT image processing method: Highlight the identified blood vessel abnormal region and soft tissue abnormal region with a color or brightness different from the surrounding tissue, add a text label near the abnormal region to indicate the type of abnormality and mark at least one of the circle, square, and arrow graphics in the abnormal region.

[0020] The present invention also discloses a chest CT image processing system, including: An image acquisition module for respectively acquiring image data one and image data two of two different energy X-rays by using dual-energy CT scanning technology; respectively obtaining the corresponding first attenuation measurement value and second attenuation measurement value at each pixel point under the two different energy X-rays through the image data one and image data two; A concentration calculation module for calculating the concentration value corresponding to the soft tissue at each pixel point according to the corresponding first attenuation measurement value at each pixel point Calculating the concentration value corresponding to the blood vessel at each pixel point according to the corresponding second attenuation measurement value at each pixel point Generate a soft tissue concentration map based on the concentration values ​​corresponding to all soft tissues at each pixel point; generate a blood vessel concentration map based on the concentration values ​​of all blood vessels at each pixel point; An image fusion module generates a soft tissue-vessel contrast initial image based on the soft tissue concentration map and the blood vessel concentration map, and calculates the initial image pixel value corresponding to each pixel point on the soft tissue-vessel contrast initial image according to the concentration value corresponding to each pixel point of the soft tissue and the concentration value corresponding to each pixel point of the blood vessel; A contrast stretching module is used to perform contrast stretching on the initial image pixel value corresponding to each pixel point on the soft tissue-vessel contrast initial image, and calculate the target image pixel value corresponding to each pixel point in the soft tissue-vessel contrast target image; The region distinction module is used to perform image segmentation on the target image pixel values ​​corresponding to each pixel point in the soft tissue-vessel contrast target image, obtain each pixel value in the binary image, and generate a binary image based on all the pixel values; by identifying and analyzing the connected regions in the binary image, the vascular region and the soft tissue region in the binary image are distinguished; The vascular region anomaly recognition module is used to extract the vascular centerline of the connected area belonging to the vascular region and calculate the local diameter of the vascular region. Preset a vascular expansion threshold and a vascular contraction threshold, compare the local diameter of the blood vessel with the vascular expansion threshold to determine whether the blood vessel is abnormally expanded, and compare the local diameter of the blood vessel with the vascular contraction threshold to determine whether the blood vessel is abnormally contracted; The soft tissue area abnormality recognition module extracts soft tissue features in the connected area belonging to the soft tissue area, and combines the extracted features into a first feature vector; a second feature vector is preset, and the first feature vector is compared with the second feature vector to determine whether the soft tissue has an abnormality; The result display module is used to automatically mark or highlight the identified abnormal situations.

[0021] (III) Beneficial effects The present invention provides a chest CT image processing method and system, which have the following beneficial effects: (1) Using dual-energy CT scanning technology to obtain image data of X-rays of different energies can more accurately measure the attenuation characteristics of soft tissues and blood vessels. By obtaining the attenuation measurement values ​​of soft tissues and blood vessels at different energies, a more accurate data basis is provided for the subsequent calculation of concentration values, which helps to more clearly display the structure of soft tissues and blood vessels and improve the recognition of lesions.

[0022] (2) After weighted fusion, the initial soft tissue-vessel contrast image is obtained, and then contrast stretching is performed to obtain the target image, which significantly enhances the contrast between soft tissue and blood vessels. This enables doctors to more clearly distinguish the boundaries and abnormalities of soft tissue and blood vessels when observing the image, thereby improving the accuracy of diagnosis.

[0023] (3) Image segmentation of the soft tissue-vessel contrast target image can enhance the contrast between soft tissue and blood vessels, better distinguish different areas, and improve the accuracy of abnormal identification.

[0024] (4) Centerline extraction and local diameter calculation of the vascular area, as well as feature extraction and threshold comparison of the soft tissue area, can effectively determine whether there are abnormalities in the blood vessels and soft tissues, providing doctors with more detailed diagnostic information; this automated processing method greatly reduces the time and workload of doctors in manually analyzing images, thereby improving diagnostic efficiency.

[0025] (5) Identified abnormalities are automatically labeled or highlighted, allowing doctors to quickly focus on areas where problems may exist, further saving diagnosis time while also reducing the risk of missed diagnosis due to human negligence. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 The figure is a schematic diagram of the working steps of a chest CT image processing method of the present invention. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] See also Figure 1 The present invention provides a chest CT image processing method, comprising: Step 1: Use dual-energy CT scanning technology to respectively obtain two sets of image data 1 and image data 2 of X-rays with different energies; obtain the first attenuation measurement value and the second attenuation measurement value corresponding to each pixel point under the two sets of X-rays with different energies through image data 1 and image data 2.

[0029] The specific steps of step one are: Step 101: Select a CT scanner that supports dual-energy scanning, and select two appropriate X-ray energies based on clinical needs and the characteristics of the target tissue. One is a lower energy, such as 80 kVp, for enhancing the contrast of soft tissues; the other is a higher energy, such as 140 kVp, for enhancing the contrast of bones and calcified tissues.

[0030] Step 102: During the scanning process, the CT scanner will automatically switch the X-ray energy and simultaneously acquire image data one and image data two.

[0031] Step 103: performing image denoising, artifact correction, energy separation and other steps on the image data 1 and the image data 2 to improve the image quality.

[0032] It should be noted that the X-ray energy can be determined through preliminary experiments or literature research to determine the optimal energy combination to produce the greatest contrast difference between soft tissue and blood vessels.

[0033] Combined with the contents of steps 101 to 103: Dual-energy CT scanning technology acquires two sets of image data with different energies through a single scan, significantly enhancing tissue contrast, optimizing image quality, and providing additional tissue information, while reducing radiation dose, improving diagnostic efficiency, and supporting more accurate medical decisions.

[0034] Step 2: Calculate the concentration value of the soft tissue at each pixel point according to the first attenuation measurement value corresponding to each pixel point Calculate the concentration value of the blood vessel at each pixel point according to the second attenuation measurement value corresponding to each pixel point A soft tissue concentration map is generated based on the concentration values ​​corresponding to all soft tissues at each pixel point; and a blood vessel concentration map is generated based on the concentration values ​​corresponding to all blood vessels at each pixel point.

[0035] The specific steps of step 2 are: Step 201: Calculate the concentration value of soft tissue at each pixel point The concentration value of blood vessels at each pixel point is based on the following formula:

[0036] in, Indicates energy In pixels The first attenuation measurement of Indicates energy In pixels The second attenuation measurement value, Is soft tissue A in energy The attenuation coefficient under , the value range is non-negative real number; It is the energy of blood vessel B The attenuation coefficient under , the value range is non-negative real number; Is soft tissue A in energy The attenuation coefficient under , the value range is non-negative real number; It is the energy of blood vessel B The attenuation coefficient under , the value range is non-negative real number; is soft tissue A at pixel point The corresponding concentration value is It is the blood vessel B at the pixel point The corresponding concentration value is and The value range is non-negative real numbers, indicating the distribution of matter in the image.

[0037] It should be noted that the principle of this formula is that in dual-energy CT data, soft tissue A and blood vessel B have different attenuation characteristics under X-rays of different energies. This difference in characteristics is the key to distinguishing different substances. and Down in pixels The measured attenuation value and The attenuation coefficient of the material at a specific energy is used to collect , , , , construct a set of equations to solve the concentration of soft tissue A and blood vessel B at this point and ; Using the information contained in the dual-energy data, the concentration information of iodine contrast agents in soft tissues and blood vessels is separated from the mixed attenuation information, thereby extracting the concentration maps of soft tissues and blood vessels from complex CT data. This provides a key data basis for subsequent analysis of the distribution and characteristics of soft tissues and blood vessels in the image.

[0038] Attenuation coefficient , , , The specific value depends on the substance and energy.

[0039] The method of forming the soft tissue concentration map and the blood vessel concentration map is: Create two 2D arrays for soft tissue and blood vessels respectively, with the same size as the original chest CT image. Assume that the image size is W×H, where W and H represent the number of rows and columns of the image, respectively. Each pixel in the image is represented by the coordinates To indicate that The value range is from 0 to W-1. The value range is 0 to H-1.

[0040] The concentration value corresponding to the soft tissue at each pixel point is stored in the corresponding position in the two-dimensional array of the soft tissue, and the concentration value corresponding to the blood vessel at each pixel point is stored in the corresponding position in the two-dimensional array of the blood vessel.

[0041] The two-dimensional array of soft tissue constitutes a soft tissue concentration map, and the two-dimensional array of blood vessels constitutes a blood vessel concentration map. Each pixel value in each concentration map represents the concentration value of the soft tissue or blood vessel at the corresponding pixel point.

[0042] Combined with the content of step 2: Through the dual-material decomposition algorithm, the attenuation measurement values ​​obtained by dual-energy CT scanning are used to accurately calculate the concentration values ​​of soft tissues and blood vessels, thereby extracting the concentration map of soft tissues and blood vessels from the dual-energy CT data. This method can effectively separate and quantify different tissue components, provide more accurate and detailed tissue information for clinical diagnosis and treatment, and help doctors make more accurate medical decisions.

[0043] Step 3: Generate a soft tissue-vessel contrast initial image based on the soft tissue concentration map and the blood vessel concentration map, and calculate the initial image pixel value corresponding to each pixel point on the soft tissue-vessel contrast initial image according to the concentration value corresponding to the soft tissue at each pixel point and the concentration value corresponding to the blood vessel at each pixel point.

[0044] The specific steps of step three are: Step 301: Calculate the initial image pixel value corresponding to each pixel point on the soft tissue-vessel contrast initial image , the formula based on is:

[0045] in, Represents the pixel point on the initial image of soft tissue-vessel contrast The corresponding initial image pixel value is yes The weight coefficient ranges from 0 to 1. is soft tissue A at pixel point The corresponding concentration value is It is the blood vessel B at the pixel point The corresponding concentration value.

[0046] It should be noted that in this formula is a weight coefficient, and its value range is usually between 0 and 1. When it is close to 0, the fused image is closer to the density map of soft tissue; when it is close to 1, the fused image is closer to the density map of blood vessels; by adjusting The value of can be used to control the display degree of the blood vessel area in the image according to specific needs.

[0047] is the weighted soft tissue A at the pixel point The corresponding concentration value and the weighted blood vessel B at pixel point The corresponding concentration value Obtained by addition.

[0048] For the concentration value corresponding to blood vessel B Weighting Operations can be used to emphasize or weaken the display of vascular areas. For example, the blood vessels can be more prominent, and the The value of makes the concentration of blood vessels account for a larger proportion in the fused image; relatively weakening the display of blood vessels can reduce The value of .

[0049] The weighted vascular concentration value is added to the soft tissue concentration value to achieve the fusion of the two material concentration maps. The advantage of this is that the information of soft tissue and blood vessels can be displayed simultaneously in one image, and the display effect of the blood vessel area can be directly controlled by adjusting the weight, so that doctors can more easily observe and analyze the relationship between soft tissue and blood vessels and the specific characteristics of blood vessels.

[0050] Step 3011: The calculation method is: Preset area threshold H; Compare with the regional threshold H to identify the blood vessel area. Specifically: when , the pixel belongs to the blood vessel area.

[0051] when , the pixel belongs to the non-vascular area.

[0052] The setting of the threshold H is determined based on experience or experimental data, and in practical applications it needs to be determined and adjusted in combination with the characteristics of the current image data.

[0053] based on And the regional threshold H is calculated , the formula based on is:

[0054] in, is the maximum weight of the blood vessel region, which determines the prominence of the blood vessel in the fused image and has a value range of ; is the weight of the non-vascular area or background area, and its value range is , The value of determines the visibility of the non-vascular area in the fused image. For example, when When it is close to 0, the display of non-vascular areas can be reduced to better highlight blood vessels; When it is close to 1, it can make the non-vascular area more obvious; Used to control the width of the transition area. The value range is positive. Small Values ​​make the transition region narrower and the weight change steeper; large Higher values ​​make the transition region wider and the weight change more gradual. , , The specific value of needs to be adjusted according to the specific situation in practical applications.

[0055] It should be noted that the purpose of this formula is to determine the weight of each pixel in the chest CT image according to the concentration value of the blood vessel B and the threshold H through dynamic weighted fusion, so as to achieve different display effects for soft tissue and blood vessels, highlight the blood vessel area while also better displaying the soft tissue area.

[0056] In this formula Indicates that when the value of the blood vessel concentration map at the coordinate point is greater than When , the weight given to this point is This usually corresponds to a higher concentration of blood vessels at that point, which may mean that the area has some specific characteristics and needs to be given a larger weight for subsequent processing.

[0057] Indicates that when the value of the blood vessel concentration map at the coordinate point is and Between, through a linear function To calculate the weight of the point. This linear function The role of is as the concentration value of the blood vessel gradually increases from close to the threshold to the threshold plus , the weights are shifted from those close to the non-vascular region Gradually increase to the maximum weight of the vascular region ; indicates a smooth transition from non-vascular area to vascular area, avoiding obvious boundaries.

[0058] covers all situations where the first two specific conditions are not met, i.e. when , it means that the concentration of blood vessels at this point is relatively low, or it is not in the range that requires special weighting, so a smaller weight is given. .

[0059] By setting This condition ensures that there is a clear weight assignment rule for all possible concentration values, which ensures that each pixel in the entire image can get a reasonable weight, thereby achieving comprehensive and accurate processing of the image.

[0060] Represents the weight of each pixel, and its value is dynamically determined according to the position of the pixel, the concentration value of the blood vessel and related parameters.

[0061] Combined with the content of step three: the vascular area is identified by preset thresholds, and the weights are dynamically adjusted based on the substance concentration and thresholds, so that the soft tissue and blood vessels can be effectively distinguished and highlighted in the image. Further, through weighted fusion and contrast stretching, the contrast between soft tissue and blood vessels is significantly enhanced, making the vascular structure clearer and more recognizable, while retaining the detailed information of soft tissue. This processing method provides doctors with more intuitive and accurate image information, which helps to improve the accuracy and efficiency of diagnosis.

[0062] Step 4: Perform contrast stretching on the initial image pixel value corresponding to each pixel point on the soft tissue-vessel contrast initial image, and calculate the target image pixel value corresponding to each pixel point in the soft tissue-vessel contrast target image.

[0063] The specific steps of step 4 are: Step 401: Calculate the target image pixel value corresponding to each pixel point in the soft tissue-vessel contrast target image, based on the formula:

[0064] in, Represents the pixel point in the target image of soft tissue-vessel contrast The corresponding target image pixel value is is the minimum pixel value of the initial image of soft tissue-vessel contrast, is the maximum pixel value of the initial soft tissue-vessel contrast image, ranging from ; M is the upper limit of the target pixel value range after stretching, m is the lower limit of the target pixel value range, ; Used to prevent the denominator from being zero. The value range is .

[0065] It should be noted that the principle of this formula is to use Will The value of is normalized to the interval [0,1], and the normalized value is stretched to the new range [m,M] by linear transformation, and then multiplied by And add To achieve this, is the upper limit of the target pixel value range after stretching, is the lower limit of the target pixel value range after stretching; the role of linear transformation is to expand the difference between different areas in the image and enhance the contrast, multiplying This achieves a proportional amplification of the normalized value, plus The enlarged range is translated to the [m, M] interval, and the pixel values ​​in the initial image are normalized and linearly transformed and stretched to the new target pixel value range to obtain the target image pixel values ​​of the soft tissue-vascular contrast target image. .

[0066] For 8-bit images, The value range is ; The value range is ; It needs to be determined according to the data type of the image. For example, for an 8-bit image, the value range is m=0, M=255.

[0067] Combined with the content of step 4: Through contrast stretching technology, the pixel value range of the initial soft tissue-vascular contrast image is linearly transformed from the original range to a wider or more suitable range for display, which significantly enhances the contrast of the image. This processing not only makes the vascular structure more prominent, but also maintains the overall brightness and details of the image, improving the visual quality and readability of the image. For 8-bit images, by setting a suitable stretching range, the best display effect of the image on the display device can be ensured, providing doctors with clearer and more intuitive image information, which helps to improve the accuracy and efficiency of diagnosis.

[0068] Step 5: Perform image segmentation on the target image pixel values ​​corresponding to each pixel point in the soft tissue-vessel contrast target image to obtain each pixel value in the binary image, and generate a binary image based on all the pixel values; distinguish the vascular area and the soft tissue area in the binary image by identifying the connected areas in the binary image and analyzing them.

[0069] The specific steps of step five are: Step 501: Calculate the value of each pixel in the binary image The formula is:

[0070] in, Represents the value of each pixel in the binary image. When the value is 1, it represents the foreground area, which represents the blood vessels and soft tissue area. When the value is 0, it represents the background area, which represents the removed area. It is the segmentation threshold, which is used to separate the foreground area from the background area.

[0071] It should be noted that the principle of this formula is to determine whether the pixel belongs to the foreground area or the background area according to the size of the pixel value, and select the threshold To divide the pixels in the image. For each pixel value in the soft tissue-vessel contrast target image , and the threshold Make a comparison.

[0072] when , the pixel is determined to be the foreground area and assigned a value of "1". It means that the pixel is considered to belong to the blood vessel and soft tissue area.

[0073] when , the pixel is determined to be the background area and is assigned a value of "0". This means that the image is divided into two parts: foreground and background.

[0074] A binary image is a digital image in which each pixel has only two possible values, represented by 0 and 1, or represented by black and white; in this embodiment, the binary image is obtained by performing image segmentation on the soft tissue-vessel contrast target image, wherein the foreground area, such as the blood vessel and soft tissue area, is represented by 1, and the background area is represented by 0.

[0075] The threshold is set by analyzing the grayscale histogram of the image, trying different thresholds, and evaluating the segmentation results to determine the best threshold. The appropriate threshold can accurately distinguish blood vessels and soft tissue areas from background areas while minimizing misclassified pixels.

[0076] Step 502: The specific method of distinguishing different blood vessels and soft tissue areas is: Step 5021: Identify and label binary images Every connected region in , for each marked connected region , extract connected regions All pixel coordinates that satisfy =1; is the index of the region.

[0077] It should be noted that the concept of connected regions is based on the connection relationship between pixels. By determining which pixels can reach each other, the foreground area is divided into different connected parts. Each connected part corresponds to a specific blood vessel or soft tissue area, which helps to more accurately analyze and describe different blood vessels and soft tissue structures, because different areas may have different shapes, sizes and characteristics.

[0078] A connected region refers to all pixels with a value of 1 in a binary image. If they can reach each other through 4-connectivity or 8-connectivity, then these pixels are considered to belong to the same connected region.

[0079] Among them, 4-connectivity: a pixel is connected to the pixels above, below, left, and right. This connection method can effectively identify and separate some simpler areas in some cases, but may miss some diagonal connections.

[0080] 8-Connected: In addition to the four directions of up, down, left, and right, it also includes the four directions on the diagonal: upper left, upper right, lower left, and lower right. That is, a pixel is connected to the 8 pixels around it. This broader connection method can more comprehensively capture the connection relationship between pixels, especially for blood vessels and soft tissue areas with irregular shapes or diagonal connections, and can more accurately identify connected areas.

[0081] Starting from any pixel with a value of 1 in the image, check whether the surrounding pixels are also 1 according to the definition of 4-connectivity or 8-connectivity. If so, these pixels are classified as the same connected region and continue to expand outward until all pixels connected to the starting pixel are marked as belonging to the same region. Select another unmarked pixel and repeat the above process until all pixels with a value of 1 in the image are assigned to different connected regions.

[0082] Step 5022: For each marked connected area , extract connected regions All pixel coordinates that satisfy =1, that is, these pixels belong to the foreground.

[0083] Step 5023: By analyzing each connected region , get the size data, the size data includes the area ,perimeter , centroid ( , )、Bounding Box Properties( , , , ), circularity , rectangularity And aspect ratio ;in: Calculate area The formula used is as follows:

[0084] Preset blood vessel area range [ and soft tissue area .

[0085] when , indicating that the area is a vascular area.

[0086] when , indicating that the area is a soft tissue area.

[0087] It should be noted that the principle of this formula is to calculate the connected area The number of foreground pixels in the connected region is used to determine the size of the region. Each pixel with a value of 1 represents a unit area, and the number of pixels is accumulated to obtain the area of ​​the connected region.

[0088] Vascular area range Source: The area distribution of vascular connectivity regions is statistically analyzed and measured through analysis of vascular regions in a large number of normal chest CT images.

[0089] Considering the possible differences in area of ​​different types of blood vessels such as arteries, veins, etc. as well as blood vessels in different parts, a reasonable area range should be determined.

[0090] For example, chest CT image samples of patients of different ages, genders, and health conditions can be selected, and the blood vessels therein can be manually labeled or automatically segmented and their areas calculated to obtain an approximate range of blood vessel areas.

[0091] Soft tissue area Source: Analysis of soft tissue areas in a large number of chest CT images; soft tissue includes muscle, fat and other tissues, and its area is usually larger than that of blood vessels and has a wider range of variation.

[0092] Combined with anatomical knowledge, the soft tissue characteristics of different parts are considered, such as the common size ranges of chest muscle tissue and fat tissue.

[0093] By marking and calculating the area of ​​the soft tissue in the sample image, the distribution range of the soft tissue area is statistically obtained.

[0094] Calculate the perimeter The formula used is as follows:

[0095] It should be noted that the principle of this formula is to estimate the perimeter by calculating the number of pixels on the boundary of the region. Boundary pixels refer to foreground pixels that are adjacent to the background. Each boundary pixel represents a unit length, and the number of boundary pixels is accumulated to obtain the approximate perimeter of the connected region.

[0096] Calculate the centroid ( , ) is based on the following formula:

[0097] It should be noted that the principle of this formula is to calculate the average centroid of all pixel coordinates in the region as an estimate of the geometric center of the region; for connected regions , respectively, for all pixels in the region and the vertical coordinate The sum is then divided by the area of ​​the pixels in the region, and the sum is the horizontal and vertical coordinates of the centroid. It reflects the average position of the pixels in the region and can be used to describe the location characteristics of the region.

[0098] Compute bounding box ( , , , ) is based on the following formula:

[0099]

[0100] It should be noted that the principle of this formula is to extract the minimum circumscribed rectangle of the area, and the four boundary coordinates of the rectangle are ( , , , ), for connected areas , find the minimum value of the horizontal coordinates of all pixels in the area , the minimum value of the vertical axis , the maximum value of the horizontal axis and the maximum value of the vertical axis ,These four values ​​determine a minimum enclosing rectangle, i.e., the bounding box, which can concisely describe the approximate range and shape of the area, facilitating subsequent analysis and processing.

[0101] Calculate circularity The formula used is as follows:

[0102] Preset blood vessel circularity threshold T and soft tissue circularity threshold Y.

[0103] when T, indicating that the area is a vascular area.

[0104] when < Y, indicating that the area is a soft tissue area.

[0105] It should be noted that this formula is used to measure the circularity of an area. The closer the value is to 1, the closer the area is to a circle. If an area is a perfect circle, then its perimeter is and area Satisfaction between , then the circularity C = 1. For the actual area, the closer its shape is to a circle, the smaller its circumference is. With area The closer the relationship is to a circle, the closer the circularity value is to 1. Conversely, if the shape of the area is irregular or deviates from a circle, the circularity value will be less than 1.

[0106] Vessel circularity threshold The method for setting T is as follows: based on the observation and analysis of the shape of normal blood vessels, blood vessels usually have a high degree of circularity, especially some larger blood vessels. The circularity of a large number of normal blood vessel connected areas is calculated to obtain an average circularity value or a typical circularity range. The upper limit of this range or a specific value is used as the blood vessel circularity threshold. When the circularity is greater than this threshold, it indicates that this area is a blood vessel.

[0107] Soft tissue circularity threshold The setting method of Y is as follows: the shape of soft tissue is usually more irregular and has a lower circularity. The circularity of the soft tissue area in the sample image is calculated, and the distribution of the circularity of the soft tissue is statistically obtained. An appropriate circularity value is determined as the soft tissue circularity threshold. When the circularity is less than this threshold, it indicates that this area is a soft tissue area.

[0108] Calculate rectangularity The formula used is as follows:

[0109] Preset blood vessel rectangularity range [ and soft tissue rectangular range [ .

[0110] when , indicating that the area is a vascular area.

[0111] when , indicating that the area is a soft tissue area.

[0112] It should be noted that this formula represents the degree to which the shape of the area is close to a rectangle. The higher the value, the closer the shape is to a rectangle. is the area of ​​the connected region, is the area of ​​the minimum enclosing rectangle of the region. If the shape of the region is completely rectangular, then its area is equal to the area of ​​the minimum enclosing rectangle, and the rectangularity is higher. The higher the value, the closer the shape of the region is to a rectangle, because the region can more tightly fill its minimum enclosing rectangle.

[0113] Vascular rectangularity range[ Origin: The shape of a blood vessel is unlikely to be close to a rectangle, so its rectangularity is usually low. The rectangularity of a large number of normal blood vessel connected areas is calculated and statistically analyzed to obtain the approximate range of blood vessel rectangularity. This range reflects the similarity between blood vessels and rectangles. When the rectangularity is within this range, it indicates that it is a blood vessel area.

[0114] Soft tissue rectangular range[ Source: The shape of the soft tissue area varies greatly, and the range of rectangularity is correspondingly wider. Considering different types of soft tissues, such as muscles may have a certain long strip shape, and fat tissue may be more irregular, the range of soft tissue rectangularity is determined by calculating and counting the rectangularity of the soft tissue in the sample image.

[0115] Calculate aspect ratio The formula used is as follows:

[0116] in, is the width of the bounding box, is the height of the bounding box.

[0117] Preset vessel aspect ratio range [ and soft tissue aspect ratio range [ .

[0118] when , indicating that the area is a vascular area.

[0119] when , indicating that the area is a soft tissue area.

[0120] It should be noted that this formula describes the proportional relationship of the region in the horizontal and vertical directions. Different aspect ratios can reflect the shape characteristics of the region, such as a strip, a square, etc.

[0121] Range of vessel aspect ratio Source: The length-to-width ratio of blood vessels, especially some larger ones, is usually close to 1, especially those that are circular or nearly circular. The length-to-width ratios of a large number of normal blood vessel connected areas are calculated and analyzed to obtain the typical range of the blood vessel length-to-width ratio. When the length-to-width ratio of a connected area is within this range, it indicates that this area is a blood vessel area.

[0122] Soft tissue aspect ratio range[ Source: The aspect ratio of soft tissue areas varies greatly, depending on their specific shape and structure. Combining anatomical knowledge and observation of soft tissue in sample images, the possible aspect ratio range of different types of soft tissues is determined. This range reflects the proportional relationship of soft tissue in the horizontal and vertical directions. When the aspect ratio of the connected area is within this range, it indicates that this area is a soft tissue area.

[0123] Step 5024: The method for determining the blood vessel region and the soft tissue region is as follows: The method for determining a vascular area is: When the area of ​​the connected region , circularity , rectangularity and aspect ratio When the area is within the blood vessel area range and the circularity is greater than or equal to the blood vessel circularity threshold T, rectangularity is within the range of blood vessel rectangularity, and aspect ratio is within the range of blood vessel aspect ratio, and the connected area is judged to belong to the blood vessel area. This shows that the area is consistent with the typical characteristics of blood vessels in all aspects and is judged to be a blood vessel area.

[0124] The method for determining the soft tissue area is: When the area of ​​the connected region , circularity , rectangularity and aspect ratio When the area is within the soft tissue area range and the circularity is less than the soft tissue circularity threshold Y, rectangularity is within the range of soft tissue rectangularity, and aspect ratio is within the range of soft tissue aspect ratio, and the connected area is judged to belong to the soft tissue area. This means that the area is consistent with the characteristics of soft tissue in all features and is judged to be a soft tissue area.

[0125] It should be noted that, assuming there is a connected area with a high circularity, which meets the circularity characteristics of the blood vessel area, that is, the circularity is greater than or equal to the blood vessel circularity threshold T. However, the rectangularity of this area is high, which exceeds the rectangularity range of blood vessels [ , but closer to the rectangular range of soft tissue. In this case, the type of the connected region cannot be determined based on a single feature. Only when the area, circularity, rectangularity and aspect ratio of the connected region simultaneously meet the characteristic range of the vascular region or the soft tissue region, a final judgment can be made.

[0126] When the area of ​​the connected region appears , circularity , rectangularity and aspect ratio When one or more of the items do not belong to the characteristic range of the soft tissue area or the vascular area at the same time, a secondary judgment is required based on the performance at different angles and positions. The judgment can be made manually or by a judgment program, and the specific situation is selected according to actual needs.

[0127] Specifically, the shape characteristics of the connected area can be further analyzed to observe its performance at different angles and positions. It can also be combined with the patient's medical history, clinical symptoms and other imaging examination results, and processed through professional doctor judgment and re-evaluation of judgment criteria to improve the accuracy of judgment of vascular areas and soft tissue areas in chest CT images.

[0128] Step Six: Extract the centerline of the blood vessels in the connected regions belonging to the blood vessel area and calculate the local blood vessel diameter Preset the blood vessel dilation threshold and the blood vessel constriction threshold, compare the local blood vessel diameter with the blood vessel dilation threshold to determine whether there is abnormal blood vessel dilation, and compare the local blood vessel diameter with the blood vessel constriction threshold to determine whether there is abnormal blood vessel constriction.

[0129] The specific steps of Step Six are as follows: The specific method for determining whether there is an abnormality in the blood vessels is as follows: For the connected regions belonging to the blood vessel area Extract the centerline of the blood vessels, and estimate the local blood vessel diameter by calculating the distance between adjacent points on the centerline .

[0130] Compare the local blood vessel diameter with the set blood vessel dilation threshold Te and blood vessel constriction threshold Tc respectively: When > Te, it is determined that the blood vessels are dilated; it indicates that there is abnormal blood vessel dilation, which may be caused by diseases such as hypertension and aneurysm.

[0131] When < Tc, it is determined that the blood vessels are constricted; it indicates that it may be caused by reasons such as blood vessel stenosis and thrombosis.

[0132] When Tc ≤ ≤ Te, it is determined to be normal; it indicates that the diameter of the blood vessels at this local position is within the normal range.

[0133] It should be noted that the specific method for estimating the local blood vessel diameter is to use a list of image coordinates to store the points on the centerline, start from the first point, and traverse the entire centerline list; for each point on the centerline, find the coordinates of its previous point and the next point, and use the Euclidean distance formula to calculate the distance between the current point and its previous point and the next point; create an array to store all the calculated distances between adjacent points; estimate the diameter of the entire blood vessel segment based on the average value of the calculated array of adjacent point distances.

[0134] Among them, for the first and the last points, since they have only one adjacent point, one can choose to only calculate the distance to one adjacent point, or consider them as boundary cases and handle them specially.

[0135] The setting of vasodilation threshold Te and vasoconstriction threshold Tc is based on multiple factors including clinical data, pathological research, imaging standards and expert experience. In practical applications, the setting of vasodilation threshold Te and vasoconstriction threshold Tc should take into account the differences among different populations, different parts of the body and different disease states. It should have a certain degree of flexibility and adjustability to adapt to different clinical scenarios and needs. It should be fully verified and evaluated to ensure its accuracy and reliability.

[0136] Step 7: Extract soft tissue features in the connected area belonging to the soft tissue area, and combine the extracted features into a first feature vector; preset a second feature vector, compare the first feature vector with the second feature vector, and determine whether there is an abnormality in the soft tissue.

[0137] The specific steps of step seven are: The specific method for determining whether there is abnormality in the soft tissue is: In the connected areas belonging to the soft tissue area Perform multiple feature extraction to obtain area , circularity , the contrast Con and energy Asm of the gray-level co-occurrence matrix, this connected area Area , circularity , the contrast Con and energy Asm of the gray-level co-occurrence matrix are combined into a first eigenvector f(x,y); the formula based on this is:

[0138] The second eigenvector F(x,y) is preset.

[0139] when When , it means that the area where the point is located is abnormal; it means that the characteristics of this area in terms of area, circularity, contrast and energy are quite different from those of the normal area, and there may be lesions or abnormalities.

[0140] when When , it means that the area where the point is located is normal, indicating that the characteristics of this area are close to those of normal soft tissue areas and there is no obvious abnormality.

[0141] It should be noted that the first eigenvector f(x, y) in the formula integrates multiple eigenvalues ​​into a mathematical representation to comprehensively judge the properties of the connected area; f(x, y) comprehensively reflects multiple characteristics of the soft tissue area, including size, shape regularity, and texture characteristics. For example, area Reflects the size of the soft tissue area; circularity It measures the circularity of the area and can determine the regularity of the area shape; contrast Describes the local variation of the grayscale value of pixels in the image, reflecting the clarity of the texture and the depth of the grooves; Energy It reflects the uniformity of image grayscale distribution and the regularity of texture.

[0142] Preset second eigenvector It is a specific value or range of values, which is determined based on the characteristic analysis of a large number of normal soft tissue areas and clinical experience. It represents the typical characteristics of normal soft tissue areas. For example, by analyzing the area of ​​a large number of known normal soft tissue areas , circularity , Contrast and energy The second eigenvector is calculated and statistically analyzed to determine a better distinction between normal and abnormal soft tissues. ,This eigenvector may be relatively fixed within a certain time and range, and can be used as a reference standard for judgment.

[0143] By comparing the corresponding elements of the first eigenvector f(x,y) with the second eigenvector The corresponding elements of are compared element by element. For example, if the first element of f(x,y) is the area value, The first element of is also the reference value of the corresponding area. Comparing the two values ​​can determine whether the area feature of the area is abnormal. , Contrast and energy Similar comparisons are made for the other elements.

[0144] when When , it means that each element of the first eigenvector is greater than the corresponding element of the second eigenvector, indicating that the characteristics of the region in terms of area, circularity, contrast and energy are quite different from those of the normal region, and there is a lesion or abnormality.

[0145] when , it means that each element of the first eigenvector is smaller than the corresponding element of the second eigenvector, and there is no abnormal performance.

[0146] When only one or more of the corresponding elements of f(x,y) are less than or equal to When the corresponding elements of the second eigenvector are equal, it means that the region as a whole has the characteristics of normal soft tissue and does not belong to the characteristics of abnormal soft tissue. For example, some elements may be slightly larger than the corresponding elements of the second eigenvector, but other elements are obviously smaller than or close to the corresponding values. On the whole, the region may still be judged as normal.

[0147] In the recognition of soft tissue abnormalities, the contrast Con and energy Asm of the gray level co-occurrence matrix are two important texture features; among them: The contrast Con of the grayscale co-occurrence matrix is ​​an indicator to measure the depth of the image texture grooves and the image clarity. In the grayscale co-occurrence matrix, the contrast reflects the total amount of local grayscale changes in the image. When the element values ​​in the matrix are distributed more dispersedly, that is, pixel pairs of different grayscale levels appear frequently, it means that the texture contrast in the image is higher and the image is clearer; conversely, when the element values ​​are distributed more concentratedly, that is, pixel pairs of the same or similar grayscale levels dominate, it means that the texture contrast in the image is lower and the image is more blurred. The calculation formula of the contrast Con of the grayscale co-occurrence matrix is:

[0148] in, It is the total number of gray levels, which determines the size of the gray-level co-occurrence matrix. It means that when calculating the energy, it participates in the calculation as a constant to ensure that images with different gray-level ranges can be evaluated relatively uniformly. is the gray-level co-occurrence matrix The value of an element, which indicates that the gray level is The pixel and gray level are The probability of pixels appearing at the same time reflects the different gray levels in the image; the frequency of the combination appears when calculating the contrast. It is one of the core factors. For different gray level combinations, the corresponding probability will affect the contrast value. If there are many pixel pairs with large gray level differences appearing simultaneously in the image, the corresponding Larger values ​​increase the contrast.

[0149] The square of the difference between two gray levels is calculated. Its function is to emphasize the difference between gray levels. When the difference between two gray levels is large, The value of will also be larger.

[0150] By traversing all possible grayscale combinations and calculating the sum of the products of their probabilities and the squares of the grayscale differences, the contrast Con of the grayscale co-occurrence matrix is ​​obtained, which comprehensively reflects the local changes in the grayscale values ​​of pixels in the image and provides an important indicator for image texture analysis.

[0151] The energy Asm of the gray-level co-occurrence matrix is ​​the sum of the squares of the element values ​​in the gray-level co-occurrence matrix, which reflects the uniformity and complexity of the image texture. When the element values ​​in the gray-level co-occurrence matrix are distributed more evenly, that is, the probability of occurrence of each pixel pair is similar, it means that the image texture is relatively simple and the energy value is low; on the contrary, when the element values ​​are distributed unevenly and there are obvious peaks, it means that the image texture is relatively complex and the energy value is high. The calculation formula of the energy Asm of the gray-level co-occurrence matrix is:

[0152] in, It is the total number of gray levels and determines the size range of the gray-level co-occurrence matrix. When calculating the energy, it is used as a constant to ensure that images with different gray-level ranges can be evaluated relatively uniformly. is the gray-level co-occurrence matrix The value of the element indicates that the gray level is The pixel and gray level are The probability of pixels appearing at the same time reflects the frequency of different grayscale combinations in the image. When calculating energy, The square of reflects the emphasis on the probability of occurrence of each gray level combination. If the probability of a gray level combination is higher, its contribution to the energy value will be greater.

[0153] The overall effect of this formula is that when the element values ​​in the gray-level co-occurrence matrix are distributed evenly, that is, the probability of occurrence of each pixel pair is similar, it means that the frequency of occurrence of pixel pairs of different gray levels in the image in a specific direction and distance is relatively consistent, and each The values ​​of are relatively small and close, and their sum of squares will also be small, so the energy value is low, indicating that the image texture is relatively simple; for example, in an image with very uniform texture, the probability of pixel pairs in various directions and distances appearing is relatively consistent, and the element values ​​in the gray-level co-occurrence matrix are not much different, so the energy value will be low.

[0154] When the element values ​​are unevenly distributed and there are obvious peaks, it means that the probability of certain grayscale combinations is much higher than other combinations. These high-probability pixel pairs will make a greater contribution to the energy calculation, making The sum of squares increases, resulting in a higher energy value, indicating that the image texture is more complex; for example, in an image with complex texture, there may be certain specific texture patterns, which lead to the frequent appearance of specific grayscale combinations, and the element values ​​in the grayscale co-occurrence matrix are unevenly distributed, resulting in a higher energy value.

[0155] Combined with the contents of steps 5 to 7: Through image segmentation, feature extraction and anomaly detection, in-depth analysis and accurate diagnosis of the soft tissue-vessel contrast target image are achieved. By calculating geometric features such as regional area, perimeter, center of mass, bounding box, and shape features such as circularity, rectangularity, and aspect ratio, combined with the texture features of the gray-level co-occurrence matrix, the health status of blood vessels and soft tissues can be comprehensively evaluated. At the same time, by setting reasonable thresholds, abnormal conditions such as vascular dilation and contraction, as well as abnormal changes in soft tissue areas, can be accurately identified, providing doctors with timely and accurate diagnostic information, which helps to detect and treat diseases at an early stage, and improve medical quality and patient survival rates.

[0156] Step 8: Automatically mark or highlight the identified abnormal situations.

[0157] The specific steps of step eight are: for the identified abnormal conditions, the identified vascular abnormal areas and soft tissue abnormal areas are highlighted with a color or brightness different from that of the surrounding tissues, a text label is added near the abnormal area, the type of abnormality is indicated, and at least one of a circle, a box, and an arrow graphic is drawn in the abnormal area to mark it.

[0158] It should be noted that highlighting, automatic annotation and graphic marking can be selected and combined according to different application scenarios and needs to maximize the doctor's or user's ability to recognize and understand abnormal situations.

[0159] Combined with the content of step 8: By automatically marking and highlighting abnormal areas, the speed and accuracy of doctors or users in identifying abnormal conditions are significantly improved. This intuitive marking method not only reduces the burden of manual inspection, but also makes the diagnosis process more efficient and accurate, helping doctors to quickly locate and evaluate lesions, thereby formulating more accurate treatment plans.

[0160] On the other hand, the present invention also discloses a chest CT image processing system, comprising: The image acquisition module is used to respectively obtain image data 1 and image data 2 of two groups of X-rays with different energies by using dual-energy CT scanning technology; and respectively obtain the first attenuation measurement value and the second attenuation measurement value corresponding to each pixel point under the two groups of X-rays with different energies by using image data 1 and image data 2.

[0161] The concentration calculation module is used to calculate the concentration value of the soft tissue at each pixel point according to the first attenuation measurement value corresponding to each pixel point. Calculate the concentration value of the blood vessel at each pixel point according to the second attenuation measurement value corresponding to each pixel point A soft tissue concentration map is generated based on the concentration values ​​corresponding to all soft tissues at each pixel point; and a blood vessel concentration map is generated based on the concentration values ​​corresponding to all blood vessels at each pixel point.

[0162] The image fusion module generates a soft tissue-vessel contrast initial image based on the soft tissue concentration map and the blood vessel concentration map, and calculates the initial image pixel value corresponding to each pixel point on the soft tissue-vessel contrast initial image according to the concentration value corresponding to the soft tissue at each pixel point and the concentration value corresponding to the blood vessel at each pixel point.

[0163] The contrast stretching module is used to perform contrast stretching on the initial image pixel value corresponding to each pixel point on the soft tissue-vessel contrast initial image, and calculate the target image pixel value corresponding to each pixel point in the soft tissue-vessel contrast target image.

[0164] The region distinction module is used to perform image segmentation on the target image pixel values ​​corresponding to each pixel point in the soft tissue-vessel contrast target image, obtain each pixel value in the binary image, and generate a binary image based on all the pixel values; by identifying and analyzing the connected areas in the binary image, the vascular area and the soft tissue area in the binary image are distinguished.

[0165] The vascular area abnormality identification module is used to extract the vascular centerline of the connected area belonging to the vascular area and calculate the local diameter of the blood vessel; preset the vascular expansion threshold and the vascular contraction threshold, compare the local diameter of the blood vessel with the vascular expansion threshold, and judge whether the blood vessel has abnormal expansion, and compare the local diameter of the blood vessel with the vascular contraction threshold to judge whether the blood vessel has abnormal contraction.

[0166] The soft tissue area abnormality recognition module extracts soft tissue features in the connected area belonging to the soft tissue area, combines the extracted features into a first feature vector; presets a second feature vector, compares the first feature vector with the second feature vector, and determines whether there is an abnormality in the soft tissue.

[0167] The result display module is used to automatically mark or highlight the identified abnormal situations.

[0168] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product. A person of ordinary skill in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented in electronic hardware or in combination with computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0169] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0170] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A chest CT image processing method, characterized in that: include: Step 1: using dual-energy CT scanning technology to respectively obtain two sets of image data 1 and image data 2 of X-rays with different energies; obtaining first attenuation measurement values ​​and second attenuation measurement values ​​corresponding to each pixel point under the two sets of X-rays with different energies through image data 1 and image data 2; Step 2: Calculate the concentration value of the soft tissue at each pixel point according to the first attenuation measurement value corresponding to each pixel point, calculate the concentration value of the blood vessel at each pixel point according to the second attenuation measurement value corresponding to each pixel point, generate a soft tissue concentration map based on the concentration values ​​corresponding to all soft tissues at each pixel point; generate a blood vessel concentration map based on the concentration values ​​of all blood vessels at each pixel point; Step 3: Generate a soft tissue-vessel contrast initial image based on the soft tissue concentration map and the blood vessel concentration map, and calculate the initial image pixel value corresponding to each pixel point on the soft tissue-vessel contrast initial image according to the concentration value corresponding to each pixel point of the soft tissue and the concentration value corresponding to each pixel point of the blood vessel; Step 4: performing contrast stretching on the initial image pixel value corresponding to each pixel point on the soft tissue-vessel contrast initial image, and calculating the target image pixel value corresponding to each pixel point in the soft tissue-vessel contrast target image; Step 5: Perform image segmentation on the target image pixel values ​​corresponding to each pixel point in the soft tissue-vessel contrast target image to obtain each pixel value in the binary image, and generate a binary image based on all the pixel values; distinguish the vascular area and the soft tissue area in the binary image by identifying and analyzing the connected areas in the binary image; Step 6: Extract the centerline of the blood vessel from the connected area belonging to the blood vessel area, and calculate the local diameter of the blood vessel; preset the blood vessel expansion threshold and the blood vessel contraction threshold, compare the local diameter of the blood vessel with the blood vessel expansion threshold, and judge whether the blood vessel has abnormal expansion; compare the local diameter of the blood vessel with the blood vessel contraction threshold, and judge whether the blood vessel has abnormal contraction; Step 7: extract soft tissue features in the connected area belonging to the soft tissue area, and combine the extracted features into a first feature vector; preset a second feature vector, compare the first feature vector with the second feature vector, and determine whether the soft tissue has abnormalities; Step 8: Automatically mark or highlight the identified abnormal situations.

2. A chest CT image processing method according to claim 1, characterized in that: The concentration values ​​of soft tissue and blood vessels at each pixel are calculated based on the following formula: in, Indicates energy In pixels The first attenuation measurement of Indicates energy In pixels The second attenuation measurement value, Is soft tissue A in energy The attenuation coefficient under , the value range is non-negative real number; It is the energy of blood vessel B The attenuation coefficient under , the value range is non-negative real number; Is soft tissue A in energy The attenuation coefficient under , the value range is non-negative real number; It is the energy of blood vessel B The attenuation coefficient under , the value range is non-negative real number; is soft tissue A at pixel point The corresponding concentration value is It is the blood vessel B at the pixel The corresponding concentration value is and The value range is non-negative real numbers, indicating the distribution of matter in the image.

3. A chest CT image processing method according to claim 2, characterized in that: The method of forming the soft tissue concentration map and the blood vessel concentration map is: Two two-dimensional arrays are created for soft tissue and blood vessels respectively, and their sizes are the same as image data one and image data two; the concentration value corresponding to the soft tissue at each pixel point is stored in the corresponding position in the two-dimensional array of the soft tissue to form a soft tissue concentration map; the concentration value corresponding to the blood vessel at each pixel point is stored in the corresponding position in the two-dimensional array of the blood vessel to form a blood vessel concentration map; each pixel value in each concentration map represents the concentration value of the soft tissue or blood vessel at the pixel point at the corresponding position.

4. A chest CT image processing method according to claim 3, characterized in that: The initial image pixel value corresponding to each pixel point on the soft tissue-vessel contrast initial image is calculated based on the formula: in, Represents the pixel point on the initial image of soft tissue-vessel contrast The corresponding initial image pixel value is yes The weight coefficient ranges from 0 to 1. is soft tissue A at pixel point The corresponding concentration value is It is the blood vessel B at the pixel The corresponding concentration value.

5. A chest CT image processing method according to claim 4, characterized in that: The target image pixel value corresponding to each pixel point in the soft tissue-vessel contrast target image is calculated based on the formula: in, Represents the pixel point in the soft tissue-vessel contrast target image The corresponding target image pixel value is is the minimum pixel value of the initial image of soft tissue-vessel contrast, is the maximum pixel value of the initial soft tissue-vessel contrast image, ranging from ; M is the upper limit of the target pixel value range after stretching, m is the lower limit of the target pixel value range, ; Used to prevent the denominator from being zero. The value range is .

6. A chest CT image processing method according to claim 5, characterized in that: Calculate the value of each pixel in a binary image The formula is: in, Represents the value of each pixel in the binary image. When the value is 1, it represents the foreground area, which represents the blood vessels and soft tissue area. When the value is 0, it represents the background area, which represents the removed area. It is the segmentation threshold, which is used to separate the foreground area from the background area.

7. A chest CT image processing method according to claim 6, characterized in that: Specific methods for distinguishing different vascular and soft tissue areas are: Identify and label each connected region in a binary image ; For each marked connected region , extract connected regions All pixel coordinates that satisfy =1; is the index of the region; By analyzing each connected region , obtain the dimension data, the dimension data at least includes the area , circularity , rectangularity And aspect ratio ; Set the blood vessel area range [ and soft tissue area , Vessel circularity threshold T and soft tissue circularity threshold Y, blood vessel rectangularity range [ and soft tissue rectangular range [ and the range of vessel aspect ratios[ and soft tissue aspect ratio range [ ; The area Respectively with the range of blood vessel area [ and soft tissue area[ Compare and determine the connected areas belongs to the vascular area or soft tissue area; The circularity The vessel circularity threshold T and soft tissue circularity threshold Y is compared to determine the connected area belongs to the vascular area or soft tissue area; The rectangular Respectively with the range of blood vessel rectangularity [ and soft tissue rectangular range [ Compare and determine the connected areas belongs to the vascular area or soft tissue area; The aspect ratio Respectively with the range of vessel length-width ratio [ and soft tissue aspect ratio range [ Compare and determine the connected areas belongs to the vascular area or soft tissue area; When the area of ​​the connected region , circularity , rectangularity and aspect ratio When all of them meet the characteristic range of the blood vessel region, it is determined that the connected region belongs to the blood vessel region; When the area of ​​the connected region , circularity , rectangularity and aspect ratio When all of them meet the characteristic range of the soft tissue area, it is determined that the connected area belongs to the soft tissue area; When the area of ​​the connected region appears , circularity , rectangularity and aspect ratio When one or several of the items do not belong to the characteristic range of the soft tissue area or the vascular area at the same time, a secondary judgment is required based on the performance of the connected area at different angles and positions.

8. A chest CT image processing method according to claim 7, characterized in that: The specific methods for determining whether there are abnormalities in blood vessels are: For connected regions belonging to vascular regions Extract the centerline of the blood vessel and estimate the local diameter of the blood vessel by calculating the distance between adjacent points on the centerline ; The local diameter of the blood vessel Compare with the set vasodilation threshold Te and vasoconstriction threshold Tc respectively: when >Te, it is judged as vasodilation; When <When Tc, it is determined as vasoconstriction; When Tc≤ When ≤Te, it is judged as normal; The specific methods for determining whether there is abnormality in soft tissue are: In the connected areas belonging to the soft tissue area Perform multiple feature extraction to obtain the area , circularity , the contrast Con and energy Asm of the gray-level co-occurrence matrix, this connected area Area , circularity , the contrast Con and energy Asm of the gray-level co-occurrence matrix are combined into a first eigenvector f(x,y); the formula based on this is: Preset the second eigenvector F(x,y); when When , it means that the area where the point is located is abnormal; when , it means the area where the point is located is normal.

9. A chest CT image processing method according to claim 8, characterized in that: The identified abnormal vascular areas and abnormal soft tissue areas are highlighted with a color or brightness different from that of the surrounding tissues, a text label is added near the abnormal area, the type of abnormality is indicated, and at least one of a circle, a box, and an arrow graphic is drawn in the abnormal area to mark it.

10. A chest CT image processing system, characterized in that: include: An image acquisition module is used to respectively acquire image data 1 and image data 2 of two groups of X-rays with different energies by using dual-energy CT scanning technology; and respectively acquire first attenuation measurement values ​​and second attenuation measurement values ​​corresponding to each pixel point under the two groups of X-rays with different energies by using image data 1 and image data 2; The concentration calculation module is used to calculate the concentration value of the soft tissue at each pixel point according to the first attenuation measurement value corresponding to each pixel point. Calculate the concentration value of the blood vessel at each pixel point according to the second attenuation measurement value corresponding to each pixel point Generate a soft tissue concentration map based on the concentration values ​​corresponding to all soft tissues at each pixel point; generate a blood vessel concentration map based on the concentration values ​​of all blood vessels at each pixel point; An image fusion module generates a soft tissue-vessel contrast initial image based on the soft tissue concentration map and the blood vessel concentration map, and calculates the initial image pixel value corresponding to each pixel point on the soft tissue-vessel contrast initial image according to the concentration value corresponding to each pixel point of the soft tissue and the concentration value corresponding to each pixel point of the blood vessel; The contrast stretching module is used to perform contrast stretching on the initial image pixel values ​​corresponding to each pixel point on the soft tissue-vessel contrast initial image, and calculate the target image pixel values ​​corresponding to each pixel point in the soft tissue-vessel contrast target image. The region distinction module is used to perform image segmentation on the target image pixel values ​​corresponding to each pixel point in the soft tissue-vessel contrast target image, obtain each pixel value in the binary image, and generate a binary image based on all the pixel values; by identifying and analyzing the connected regions in the binary image, the vascular region and the soft tissue region in the binary image are distinguished; The vascular region anomaly recognition module is used to extract the vascular centerline of the connected area belonging to the vascular region and calculate the local diameter of the vascular region. Preset a vascular expansion threshold and a vascular contraction threshold, compare the local diameter of the blood vessel with the vascular expansion threshold to determine whether the blood vessel is abnormally expanded, and compare the local diameter of the blood vessel with the vascular contraction threshold to determine whether the blood vessel is abnormally contracted; The soft tissue area abnormality recognition module extracts soft tissue features in the connected area belonging to the soft tissue area, and combines the extracted features into a first feature vector; a second feature vector is preset, and the first feature vector is compared with the second feature vector to determine whether the soft tissue has an abnormality; The result display module is used to automatically mark or highlight the identified abnormal situations.

Images