WO2023050691A1 - 图像处理方法及装置、电子设备、存储介质和程序 - Google Patents

图像处理方法及装置、电子设备、存储介质和程序 Download PDF

Info

Publication number
WO2023050691A1
WO2023050691A1 PCT/CN2022/077185 CN2022077185W WO2023050691A1 WO 2023050691 A1 WO2023050691 A1 WO 2023050691A1 CN 2022077185 W CN2022077185 W CN 2022077185W WO 2023050691 A1 WO2023050691 A1 WO 2023050691A1
Authority
WO
WIPO (PCT)
Prior art keywords
vertebra
image
rib
vertebral
segmented
Prior art date
Application number
PCT/CN2022/077185
Other languages
English (en)
French (fr)
Inventor
吴宇
赵亮
Original Assignee
上海商汤智能科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 上海商汤智能科技有限公司 filed Critical 上海商汤智能科技有限公司
Publication of WO2023050691A1 publication Critical patent/WO2023050691A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T9/00—Image coding
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10081—Computed x-ray tomography [CT]

Definitions

  • the present disclosure relates to the technical field of computers, and in particular to an image processing method and device, electronic equipment, storage media and programs.
  • the vertebrae are an important part of the human skeletal system. While maintaining and supporting body structures and organs, they also protect the central nervous system from damage caused by mechanical shocks. However, due to factors such as osteoporosis and external force, it is easy to cause vertebral fractures. In the diagnosis of vertebral fractures, doctors need to give the specific location of the vertebra where the fracture is located based on medical images of the chest (for example, chest CT images). In the related art, a neural network is used to directly segment vertebrae. However, due to the high similarity in shape between adjacent vertebrae, it is easy to cause mis-segmentation between adjacent vertebrae, resulting in low accuracy of vertebral segmentation.
  • the disclosure proposes a technical solution of an image processing method and device, electronic equipment, storage medium and program.
  • the present disclosure provides an image processing method, which is executed by an electronic device, and includes: performing image segmentation on an image of a vertebra to be segmented to obtain a binary segmentation result corresponding to the image of a vertebra to be segmented; Segment the vertebral image and perform feature encoding to obtain the pixel point embedding vector corresponding to each pixel point in the vertebral image to be segmented; determine the pixel point embedding vector corresponding to the pixel point to be divided according to the binary segmentation result A pixel embedding vector corresponding to the vertebra area in the vertebra image; performing clustering on the pixel embedding vector corresponding to the vertebra area to obtain a vertebra segmentation result corresponding to the vertebra image to be segmented.
  • an image processing device including:
  • An image segmentation module configured to perform image segmentation on the image of the vertebra to be segmented to obtain a binary segmentation result corresponding to the image of the vertebra to be segmented;
  • a feature encoding module configured to perform feature encoding on the image of the vertebra to be segmented to obtain a pixel embedding vector corresponding to each pixel in the image of the vertebra to be segmented;
  • the first determination module is configured to determine the pixel embedding vector corresponding to the vertebral region in the vertebra image to be segmented according to the binary segmentation result and the pixel embedding vector corresponding to each pixel;
  • the clustering module is configured to perform clustering on the pixel point embedding vectors corresponding to the vertebrae region to obtain the vertebral segmentation result corresponding to the vertebral image to be segmented.
  • the present disclosure provides an electronic device, including: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to call the instructions stored in the memory to perform the above method .
  • the present disclosure provides a computer-readable storage medium on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the above method is implemented.
  • an embodiment of the present disclosure provides a computer program, including computer readable code, when the computer readable code is run in an electronic device, a processor in the electronic device executes any one of the above an image processing method.
  • the image of the vertebra to be segmented is segmented to obtain the binary segmentation result corresponding to the image of the vertebra to be segmented; the image of the vertebra to be segmented is subjected to feature encoding to obtain the pixel point embedding vector corresponding to each pixel in the image of the vertebra to be segmented ; According to the binary segmentation result and the corresponding pixel point embedding vector of each pixel point, determine the pixel point embedding vector corresponding to the vertebral region in the vertebral image to be segmented; because the pixel point embedding vector has higher semantic expression ability, therefore, for the vertebral region
  • the corresponding pixel embedding vectors are clustered, which can reduce the probability of misclassification of adjacent vertebrae, obtain a vertebral segmentation result with high accuracy corresponding to the vertebral image to be segmented, and effectively improve the accuracy of vertebral segmentation.
  • FIG. 1 shows a flowchart of an image processing method according to an embodiment of the present disclosure
  • Fig. 2 shows a schematic diagram of a vertebral convex region according to an embodiment of the present disclosure
  • FIG. 3 shows a schematic diagram of a preset vertebral convex hull segmentation network according to an embodiment of the present disclosure
  • Fig. 4 shows a schematic diagram of a vertebral segmentation result of an embodiment of the present disclosure
  • FIG. 5 shows a schematic diagram of a preset vertebral instance segmentation network according to an embodiment of the present disclosure
  • Figure 6 shows a schematic diagram of rib endpoints and vertebral centroids of an embodiment of the present disclosure
  • Fig. 7 shows a schematic diagram of a target rib-vertebrae matching result of an embodiment of the present disclosure
  • Fig. 8 shows a schematic diagram of vertebral labeling results of an embodiment of the present disclosure
  • FIG. 9 shows a block diagram of an image processing device according to an embodiment of the present disclosure.
  • FIG. 10 shows a block diagram of an electronic device according to an embodiment of the present disclosure
  • Fig. 11 shows a block diagram of an electronic device according to an embodiment of the present disclosure.
  • the image processing method provided by the embodiments of the present disclosure may be executed by electronic devices such as terminal devices or servers, and the terminal devices may be user equipment (User Equipment, UE), mobile devices, user terminals, cellular phones, cordless phones, personal digital assistants ( Personal Digital Assistant, PDA), handheld devices, computing devices, vehicle-mounted devices, wearable devices, etc.
  • the image processing method can be realized by calling the computer-readable instructions stored in the memory by the processor.
  • the image processing method may be performed by a server.
  • the image processing method provided by the embodiments of the present disclosure can accurately locate the anatomical label of the vertebra where the fracture or lesion is located, and provide the 3D reconstruction visualization result of the spine and each vertebra. It can provide technical support for vertebral fracture detection, quantitative analysis of vertebral parameters, vertebral disease screening, lesion detection and other projects, thereby reducing the workload and error rate of doctors and shortening the diagnosis cycle.
  • Fig. 1 shows a flowchart of an image processing method according to an embodiment of the present disclosure.
  • the image processing method provided by the embodiment of the present disclosure the method is executed by an electronic device, may include the following steps:
  • step S11 image segmentation is performed on the image of the vertebra to be segmented to obtain a binary segmentation result corresponding to the image of the vertebra to be segmented.
  • step S12 feature encoding is performed on the image of the vertebra to be segmented to obtain a pixel embedding vector corresponding to each pixel in the image of the vertebra to be segmented.
  • step S13 according to the binary segmentation result and the pixel embedding vector corresponding to each pixel, the pixel embedding vector corresponding to the vertebra region in the vertebra image to be segmented is determined.
  • step S14 the pixel point embedding vectors corresponding to the vertebral region are clustered to obtain a vertebral segmentation result corresponding to the vertebral image to be segmented.
  • the image of the vertebra to be segmented is segmented to obtain the binary segmentation result corresponding to the image of the vertebra to be segmented; the image of the vertebra to be segmented is subjected to feature encoding to obtain the pixel point embedding vector corresponding to each pixel in the image of the vertebra to be segmented ; According to the binary segmentation result and the corresponding pixel point embedding vector of each pixel point, determine the pixel point embedding vector corresponding to the vertebral region in the vertebral image to be segmented; because the pixel point embedding vector has higher semantic expression ability, therefore, for the vertebral region
  • the corresponding pixel embedding vectors are clustered, which can reduce the probability of misclassification of adjacent vertebrae, obtain a vertebral segmentation result with high accuracy corresponding to the vertebral image to be segmented, and effectively improve the accuracy of vertebral segmentation.
  • the image processing method further includes the following steps: acquiring an original chest scan image; performing image preprocessing on the original chest scan image to obtain an initial vertebral image; performing convex hull segmentation on the initial vertebral image to obtain an initial The vertebral convex region in the vertebral image; according to the vertebral convex region, the initial vertebral image is cut to obtain the vertebral image to be segmented.
  • the convex hull segmentation of the initial vertebral image can be performed to determine the vertebral convex region in the initial vertebral image, so that according to the vertebral convex region, from The vertebral image to be segmented is obtained by cropping the initial vertebral image, so that subsequent vertebral segmentation can be performed based on the cropped vertebral image to be segmented, which can effectively reduce the consumption of computing resources and improve the efficiency of vertebral segmentation.
  • the original chest scan image may be a chest computed tomography (CT) image I.
  • CT computed tomography
  • chest CT images I are usually used as medical images for the diagnosis of vertebral fractures.
  • image preprocessing is performed on the chest CT image I to obtain an initial vertebral image I n .
  • Image preprocessing may include: one or more of reorientation, cropping, normalization, etc., and the embodiment of the present disclosure does not limit the specific manner of image preprocessing.
  • the chest CT image I is pre-processed according to the preset identity matrix to obtain the initial vertebral image I n , so that the initial vertebral image I n
  • the direction of the vertebrae in is consistent with the preset coordinate axes (x/y/z axes) to improve the efficiency of subsequent processing.
  • the preset unit matrix can be set according to actual conditions, and the embodiment of the present disclosure does not limit the specific form of the preset unit matrix.
  • the chest CT image I in addition to the bone region, also includes large areas of other background parts.
  • the chest CT image I is cropped using a preset grayscale threshold . Specifically, it may include: performing binarization processing on the chest CT image I based on a preset gray threshold to obtain a binary image. Wherein, the pixel in the chest CT image I whose gray value is greater than or equal to the preset gray threshold has a corresponding pixel value of 1 in the binary image; the pixel in the chest CT image I whose gray value is less than the preset gray threshold point, the corresponding pixel value in the binary image is 0.
  • the chest CT image I is cropped to obtain the initial vertebral image I n to reduce the image size, thereby effectively reducing subsequent computing resource consumption and improving processing efficiency .
  • the specific value of the preset grayscale threshold can be set according to actual conditions, which is not specifically limited in this embodiment of the present disclosure.
  • the preset gray value normalization window is used for normalization processing, so that the gray value of the initial vertebral image I n is within a reasonable gray value Within the range, the accuracy of vertebral segmentation is improved.
  • the preset gray value normalization window can be set according to the actual situation.
  • the preset gray value normalization window is [-1000, 2000].
  • the preset gray value normalization The actual value range of the normalization window is not specifically limited.
  • the initial vertebral image I since the scanning area of the chest CT image I is very large, the initial vertebral image I includes not only the vertebrae, but also other bones such as ribs, hip bones, femurs, etc., and the vertebral area only accounts for a small part of it. Therefore, it can be based on Convex hull algorithm, segmenting the initial vertebral image I n to obtain the vertebral convex hull area in the initial vertebral image I n , and then cutting the initial vertebral image I n based on the vertebral convex hull area to obtain the vertebra to be divided including the vertebral area Image IV .
  • the subsequent vertebral segmentation based on the vertebral image Iv to be segmented can reduce the waste of computing resources and improve the efficiency of vertebral segmentation.
  • the initial vertebral image I n can be segmented based on a preset vertebral convex hull segmentation network to obtain the vertebral convex hull region in the initial vertebral image I n .
  • the initial vertebral image I n may be resampled to obtain the first vertebral image I sp3 , where the resolution of the first vertebral image I sp3 is the first resolution.
  • a specific value of the first resolution may be determined according to an actual situation, which is not specifically limited in this embodiment of the present disclosure.
  • the first resolution is 3mm ⁇ 3mm ⁇ 3mm, that is, the actual physical size corresponding to each pixel in the first vertebra image I sp3 is 3mm ⁇ 3mm ⁇ 3mm.
  • FIG. 2 shows a schematic view of a vertebral convex region according to an embodiment of the disclosure.
  • a detection frame including the entire area of the vertebra can be determined.
  • the detection frame may be the smallest three-dimensional rectangular frame including the convex hull region H of the vertebrae.
  • the initial vertebral image In is cropped to obtain the vertebral image Iv to be segmented including the whole area of the vertebrae. Based on the image I v of the vertebra to be segmented, subsequent vertebrae segmentation is performed.
  • the preset vertebral convex hull segmentation network can be a 3D-U network, including an encoder consisting of several convolutional layers and downsampling layers, and an encoder consisting of several convolutional layers and upsampling layers.
  • FIG. 3 shows a schematic diagram of a preset vertebral convex hull segmentation network according to an embodiment of the present disclosure.
  • the learning rate setting strategy of warmup and cosine annealing can be used, and cross entropy (cross entropy loss) and dice loss can be used as the loss function to perform several operations on the preset vertebral convex hull segmentation network. Training rounds of training.
  • the training of the preset vertebral convex hull segmentation network can be realized by minimizing the loss function L 1 shown in the following formula (1).
  • y i in formula (1) is the segmentation label corresponding to the training sample image i, is the segmentation prediction probability corresponding to the training sample image i determined according to the preset vertebral convex hull segmentation network, Y is the real segmentation result corresponding to the training sample image i, is the predicted segmentation result corresponding to the training sample image i determined according to the preset vertebral convex hull segmentation network.
  • the preset training samples of the vertebral convex hull segmentation network include vertebral sample images and vertebral convex hull area labels corresponding to the vertebral sample images.
  • the vertebra convex hull region label may be the vertebra binary segmentation label obtained by performing binary segmentation on the vertebra sample image, and obtained after several times of Gaussian smoothing threshold expansion.
  • the label of the vertebral convex region can also be determined by other means in the related art, which is not specifically limited in this embodiment of the present disclosure.
  • a convex hull algorithm may be used to perform convex hull segmentation on the initial vertebra image I n to obtain the vertebra convex hull area in the initial vertebra image I n .
  • the specific algorithm form of the convex hull algorithm can be flexibly set according to the actual situation, which is not specifically limited in the embodiments of the present disclosure.
  • the image processing method further includes: performing position encoding on the pixels in the image of the vertebra to be segmented to obtain a position-encoded image.
  • the following formula (2) can be used to perform position coding on the image I v of the vertebra to be segmented to obtain the position coding image I c .
  • (i, j, k) are the corresponding pixels in the vertebral image I v to be segmented and the position coding image I c , ( ⁇ x , ⁇ y , ⁇ z ) are the image center pixels of the vertebral image I v to be segmented , W x , W y and W z are preset hyperparameters. Specific values of the preset hyperparameters W x , W y and W z may be determined according to actual conditions, and are not specifically limited in this embodiment of the present disclosure.
  • the image I v of the vertebra to be segmented may also be resampled to obtain the second image I sp1.5 of the vertebra, wherein the resolution of the second image I sp1.5 of the vertebra is the second resolution.
  • a specific value of the second resolution may be determined according to an actual situation, which is not specifically limited in this embodiment of the present disclosure.
  • the second resolution is 1.5 mm ⁇ 1.5 mm ⁇ 1.5 mm, that is, the actual physical size corresponding to each pixel in the second vertebra image I sp1.5 is 1.5 mm ⁇ 1.5 mm ⁇ 1.5 mm.
  • the above formula (1) is used to perform position coding on the second vertebral image I sp1.5 to obtain a position coding image I c .
  • the vertebrae segmentation of the vertebral image I v to be segmented is realized by using the second vertebral image Isp1.5 and the position coded image Ic .
  • performing image segmentation on the image of the vertebra to be segmented to obtain a binary segmentation result corresponding to the image of the vertebra to be segmented includes: performing image segmentation on the image of the vertebra to be segmented based on the position coded image to obtain a binary segmentation result.
  • feature coding is performed on the image of the vertebra to be segmented to obtain the pixel embedding vector corresponding to each pixel in the image of the vertebra to be segmented, including: based on the position coded image, performing feature coding on the image of the vertebra to be segmented to obtain The pixel point embedding vector corresponding to each pixel point in the segmented vertebral image.
  • the image processing method further includes: filtering the non-vertebra region in the binary segmentation result according to the convex hull region of the vertebra.
  • instance segmentation can be performed on the vertebral image I v to be segmented based on a preset vertebral instance segmentation network. Input the vertebral image I v to be segmented (or the second vertebral image I sp1.5 ) and the position-coded image I c into the preset vertebral instance segmentation network at the same time.
  • the preset vertebral instance segmentation network includes two branches, one branch is used for binary segmentation of the vertebral image I v to be segmented (or the second vertebral image I sp1.5 ), to obtain the vertebral image I v to be segmented (or The binary segmentation result A b corresponding to the second vertebra image I sp1.5 ).
  • the vertebral convex hull region H is used to filter the binary segmentation result A b to eliminate the false positive part (non-vertebral region) that is mistakenly segmented into the vertebral region in the binary segmentation result A b , and improve the binary segmentation result Accuracy of A b .
  • the binary segmentation result A b A b ⁇ H.
  • another branch of the preset vertebral instance segmentation network is used to determine the pixel embedding corresponding to each pixel in the vertebral image Iv to be segmented (or the second vertebral image I sp1.5 ) based on the position coded image Ic Vector A e .
  • the dimension of the embedding vector may be 8 dimensions, or may be set to other dimensions according to actual conditions, which is not specifically limited in this embodiment of the present disclosure.
  • the mean-shift clustering algorithm is used to cluster the pixel embedding vector A re corresponding to the vertebral region, and the vertebral segmentation result A ins corresponding to the vertebral image I v to be segmented is obtained.
  • FIG. 4 shows a schematic diagram of vertebral segmentation results according to an embodiment of the present disclosure.
  • a ins in the vertebral segmentation result includes multiple vertebrae in the vertebral image Iv to be segmented.
  • the vertebral segmentation result A ins cannot determine the vertebral identity of each vertebrae.
  • the preset vertebral instance segmentation network may be a 3D-U network, two encoder branches share a decoder, and a non-local module is embedded between the encoder and decoder.
  • one encoder branch is used to perform binary segmentation on the vertebral image I v to be segmented (or the second vertebral image I sp1.5 ) to obtain the binary segmentation result A b
  • the other encoder branch is used to determine the vertebral image to be segmented
  • the pixel embedding vector A e corresponding to each pixel in I v (or the second vertebra image I sp1.5 ).
  • FIG. 5 shows a schematic diagram of a preset vertebral instance segmentation network according to an embodiment of the present disclosure.
  • the encoder branch for binary segmentation can be trained for several training rounds using cross entropy and dice loss as loss functions.
  • the specific training formula can refer to the above formula (1).
  • the encoder branch for embedding vector prediction can be trained for several training epochs with discriminative loss as the loss function.
  • the training of the encoder branch for embedding vector prediction can be realized by minimizing the loss function L d shown in the following formula (3).
  • C in formula (3) is the total number of categories in the training sample image
  • N c is the number of pixels belonging to the same category in the training sample image
  • ⁇ c is the mean vector of the category
  • ⁇ c is the pixel point in the training sample image Embedding vector for i.
  • vertebral identification anatomical labels of vertebrae, for example, No. 3 vertebra
  • doctors usually need to count the vertebrae from top to bottom (or from bottom to top) to determine the vertebral identification, which leads to a complicated vertebral identification process and low accuracy. Therefore, after the vertebral segmentation results are obtained based on the methods of the above embodiments, it is also necessary to provide a method capable of automatically marking vertebrae to improve the efficiency and accuracy of vertebral marking.
  • the vertebral segmentation result includes multiple vertebrae in the vertebral image to be segmented; the image processing method further includes: obtaining the rib labeling result corresponding to the vertebral region in the vertebral image to be segmented, wherein the rib labeling
  • the result includes a plurality of ribs and a rib identification corresponding to each rib; according to the plurality of vertebrae and the plurality of ribs, a target rib-vertebra matching result is determined, wherein the target rib-vertebra matching result includes at least one rib-vertebra matching pair; According to the rib identification corresponding to each rib and the target rib-vertebra matching result, the vertebra identification corresponding to each vertebra is determined.
  • the multiple vertebrae in the vertebral segmentation result are matched with the multiple ribs in the rib marking result to obtain the target rib-vertebrae matching result, and then the rib corresponding to each rib in the rib marking result is used Rib identification, to determine the corresponding vertebral identification of each vertebra, so as to effectively obtain the result of vertebral identification with high accuracy.
  • the rib marking result is obtained after segmenting and marking the rib image to be segmented, the rib image to be segmented and the vertebral image to be segmented correspond to the same target object, for example, the rib image to be segmented and the vertebral image to be segmented are both derived from the same Chest CT image obtained after chest computed tomography scan of target subject.
  • the rib image to be segmented is segmented and marked, and the rib marking result corresponding to the rib image to be segmented is determined, including: performing semantic category segmentation on the rib image to be segmented to obtain the semantic category segmentation result corresponding to the rib image to be segmented; the rib image to be segmented Carry out instance segmentation to obtain the instance segmentation result corresponding to the rib image to be segmented; determine the rib labeling result corresponding to the rib image to be segmented according to the semantic category segmentation result and the instance segmentation result, wherein the rib labeling result includes multiple ribs in the rib image to be segmented The root rib and the corresponding rib ID for each rib.
  • the rib image to be segmented is segmented and marked, and the rib marking result corresponding to the rib image to be segmented is determined, which further includes: when the semantic category segmentation and instance segmentation are performed at the first resolution, at the second resolution Next, binary segmentation is performed on the rib image to be segmented to obtain a fine binary segmentation result corresponding to the rib image to be segmented, wherein the second resolution is greater than the first resolution; and the rib marking result is updated according to the fine binary segmentation result.
  • the segmentation result of each rib in the rib labeling result is relatively rough. Therefore, based on the higher resolution fine binary segmentation result , update the rib marking result, and obtain the rib marking result with higher resolution and higher segmentation accuracy.
  • any manner of rib segmentation and marking in the related art may be used for the manner of segmenting and marking the rib image to be segmented, which is not specifically limited in this embodiment of the present disclosure.
  • determining the target rib-vertebrae matching result according to multiple vertebrae and multiple ribs includes: for any one vertebra, determining the candidate rib-vertebrae matching result corresponding to the vertebrae, wherein the candidate rib- The vertebra matching result includes at least one rib-vertebra matching pair; according to the spatial geometric relationship of at least one rib-vertebra matching pair included in each candidate rib-vertebra matching result, determine the matching degree corresponding to each candidate rib-vertebra matching result ; Determining the candidate rib-vertebra matching result whose matching degree meets the preset matching condition as the target rib-vertebra matching result.
  • the matching degree corresponding to each candidate rib-vertebra matching result is determined according to the spatial geometric relationship of at least one rib-vertebra matching pair included in each candidate rib-vertebra matching result, including: According to the result of rib marking, determine the rib endpoint closest to the vertebra in each rib; according to the vertebral segmentation result, determine the centroid of each vertebra corresponding to each vertebra, and the tangent vector corresponding to each vertebra centroid; according to the rib endpoint corresponding to each rib, The centroid of each vertebra corresponding to the vertebra and the tangent vector corresponding to each vertebra centroid determine the mean cosine distance corresponding to each candidate rib-vertebra matching result; the mean cosine distance corresponding to each candidate rib-vertebra matching result is determined as The matching degree corresponding to each candidate rib-vertebrae matching result.
  • the vertebral segmentation result includes a vertebral set ⁇ V i
  • Determine the tangent vector ⁇ of each vertebral centroid on the central line of the spine, and obtain the tangent vector set T ( ⁇ 1 , ⁇ 2 ,..., ⁇ N ).
  • 6 shows a schematic diagram of rib endpoints and vertebral centroids, according to an embodiment of the disclosure.
  • all possible rib-vertebrae matching methods are enumerated, and the candidate rib-vertebrae matching results corresponding to each vertebra are obtained
  • the vertebra V sv assuming that the vertebra V sv and the rib R sr are a rib-vertebra matching pair, then according to the sorting in the rib sequence R and the vertebra sequence V, the vertebra V sv+1 and the rib R sr+1 are A rib-vertebrae pair.
  • the candidate rib-vertebra matching results corresponding to the vertebra V sv are obtained Traverse each vertebra in the vertebrae set to get a set of candidate rib-vertebrae matching results
  • p is the candidate rib-vertebrae matching result
  • the number of rib-vertebra matching pairs included in (R i , V j ) is the candidate rib-vertebra matching result
  • a rib-vertebra matching pair in , e i is the rib endpoint corresponding to rib R i
  • m j is the vertebra centroid corresponding to vertebra V j
  • ⁇ j is the tangent vector corresponding to vertebra centroid m j .
  • the mean cosine distance corresponding to each candidate rib-vertebrae matching result is determined as its corresponding matching degree, and then, the candidate rib-vertebrae matching result whose matching degree meets the preset matching conditions is determined as the target rib-vertebrae matching result.
  • the vertebra matching result that is, the candidate rib-vertebra matching result whose mean cosine distance meets the preset matching condition is determined as the target rib-vertebra matching result.
  • the candidate rib-vertebrae matching result whose matching degree meets the preset matching condition is determined as the target rib-vertebrae matching result, including: according to the mean cosine distance corresponding to each candidate rib-vertebrae matching result , determine whether there is a target rib-vertebrae matching result set, wherein the mean cosine distance corresponding to each candidate rib-vertebrae matching result set in the target rib-vertebrae matching result set is greater than 0; if there is a target rib-vertebrae matching result set In this case, the candidate rib-vertebra matching result corresponding to the smallest cosine distance mean in the target rib-vertebra matching result set is determined as the target rib-vertebra matching result.
  • the image processing method further includes: in the absence of a target rib-vertebrae matching result set, determining the candidate rib-vertebrae matching result corresponding to the maximum cosine distance mean value as the target Rib-vertebrae matching results.
  • the set of candidate rib-vertebrae matching results ⁇ is divided into a positive-cosine matching set ⁇ + and a negative cosine matching set ⁇ ⁇ .
  • the mean cosine distances corresponding to the candidate rib-vertebra matching results included in the sine-cosine matching set ⁇ + are all greater than 0, and the mean cosine distances corresponding to the candidate rib-vertebra matching results included in the negative cosine matching set ⁇ - are all less than or equal to 0.
  • the sine-cosine matching set ⁇ + is a non-empty set, that is, there is a target rib-vertebra matching result set ⁇ + .
  • the mean cosine distance corresponding to the vertebra matching result is greater than 0, and the smaller the value of the cosine distance mean of the candidate rib-vertebra matching result, it can represent the rib-vertebra matching rib-vertebra included in the candidate rib-vertebra matching result. The closer the distance, the higher the matching degree. Therefore, the candidate rib-vertebrae matching result corresponding to the minimum cosine distance mean in the target rib-vertebrae matching result set ⁇ + can be determined as the target rib-vertebrae matching result
  • the negative cosine matching set ⁇ + is an empty set, that is, there is no target rib-vertebrae matching result set ⁇ + , only the negative cosine matching set ⁇ - exists, because the negative cosine matching set ⁇ - includes The mean cosine distance corresponding to the candidate rib-vertebra matching result is less than or equal to 0, and the larger the value of the cosine distance mean value of the candidate rib-vertebra matching result, it can represent the rib-vertebra matching pair included in the candidate rib-vertebra matching result The closer the rib-vertebra distance, the higher the matching degree. Therefore, the candidate rib-vertebra matching result corresponding to the largest cosine distance mean in the negative cosine matching set ⁇ - can be determined as the target rib-vertebra matching result
  • the target rib-vertebrae matching result can be determined according to the following formula (5):
  • FIG. 7 shows a schematic diagram of target rib-vertebrae matching results according to an embodiment of the present disclosure. As shown in Figure 7, the distance between the rib endpoint and the centroid of the vertebra in each rib-vertebra matching pair in the target rib-vertebra matching result is the closest.
  • At least one rib-vertebrae matching pair is included in the target rib-vertebrae matching result, according to the rib identification of the ribs in each rib-vertebrae matching pair, and the rib-vertebrae in anatomical
  • determine the vertebral identity of each rib-vertebrae matching pair and obtain the vertebral identity of at least one vertebra, and then get the vertebral identity of each vertebra according to the sorting of multiple vertebrae in the vertebral set, and get The vertebral labeling result with high accuracy can effectively realize the vertebral labeling of the vertebral image to be segmented and improve the accuracy of vertebral labeling.
  • FIG. 8 shows a schematic diagram of vertebral labeling results according to an embodiment of the present disclosure.
  • the vertebral labeling result includes multiple vertebrae in the vertebral image I v to be segmented and the vertebral identification corresponding to each vertebrae.
  • each vertebra and the vertebra identification corresponding to each vertebra may be indicated by different colors. For example, gray is used to indicate vertebra number 1, green is used to indicate vertebra number 2, and so on.
  • vertebral labeling result may indicate each vertebra and the vertebral identifier corresponding to each vertebra, which is not specifically limited in this embodiment of the present disclosure.
  • embodiments of the present disclosure also provide image processing devices, electronic equipment, computer-readable storage media, and programs, all of which can be used to implement any image processing method provided by the embodiments of the present disclosure, corresponding technical solutions and descriptions, and refer to methods Part of the corresponding record.
  • FIG. 9 shows a block diagram of an image processing device according to an embodiment of the present disclosure. As shown in Figure 9, the device 90 includes:
  • the image segmentation module 91 is configured to perform image segmentation on the image of the vertebra to be segmented to obtain a binary segmentation result corresponding to the image of the vertebra to be segmented;
  • the feature encoding module 92 is configured to perform feature encoding on the vertebral image to be segmented to obtain a pixel point embedding vector corresponding to each pixel in the vertebral image to be segmented;
  • the first determination module 93 is configured to determine the pixel embedding vector corresponding to the vertebral region in the vertebra image to be segmented according to the binary segmentation result and the pixel embedding vector corresponding to each pixel;
  • the clustering module 94 is configured to perform clustering on the pixel point embedding vectors corresponding to the vertebral region to obtain a vertebral segmentation result corresponding to the vertebral image to be segmented.
  • the device 90 also includes:
  • a first acquisition module configured to acquire an original chest scan image
  • An image preprocessing module configured to perform image preprocessing on the original chest scan image to obtain an initial vertebral image
  • the convex hull segmentation module is configured to perform convex hull segmentation on the initial vertebra image to obtain the vertebra convex hull area in the initial vertebra image;
  • the clipping module is configured to clip the initial vertebra image according to the convex hull area of the vertebra to obtain the vertebra image to be segmented.
  • the device 90 also includes:
  • the position encoding module is configured to perform position encoding on the pixels in the vertebra image to be segmented to obtain a position encoded image.
  • the image segmentation module 91 is configured as:
  • image segmentation is performed on the vertebral image to be segmented, and a binary segmentation result is obtained.
  • the feature encoding module 92 is configured as:
  • feature encoding is performed on the image of the vertebra to be segmented, and the pixel embedding vector corresponding to each pixel is obtained.
  • the device 90 also includes:
  • the filtering module is configured to filter the non-vertebra region in the binary segmentation result according to the convex hull region of the vertebra.
  • the vertebral segmentation result includes multiple vertebrae in the vertebral image to be segmented
  • the device 70 also includes:
  • the second acquisition module is configured to acquire rib marking results corresponding to the vertebral region in the vertebra image to be segmented, wherein the rib marking results include multiple ribs and a rib identification corresponding to each rib;
  • the second determination module is configured to determine a target rib-vertebrae matching result according to multiple vertebrae and multiple ribs, wherein the target rib-vertebrae matching result includes at least one rib-vertebrae matching pair;
  • the third determining module is configured to determine the vertebral identifier corresponding to each vertebra according to the rib identifier corresponding to each rib and the target rib-vertebrae matching result.
  • the second determining module includes:
  • the first determining submodule is configured to, for any vertebra, determine a candidate rib-vertebra matching result corresponding to the vertebra, wherein the candidate rib-vertebra matching result includes at least one rib-vertebra matching pair;
  • the second determining submodule is configured to determine the matching degree corresponding to each candidate rib-vertebra matching result according to the spatial geometric relationship of at least one rib-vertebra matching pair included in each candidate rib-vertebra matching result;
  • the third determination sub-module is configured to determine the candidate rib-vertebrae matching result whose matching degree meets the preset matching condition as the target rib-vertebrae matching result.
  • the second determining submodule includes:
  • the first determining unit is configured to determine the rib endpoint closest to the vertebra in each rib according to the rib marking result
  • the second determination unit is configured to determine the vertebral centroid corresponding to each vertebra and the tangent vector corresponding to each vertebral centroid according to the vertebral segmentation result;
  • the third determining unit is configured to determine the mean cosine distance corresponding to each candidate rib-vertebra matching result according to the rib endpoint corresponding to each rib, the vertebra centroid corresponding to each vertebra, and the tangent vector corresponding to each vertebra centroid;
  • the fourth determination unit is configured to determine the mean cosine distance corresponding to each candidate rib-vertebra matching result as the matching degree corresponding to each candidate rib-vertebra matching result.
  • the third determination submodule includes:
  • the fifth determination unit is configured to determine whether there is a target rib-vertebra matching result set according to the mean cosine distance corresponding to each candidate rib-vertebra matching result, wherein each candidate rib-vertebrae in the target rib-vertebra matching result set The average cosine distance corresponding to the matching results is greater than 0;
  • the sixth determination unit is configured to determine the candidate rib-vertebra matching result corresponding to the mean cosine distance with the smallest value in the target rib-vertebra matching result set as the target rib-vertebra matching result set if there is a target rib-vertebra matching result set. Vertebral matching results.
  • the device 90 also includes:
  • the fourth determining module is configured to determine the candidate rib-vertebra matching result corresponding to the maximum cosine distance mean value as the target rib-vertebra matching result when there is no target rib-vertebra matching result set.
  • the functions or modules included in the apparatus provided by the embodiments of the present disclosure may be used to execute the methods described in the above method embodiments, and for specific implementation, refer to the descriptions of the above method embodiments.
  • Embodiments of the present disclosure also provide a computer-readable storage medium, on which computer program instructions are stored, and the above-mentioned method is implemented when the computer program instructions are executed by a processor.
  • Computer readable storage media may be volatile or nonvolatile computer readable storage media.
  • An embodiment of the present disclosure also proposes an electronic device, including: a processor; a memory for storing instructions executable by the processor; wherein the processor is configured to invoke the instructions stored in the memory to execute the above method.
  • An embodiment of the present disclosure also provides a computer program product, including computer-readable codes, or a non-volatile computer-readable storage medium carrying computer-readable codes, when the computer-readable codes are stored in a processor of an electronic device When running in the electronic device, the processor in the electronic device executes the above method.
  • Electronic devices may be provided as terminals, servers, or other forms of devices.
  • Fig. 10 shows a block diagram of an electronic device according to an embodiment of the present disclosure.
  • the electronic device 800 may be a terminal such as a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, or a personal digital assistant.
  • electronic device 1000 may include one or more of the following components: processing component 1002, memory 1004, power supply component 1006, multimedia component 1008, audio component 1010, input/output (I/O) interface 1012, sensor component 1014 , and the communication component 1016.
  • the processing component 1002 generally controls the overall operations of the electronic device 1000, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations.
  • the processing component 1002 may include one or more processors 1020 to execute instructions to complete all or part of the steps of the above method. Additionally, processing component 1002 may include one or more modules that facilitate interaction between processing component 1002 and other components. For example, processing component 1002 may include a multimedia module to facilitate interaction between multimedia component 1008 and processing component 1002 .
  • the memory 1004 is configured to store various types of data to support operations at the electronic device 1000 . Examples of such data include instructions for any application or method operating on the electronic device 1000, contact data, phonebook data, messages, pictures, videos, and the like.
  • the memory 1004 can be implemented by any type of volatile or non-volatile memory device or their combination, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable Programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), Magnetic Memory, Flash Memory, Magnetic or Optical Disk.
  • SRAM static random access memory
  • EEPROM electrically erasable programmable read-only memory
  • EPROM erasable Programmable Read Only Memory
  • PROM Programmable Read Only Memory
  • ROM Read Only Memory
  • Magnetic Memory Flash Memory
  • Magnetic or Optical Disk Magnetic Disk
  • the power supply component 1006 provides power to various components of the electronic device 1000 .
  • Power supply components 1006 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for electronic device 1000 .
  • the multimedia component 1008 includes a screen providing an output interface between the electronic device 1000 and the user.
  • the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user.
  • the touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor may not only sense a boundary of a touch or swipe action, but also detect duration and pressure associated with the touch or swipe action.
  • the multimedia component 1008 includes a front camera and/or a rear camera. When the electronic device 1000 is in an operation mode, such as a shooting mode or a video mode, the front camera and/or the rear camera can receive external multimedia data. Each front camera and rear camera can be a fixed optical lens system or have focal length and optical zoom capability.
  • the audio component 1010 is configured to output and/or input audio signals.
  • the audio component 1010 includes a microphone (MIC), which is configured to receive an external audio signal when the electronic device 1000 is in an operation mode, such as a calling mode, a recording mode and a voice recognition mode. Received audio signals may be further stored in memory 1004 or sent via communication component 1016 .
  • the audio component 1010 also includes a speaker for outputting audio signals.
  • the I/O interface 1012 provides an interface between the processing component 1002 and a peripheral interface module, which may be a keyboard, a click wheel, a button, and the like. These buttons may include, but are not limited to: a home button, volume buttons, start button, and lock button.
  • Sensor assembly 1014 includes one or more sensors for providing status assessments of various aspects of electronic device 1000 .
  • the sensor component 1014 can detect the open/closed state of the electronic device 1000, the relative positioning of components, for example, the components are the display and the keypad of the electronic device 1000, the sensor component 1014 can also detect the electronic device 1000 or a Changes in position of components, presence or absence of user contact with electronic device 1000 , electronic device 1000 orientation or acceleration/deceleration and temperature changes in electronic device 1000 .
  • the sensor assembly 1014 may include a proximity sensor configured to detect the presence of nearby objects in the absence of any physical contact.
  • the sensor assembly 1014 may also include an optical sensor, such as a complementary metal-oxide-semiconductor (CMOS) or charge-coupled device (CCD) image sensor, for use in imaging applications.
  • CMOS complementary metal-oxide-semiconductor
  • CCD charge-coupled device
  • the sensor component 1014 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor or a temperature sensor.
  • the communication component 1016 is configured to facilitate wired or wireless communication between the electronic device 1000 and other devices.
  • the electronic device 1000 can access a wireless network based on a communication standard, such as a wireless network (WiFi), a second generation mobile communication technology (2G) or a third generation mobile communication technology (3G), or a combination thereof.
  • the communication component 1016 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel.
  • the communication component 1016 also includes a near field communication (NFC) module to facilitate short-range communication.
  • the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, Infrared Data Association (IrDA) technology, Ultra Wide Band (UWB) technology, Bluetooth (BT) technology and other technologies.
  • RFID Radio Frequency Identification
  • IrDA Infrared Data Association
  • UWB Ultra Wide Band
  • Bluetooth Bluetooth
  • electronic device 1000 may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGA), controller, microcontroller, microprocessor, or other electronic component implementation for performing the method described above.
  • ASICs application specific integrated circuits
  • DSPs digital signal processors
  • DSPDs digital signal processing devices
  • PLDs programmable logic devices
  • FPGA field programmable gate arrays
  • controller microcontroller, microprocessor, or other electronic component implementation for performing the method described above.
  • An embodiment of the present disclosure provides a non-volatile computer-readable storage medium, such as the memory 1004 including computer program instructions, which can be executed by the processor 1020 of the electronic device 1000 to complete the above method.
  • Fig. 11 shows a block diagram of an electronic device according to an embodiment of the present disclosure.
  • an electronic device 1100 may be provided as a server.
  • electronic device 1100 includes processing component 1122 , which further includes one or more processors, and a memory resource represented by memory 1132 for storing instructions executable by processing component 1122 , such as application programs.
  • the application program stored in memory 1132 may include one or more modules each corresponding to a set of instructions.
  • the processing component 1122 is configured to execute instructions to perform the above method.
  • Electronic device 1100 may also include a power supply component 1126 configured to perform power management of electronic device 1100, a wired or wireless network interface 1150 configured to connect electronic device 1100 to a network, and an input-output (I/O) interface 1158 .
  • the electronic device 1100 can operate based on the operating system stored in the memory 1132, such as the Microsoft server operating system (Windows ServerTM), the graphical user interface-based operating system (Mac OS XTM) introduced by Apple Inc., the multi-user and multi-process computer operating system (UnixTM). ), a free and open source Unix-like operating system (LinuxTM), an open source Unix-like operating system (FreeBSDTM), or similar.
  • An embodiment of the present disclosure provides a non-volatile computer-readable storage medium, such as a memory 1132 including computer program instructions, which can be executed by the processing component 1122 of the electronic device 1100 to complete the above method.
  • a non-volatile computer-readable storage medium such as a memory 1132 including computer program instructions, which can be executed by the processing component 1122 of the electronic device 1100 to complete the above method.
  • An embodiment of the present disclosure provides a computer program product, the computer program product carries a program code, and instructions included in the program code can be used to execute the steps of the image processing method in the above method embodiment, for details, refer to the above method embodiment.
  • Embodiments of the present disclosure may be systems, methods and/or computer program products.
  • a computer program product may include a computer readable storage medium having computer readable program instructions thereon for causing a processor to implement various aspects of the present disclosure.
  • a computer readable storage medium may be a tangible device that can retain and store instructions for use by an instruction execution device.
  • a computer readable storage medium may be, for example, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing.
  • Computer-readable storage media include: portable computer diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or flash memory), static random access memory (SRAM), compact disc read only memory (CD-ROM), digital versatile disc (DVD), memory stick, floppy disk, mechanically encoded device, such as a printer with instructions stored thereon A hole card or a raised structure in a groove, and any suitable combination of the above.
  • RAM random access memory
  • ROM read-only memory
  • EPROM erasable programmable read-only memory
  • flash memory static random access memory
  • SRAM static random access memory
  • CD-ROM compact disc read only memory
  • DVD digital versatile disc
  • memory stick floppy disk
  • mechanically encoded device such as a printer with instructions stored thereon
  • a hole card or a raised structure in a groove and any suitable combination of the above.
  • computer-readable storage media are not to be construed as transient signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., pulses of light through fiber optic cables), or transmitted electrical signals.
  • Computer-readable program instructions described herein may be downloaded from a computer-readable storage medium to a respective computing/processing device, or downloaded to an external computer or external storage device over a network, such as the Internet, a local area network, a wide area network, and/or a wireless network.
  • the network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and/or edge servers.
  • a network adapter card or a network interface in each computing/processing device receives computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in each computing/processing device .
  • Computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or Source or object code written in any combination, including object-oriented programming languages—such as Smalltalk, C++, etc., and conventional procedural programming languages—such as the “C” language or similar programming languages.
  • Computer readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server implement.
  • the remote computer can be connected to the user computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (such as via the Internet using an Internet service provider). connect).
  • electronic circuits such as programmable logic circuits, field programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), are customized by utilizing state information of computer-readable program instructions.
  • Computer readable program instructions may be executed to implement various aspects of the present disclosure.
  • These computer-readable program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine such that when executed by the processor of the computer or other programmable data processing apparatus , producing an apparatus for realizing the functions/actions specified in one or more blocks in the flowchart and/or block diagram.
  • These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause computers, programmable data processing devices and/or other devices to work in a specific way, so that the computer-readable medium storing instructions includes An article of manufacture comprising instructions for implementing various aspects of the functions/acts specified in one or more blocks in flowcharts and/or block diagrams.
  • each block in a flowchart or block diagram may represent a module, a portion of a program segment, or an instruction that includes one or more Executable instructions.
  • the functions noted in the block may occur out of the order noted in the figures. For example, two blocks in succession may, in fact, be executed substantially concurrently, or they may sometimes be executed in the reverse order, depending upon the functionality involved.
  • each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations can be implemented by a dedicated hardware-based system that performs the specified function or action , or may be implemented by a combination of dedicated hardware and computer instructions.
  • the computer program product can be specifically realized by means of hardware, software or a combination thereof.
  • the computer program product is embodied as a computer storage medium, and in another optional embodiment, the computer program product is embodied as a software product, such as a software development kit (Software Development Kit, SDK) etc. wait.
  • a software development kit Software Development Kit, SDK
  • Embodiments of the present disclosure relate to an image processing method and device, electronic equipment, and a storage medium.
  • the method includes: performing image segmentation on an image of a vertebra to be segmented to obtain a binary segmentation result corresponding to the image of a vertebra to be segmented; Segment the vertebral image and perform feature encoding to obtain the pixel point embedding vector corresponding to each pixel point in the vertebral image to be segmented; determine the pixel point embedding vector corresponding to the pixel point to be divided according to the binary segmentation result A pixel embedding vector corresponding to the vertebra area in the vertebra image; performing clustering on the pixel embedding vector corresponding to the vertebra area to obtain a vertebra segmentation result corresponding to the vertebra image to be segmented.
  • the embodiments of the present disclosure can reduce the probability of misclassification of adjacent vertebrae and improve the accuracy of vertebrae segmentation.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Multimedia (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

本公开涉及一种图像处理方法及装置、电子设备和存储介质，所述方法包括：对待分割椎骨图像进行图像分割，得到所述待分割椎骨图像对应的二值分割结果；对所述待分割椎骨图像进行特征编码，得到所述待分割椎骨图像中各像素点对应的像素点嵌入向量；根据所述二值分割结果和所述各像素点对应的像素点嵌入向量，确定所述待分割椎骨图像中椎骨区域对应的像素点嵌入向量；对所述椎骨区域对应的像素点嵌入向量进行聚类，得到所述待分割椎骨图像对应的椎骨分割结果。本公开实施例可以降低相邻椎骨被误分的概率，提高椎骨分割精度。

Description

图像处理方法及装置、电子设备、存储介质和程序
相关申请的交叉引用
本专利申请要求2021年09月28日提交的中国专利申请号为202111145348.6、申请人为上海商汤智能科技有限公司，申请名称为“图像处理方法及装置、电子设备和存储介质”的优先权，该申请的全文以引用的方式并入本申请中。
技术领域
本公开涉及计算机技术领域，尤其涉及一种图像处理方法及装置、电子设备、存储介质和程序。
背景技术
椎骨是人体骨骼系统的重要组成部分，在维持和支撑身体结构及器官的同时，还保护着中枢神经免受机械冲击造成伤害。然而由于骨质疏松、外力等因素，很容易造成椎骨骨折。在椎骨骨折诊断中，医生需要根据胸部的医疗影像(例如，胸部CT图像)，给出骨折所在椎骨的具体位置。相关技术中，使用神经网络直接进行椎骨分割。但是，由于相邻椎骨之间在外形上具有高度相似性，使得相邻椎骨之间容易造成误分割，导致椎骨分割的精度较低。
发明内容
本公开提出了一种图像处理方法及装置、电子设备、存储介质和程序的技术方案。
第一方面，本公开提供了一种图像处理方法，所述方法由电子设备执行，包括：对待分割椎骨图像进行图像分割，得到所述待分割椎骨图像对应的二值分割结果；对所述待分割椎骨图像进行特征编码，得到所述待分割椎骨图像中各像素点对应的像素点嵌入向量；根据所述二值分割结果和所述各像素点对应的像素点嵌入向量，确定所述待分割椎骨图像中椎骨区域对应的像素点嵌入向量；对所述椎骨区域对应的像素点嵌入向量进行聚类，得到所述待分割椎骨图像对应的椎骨分割结果。
第二方面，本公开提供了一种图像处理装置，包括：
图像分割模块，配置为对待分割椎骨图像进行图像分割，得到所述待分割椎骨图像对应的二值分割结果；
特征编码模块，配置为对所述待分割椎骨图像进行特征编码，得到所述待分割椎骨图像中各像素点对应的像素点嵌入向量；
第一确定模块，配置为根据所述二值分割结果和所述各像素点对应的像素点嵌入向量，确定所述待分割椎骨图像中椎骨区域对应的像素点嵌入向量；
聚类模块，配置为对所述椎骨区域对应的像素点嵌入向量进行聚类，得到所述待分割椎骨图像对应的椎骨分割结果。
第三方面，本公开提供了一种电子设备，包括：处理器；用于存储处理器可执行指令的存储器；其中，所述处理器被配置为调用所述存储器存储的指令，以执行上述方法。
第四方面，本公开提供了一种计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令被处理器执行时实现上述方法。
第五方面，本公开实施例提供了一种计算机程序，包括计算机可读代码，当所述计算机可读代码在电子设备中运行时，所述电子设备中的处理器执行用于实现上述任意一种图像处理方法。
在本公开实施例中，对待分割椎骨图像进行图像分割，得到待分割椎骨图像对应的二值分割结果；对待分割椎骨图像进行特征编码，得到待分割椎骨图像中各像素点对应的像素点嵌入向量；根据二值分割结果和各像素点对应的像素点嵌入向量，确定待分割椎骨图像中椎骨区域对应的像素点嵌入向量；由于像素点嵌入向量具有更高的语义表达能力，因此，对椎骨区域对应的像素点嵌入向量进行聚类，可以降低相邻椎骨被误分的概率，得到待分割椎骨图像对应的准确率较高的椎骨分割结果，有效提高了椎骨分割精 度。
应当理解的是，以上的一般描述和后文的细节描述仅是示例性和解释性的，而非限制本公开。根据下面参考附图对示例性实施例的详细说明，本公开的其它特征及方面将变得清楚。
附图说明
此处的附图被并入说明书中并构成本说明书的一部分，这些附图示出了符合本公开的实施例，并与说明书一起用于说明本公开的技术方案。
图1示出本公开实施例的一种图像处理方法的流程图；
图2示出本公开实施例的一种椎骨凸包区域的示意图；
图3示出本公开实施例的预设椎骨凸包分割网络的示意图
图4示出本公开实施例的椎骨分割结果的示意图；
图5示出本公开实施例的预设椎骨实例分割网络的示意图；
图6示出本公开实施例的肋骨端点和椎骨质心的示意图；
图7示出本公开实施例的目标肋骨-椎骨匹配结果的示意图；
图8示出本公开实施例的椎骨标记结果的示意图；
图9示出本公开实施例的一种图像处理装置的框图；
图10示出本公开实施例的一种电子设备的框图；
图11示出本公开实施例的一种电子设备的框图。
具体实施方式
以下将参考附图详细说明本公开的各种示例性实施例、特征和方面。附图中相同的附图标记表示功能相同或相似的元件。尽管在附图中示出了实施例的各种方面，但是除非特别指出，不必按比例绘制附图。
在这里专用的词“示例性”意为“用作例子、实施例或说明性”。这里作为“示例性”所说明的任何实施例不必解释为优于或好于其它实施例。
本文中术语“和/或”，仅仅是一种描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。另外，本文中术语“至少一种”表示多种中的任意一种或多种中的至少两种的任意组合，例如，包括A、B、C中的至少一种，可以表示包括从A、B和C构成的集合中选择的任意一个或多个元素。
另外，为了更好地说明本公开，在下文的具体实施方式中给出了众多的具体细节。本领域技术人员应当理解，没有某些具体细节，本公开同样可以实施。在一些实例中，对于本领域技术人员熟知的方法、手段、元件和电路未作详细描述，以便于凸显本公开的主旨。
本公开实施例提供的图像处理方法，可以由终端设备或服务器等电子设备执行，终端设备可以为用户设备(User Equipment，UE)、移动设备、用户终端、蜂窝电话、无绳电话、个人数字助理(Personal Digital Assistant，PDA)、手持设备、计算设备、车载设备、可穿戴设备等，该图像处理方法可以通过处理器调用存储器中存储的计算机可读指令的方式来实现。或者，可通过服务器执行该图像处理方法。
本公开实施例提供的图像处理方法，可以准确地定位骨折或病变区域所在椎骨的解剖学标签，给出脊柱和各块椎骨的3D重建可视化结果。能够为椎骨骨折检测、椎骨参数定量化分析、椎骨疾病筛查、病灶检测等项目提供技术支持，从而，降低医生的工作量和错误率，缩短了诊断周期。
图1示出根据本公开实施例的一种图像处理方法的流程图。如图1所示，本公开实施例提供的图像处理方法，所述方法由电子设备执行，可以包括以下步骤：
在步骤S11中，对待分割椎骨图像进行图像分割，得到待分割椎骨图像对应的二值分 割结果。
在步骤S12中，对待分割椎骨图像进行特征编码，得到待分割椎骨图像中各像素点对应的像素点嵌入向量。
在步骤S13中，根据二值分割结果和各像素点对应的像素点嵌入向量，确定待分割椎骨图像中椎骨区域对应的像素点嵌入向量。
在步骤S14中，对椎骨区域对应的像素点嵌入向量进行聚类，得到待分割椎骨图像对应的椎骨分割结果。
在本公开实施例中，对待分割椎骨图像进行图像分割，得到待分割椎骨图像对应的二值分割结果；对待分割椎骨图像进行特征编码，得到待分割椎骨图像中各像素点对应的像素点嵌入向量；根据二值分割结果和各像素点对应的像素点嵌入向量，确定待分割椎骨图像中椎骨区域对应的像素点嵌入向量；由于像素点嵌入向量具有更高的语义表达能力，因此，对椎骨区域对应的像素点嵌入向量进行聚类，可以降低相邻椎骨被误分的概率，得到待分割椎骨图像对应的准确率较高的椎骨分割结果，有效提高了椎骨分割精度。
在一种可能的实现方式中，该图像处理方法还包括以下步骤：获取原始胸部扫描图像；对原始胸部扫描图像进行图像预处理，得到初始椎骨图像；对初始椎骨图像进行凸包分割，得到初始椎骨图像中的椎骨凸包区域；根据椎骨凸包区域，对初始椎骨图像进行裁剪，得到待分割椎骨图像。
在示例中，在对原始胸部扫描图像进行图像预处理得到初始椎骨图像后，可以对初始椎骨图像进行凸包分割，确定初始椎骨图像中的椎骨凸包区域，从而可以根据椎骨凸包区域，从初始椎骨图像中裁剪得到待分割椎骨图像，以使得后续基于裁剪后的待分割椎骨图像进行椎骨分割，可以有效降低对计算资源的消耗，提高椎骨分割效率。
在一种可能的实现方式中，原始胸部扫描图像可以是胸部计算机X线断层扫描(CT)图像I。由于CT图像具有良好的骨-软组织对比度，因此，胸部CT图像I通常用于作为进行椎骨骨折诊断的医疗影像。
在示例中，对胸部CT图像I进行图像预处理，得到初始椎骨图像I n。图像预处理可以包括：重定向、裁剪、归一化等中的一种或多种，本公开实施例对图像预处理的具体方式不做限定。
应理解，由于拍摄角度不同，不同胸部CT图像中的椎骨方向可能不同，根据预设单位矩阵对胸部CT图像I进行重定向的预处理，得到初始椎骨图像I n，以使得初始椎骨图像I n中的椎骨方向与预设的坐标轴(x/y/z轴)一致，以提高后续处理效率。预设单位矩阵可以根据实际情况进行设置，本公开实施例对预设单位矩阵的具体形式不做限定。
在示例中，胸部CT图像I中除了包括骨区域以外，还包括大面积的其它背景部分，为了降低后续的计算资源消耗，提高处理效率，利用预设灰度阈值对胸部CT图像I进行裁剪处理。具体可以包括：基于预设灰度阈值对胸部CT图像I进行二值化处理，得到二值图像。其中，胸部CT图像I中灰度值大于或等于预设灰度阈值的像素点，在二值图像中对应的像素值为1；胸部CT图像I中灰度值小于预设灰度阈值的像素点，在二值图像中对应的像素值为0。根据二值图像中像素值为1的像素点构成的包围框，对胸部CT图像I进行裁剪，得到初始椎骨图像I n，以减小图像尺寸，从而有效降低后续的计算资源消耗，提高处理效率。预设灰度阈值的具体取值可以根据实际情况进行设置，本公开实施例对此不做具体限定。
在实际应用中，若拍摄胸部CT图像I的患者佩带金属等物品时，会使得拍摄得到的胸部CT图像I中存在灰度值过高的像素点，影响后续处理精度。因此，针对上述胸部CT 图像I裁剪后得到的初始椎骨图像I n，利用预设灰度值归一化窗口进行归一化处理，以使得初始椎骨图像I n的灰度值在合理灰度值范围内，提高椎骨分割精度。
在示例中，预设灰度值归一化窗口可以根据实际情况进行设置，例如，预设灰度值归一化窗口是[-1000，2000]，本公开实施例对预设灰度值归一化窗口的实际取值范围不作具体限定。
应理解，由于胸部CT图像I的扫描区域很大，初始椎骨图像I n中除了包括椎骨以外，还包括肋骨、髋骨、股骨等其它骨头，椎骨区域只占其中很小一部分，因此，可以基于凸包算法，对初始椎骨图像I n进行分割，以得到初始椎骨图像I n中的椎骨凸包区域，进而基于椎骨凸包区域对初始椎骨图像I n进行裁剪，得到包括椎骨区域的待分割椎骨图像I v。
这样，后续基于待分割椎骨图像I v进行椎骨分割，可以降低对计算资源的浪费，提高椎骨分割效率。
在示例中，可以基于预设椎骨凸包分割网络，对初始椎骨图像I n进行分割，得到初始椎骨图像I n中的椎骨凸包区域。为了提高分割效率，可以对初始椎骨图像I n进行重采样，得到第一椎骨图像I sp3，其中，第一椎骨图像I sp3的分辨率是第一分辨率。第一分辨率的具体取值可以根据实际情况确定，本公开实施例对此不作具体限定。
例如，第一分辨率是3mm×3mm×3mm，即，第一椎骨图像I sp3中的每个像素点对应的实际物理尺寸是3mm×3mm×3mm。
在示例中，将第一椎骨图像I sp3输入预设椎骨凸包分割网络，预设椎骨凸包分割网络对第一椎骨图像I sp3进行分割后，得到椎骨凸包区域H。图2示出根据本公开实施例的椎骨凸包区域的示意图。
在示例中，根据椎骨凸包区域H，可以确定包括椎骨整体区域的检测框。例如，该检测框可以是包括椎骨凸包区域H的最小三维矩形框。根据该检测框，对初始椎骨图像I n进行裁剪，得到包括椎骨整体区域的待分割椎骨图像I v。基于待分割椎骨图像I v，执行后续的椎骨分割。
在一种可能的实现方式中，预设椎骨凸包分割网络可以是3D-U型网络，包括由若干个卷积层和下采样层构成的编码器，以及由若干个卷积层和上采样层构成的解码器，编码器和解码器之间嵌入non-local模块，编码器和解码器对应阶段引入跳跃连接。
在示例中，图3示出根据本公开实施例的预设椎骨凸包分割网络的示意图。对预设椎骨凸包分割网络进行训练时，可以采用warmup和cosine annealing的学习率设置策略，以及采用cross entropy(交叉熵损失)和dice loss作为损失函数，对预设椎骨凸包分割网络进行若干训练回合的训练。
例如，可以通过最小化下述公式(1)所示的损失函数L 1，实现对预设椎骨凸包分割网络的训练。
Figure PCTCN2022077185-appb-000001
其中，公式(1)中的y i是训练样本图像i对应的分割标签，
Figure PCTCN2022077185-appb-000002
是根据预设椎骨凸包 分割网络确定的训练样本图像i对应的分割预测概率，Y是训练样本图像i对应的真实分割结果，
Figure PCTCN2022077185-appb-000003
是根据预设椎骨凸包分割网络确定的训练样本图像i对应的预测分割结果。
应理解，预设椎骨凸包分割网络的具体网络结构和训练过程可以采用相关技术中的其它网络结构和训练方式，本公开实施例对此不作具体限定。
在一种可能的实现方式中，预设椎骨凸包分割网络的训练样本包括椎骨样本图像，以及椎骨样本图像对应的椎骨凸包区域标签。其中，椎骨凸包区域标签可以是通过对椎骨样本图像进行二值分割得到的椎骨二值分割标签，经过若干次高斯平滑阈值膨胀得到的。椎骨凸包区域标签还可以采用相关技术中的其它方式确定得到，本公开实施例对此不作具体限定。
在示例中，可以利用凸包算法，对初始椎骨图像I n进行凸包分割，得到初始椎骨图像I n中的椎骨凸包区域。其中，凸包算法的具体算法形式，可以根据实际情况灵活设置，本公开实施例对此不作具体限定。
在一种可能的实现方式中，该图像处理方法还包括：对待分割椎骨图像中的像素点进行位置编码，得到位置编码图像。
在示例中，可以利用下述公式(2)对待分割椎骨图像I v进行位置编码，得到位置编码图像I c。
Figure PCTCN2022077185-appb-000004
其中，(i,j,k)是待分割椎骨图像I v和位置编码图像I c中的对应像素点，(μ x,μ y,μ z)是待分割椎骨图像I v的图像中心像素点，W x、W y和W z为预设超参数。预设超参数W x、W y和W z的具体取值可以根据实际情况确定，本公开实施例对此不作具体限定。
在示例中，为了提高分割效率，还可以对待分割椎骨图像I v进行重采样，得到第二椎骨图像I sp1.5，其中，第二椎骨图像I sp1.5的分辨率是第二分辨率。第二分辨率的具体取值可以根据实际情况确定，本公开实施例对此不作具体限定。
例如，第二分辨率是1.5mm×1.5mm×1.5mm，即，第二椎骨图像I sp1.5中的每个像素点对应的实际物理尺寸是1.5mm×1.5mm×1.5mm。利用上述公式(1)对第二椎骨图像I sp1.5进行位置编码，得到位置编码图像I c。进而利用第二椎骨图像I sp1.5和位置编码图像I c，实现对待分割椎骨图像I v的椎骨分割。
在一种可能的实现方式中，对待分割椎骨图像进行图像分割，得到待分割椎骨图像对应的二值分割结果，包括：基于位置编码图像，对待分割椎骨图像进行图像分割，得到二值分割结果。
在一种可能的实现方式中，对待分割椎骨图像进行特征编码，得到待分割椎骨图像中各像素点对应的像素点嵌入向量，包括：基于位置编码图像，对待分割椎骨图像进行特征编码，得到待分割椎骨图像中各像素点对应的像素点嵌入向量。
在一种可能的实现方式中，该图像处理方法还包括：根据椎骨凸包区域，对二值分割结果中的非椎骨区域进行过滤。
在示例中，可以基于预设椎骨实例分割网络，对待分割椎骨图像I v进行实例分割。 将待分割椎骨图像I v(或第二椎骨图像I sp1.5)和位置编码图像I c同时输入预设椎骨实例分割网络。
在示例中，预设椎骨实例分割网络包括两个分支，一个分支用于对待分割椎骨图像I v(或第二椎骨图像I sp1.5)进行二值分割，得到待分割椎骨图像I v(或第二椎骨图像I sp1.5)对应的二值分割结果A b。
在示例中，利用椎骨凸包区域H，对二值分割结果A b进行过滤，消除二值分割结果A b中被误分割为椎骨区域的假阳部分(非椎骨区域)，提高二值分割结果A b的精度。例如，二值分割结果A b＝A b∩H。
在示例中，预设椎骨实例分割网络的另一个分支用于基于位置编码图像I c，确定待分割椎骨图像I v(或第二椎骨图像I sp1.5)中各像素点对应的像素点嵌入向量A e。嵌入向量的维度可以是8维，也可以根据实际情况设置为其它维度，本公开实施例对此不作具体限定。
在示例中，利用二值分割结果A b和像素点嵌入向量A e，可以得到待分割椎骨图像I v(或第二椎骨图像I sp1.5)中的椎骨区域对应的像素点嵌入向量A re＝A b∩A e。然后，利用mean-shift聚类算法对椎骨区域对应的像素点嵌入向量A re进行聚类，得到待分割椎骨图像I v对应的椎骨分割结果A ins。
应理解，聚类算法除了可以采用mean-shift聚类算法，还可以采用相关技术中的其它聚类算法，本公开实施例对此不作具体限定。
图4示出根据本公开实施例的椎骨分割结果的示意图。如图4所示，椎骨分割结果中A ins中包括待分割椎骨图像I v中的多根椎骨。但是，椎骨分割结果A ins无法确定每根椎骨的椎骨标识。
在一种可能的实现方式中，预设椎骨实例分割网络可以是3D-U型网络，两个编码器分支共享一个解码器，编码器和解码器之间嵌入non-local模块。其中，一个编码器分支用于对待分割椎骨图像I v(或第二椎骨图像I sp1.5)进行二值分割，得到二值分割结果A b，另一个编码器分支用于确定待分割椎骨图像I v(或第二椎骨图像I sp1.5)中各像素点对应的像素点嵌入向量A e。
在示例中，图5示出根据本公开实施例的预设椎骨实例分割网络的示意图。对预设椎骨实例分割网络进行训练时，用于进行二值分割的编码器分支可以采用cross entropy和dice loss作为损失函数进行若干训练回合的训练。具体训练公式可以参考上述公式(1)。
在示例中，用于嵌入向量预测的编码器分支可以采用discriminative loss作为损失函数进行若干训练回合的训练。例如，可以通过最小化下述公式(3)所示的损失函数L d，实现对用于嵌入向量预测的编码器分支的训练。
Figure PCTCN2022077185-appb-000005
其中，公式(3)中的C是训练样本图像中的类别总数，N c是训练样本图像中属于同一类别的像素点数量，μ c是类别的均值向量，μ c是训练样本图像中像素点i的嵌入向量。
应理解，预设椎骨实例分割网络的具体网络结构和训练过程可以采用相关技术中的其它网络结构和训练方式，本公开实施例对此不作具体限定。
在椎骨骨折诊断中，医生需要根据胸部的医疗影像(例如，胸部CT图像)，给出骨折所在椎骨的准确椎骨标识(椎骨的解剖学标签，例如，3号椎骨)。相关技术中，医生通常需要一块块椎骨的从上往下数(或从下往上数)，以此确定椎骨标识，导致椎骨标识过程复杂且准确率较低。因此，在基于上述实施例方法得到椎骨分割结果之后，还需要提供一种能够自动标记椎骨标识的方法，提高椎骨标记效率以及准确率。
在一种可能的实现方式中，椎骨分割结果中包括待分割椎骨图像中的多根椎骨；该图像处理方法还包括：获取待分割椎骨图像中的椎骨区域对应的肋骨标记结果，其中，肋骨标记结果中包括多根肋骨以及每根肋骨对应的肋骨标识；根据多根椎骨和多根肋骨，确定目标肋骨-椎骨匹配结果，其中，目标肋骨-椎骨匹配结果中包括至少一个肋骨-椎骨匹配对；根据每根肋骨对应的肋骨标识，以及目标肋骨-椎骨匹配结果，确定每根椎骨对应的椎骨标识。
基于椎骨和肋骨的空间几何关系，将椎骨分割结果中的多根椎骨，与肋骨标记结果中的多根肋骨进行匹配，得到目标肋骨-椎骨匹配结果，进而利用肋骨标记结果中每根肋骨对应的肋骨标识，确定每根椎骨对应的椎骨标识，从而有效得到准确率较高的椎骨标记结果。
其中，肋骨标记结果是对于对待分割肋骨图像进行分割标记后得到的，待分割肋骨图像和待分割椎骨图像对应相同的目标对象，例如，待分割肋骨图像和待分割椎骨图像均是来源于对同一目标对象进行胸部计算机X线断层扫描后得到的胸部CT图像。
在示例中，对待分割肋骨图像进行分割标记，确定待分割肋骨图像对应的肋骨标记结果，包括：对待分割肋骨图像进行语义类别分割，得到待分割肋骨图像对应的语义类别分割结果；对待分割肋骨图像进行实例分割，得到待分割肋骨图像对应的实例分割结果；根据语义类别分割结果和实例分割结果，确定待分割肋骨图像对应的肋骨标记结果，其中，肋骨标记结果中包括待分割肋骨图像中的多根肋骨以及每根肋骨对应的肋骨标识。
基于图像全局语义信息，对待分割肋骨图像进行语义类别分割，得到待分割肋骨图像对应的语义类别分割结果；基于图像局部几何信息，对待分割肋骨图像进行实例分割，得到待分割肋骨图像对应的实例分割结果；由于综合考虑了图像全局语义信息和图像局部几何信息，因此，基于语义类别分割结果和实例分割结果，确定的包括待分割肋骨图像中的多根肋骨以及每根肋骨对应的肋骨标识的目标标记结果具有较高的准确率，从而有效提高了肋骨的分割和标记精度。
在示例中，对待分割肋骨图像进行分割标记，确定待分割肋骨图像对应的肋骨标记结果，还包括：在语义类别分割和实例分割是在第一分辨率下进行的情况下，在第二分辨率下对待分割肋骨图像进行二值分割，得到待分割肋骨图像对应的细二值分割结果， 其中，第二分辨率大于第一分辨率；根据细二值分割结果，更新肋骨标记结果。
对待分割肋骨图像分别进行语义类别分割和实例分割是在第一分辨率下进行的情况下，肋骨标记结果中每根肋骨的分割结果比较粗糙，因此，基于分辨率更高的细二值分割结果，更新肋骨标记结果，得到分辨率更高以及分割精度更高的肋骨标记结果。
应理解，对待分割肋骨图像进行分割标记的方式，除了可以采用上述方式之外，还可以采用相关技术中任意的肋骨分割标记方式，本公开实施例对此不作具体限定。
在一种可能的实现方式中，根据多根椎骨和多根肋骨，确定目标肋骨-椎骨匹配结果，包括：针对任意一根椎骨，确定椎骨对应的候选肋骨-椎骨匹配结果，其中，候选肋骨-椎骨匹配结果中包括至少一个肋骨-椎骨匹配对；根据每个候选肋骨-椎骨匹配结果中包括的至少一个肋骨-椎骨匹配对的空间几何关系，确定每个候选肋骨-椎骨匹配结果对应的匹配度；将匹配度符合预设匹配条件的候选肋骨-椎骨匹配结果，确定为目标肋骨-椎骨匹配结果。
在一种可能的实现方式中，根据每个候选肋骨-椎骨匹配结果中包括的至少一个肋骨-椎骨匹配对的空间几何关系，确定每个候选肋骨-椎骨匹配结果对应的匹配度，包括：根据肋骨标记结果，确定每根肋骨中与椎骨距离最近的肋骨端点；根据椎骨分割结果，确定每根椎骨对应的椎骨质心，以及每个椎骨质心对应的切向量；根据每根肋骨对应的肋骨端点、每根椎骨对应的椎骨质心，以及每个椎骨质心对应的切向量，确定每个候选肋骨-椎骨匹配结果对应的余弦距离均值；将每个候选肋骨-椎骨匹配结果对应的余弦距离均值，确定为每个候选肋骨-椎骨匹配结果对应的匹配度。
在示例中，椎骨分割结果中包括多根椎骨构成的椎骨集合{V i|V i∈A ins}，确定每根椎骨对应的椎骨质心m。根据每根椎骨对应的质心的坐标，对椎骨集合中的多根椎骨进行排序，得到椎骨序列V＝(V 1,V 2,...,V K)，以及椎骨质心序列M＝(m 1,m 2,...,m N)。将N个椎骨质心进行曲线拟合，得到脊柱中心线。确定每个椎骨质心在脊柱中心线上的切向量τ，得到切向量集合T＝(τ 1,τ 2,...,τ N)。
在示例中，肋骨标记结果中包括多根肋骨构成的肋骨序列R＝(R 1,R 2,...,R K)，确定每根肋骨中与椎骨距离最近的肋骨端点，即每根肋骨中距离脊柱中心线距离最近的肋骨端点，得到肋骨端点集合E＝(e 1,e 2,...,e K)。图6示出根据本公开实施例的肋骨端点和椎骨质心的示意图。
在示例中，枚举所有可能的肋骨-椎骨匹配方式，得到每根椎骨对应的候选肋骨-椎骨匹配结果
Figure PCTCN2022077185-appb-000006
例如，针对椎骨V sv，假设椎骨V sv与肋骨R sr是一个肋骨-椎骨匹配对，则按照肋骨序列R和椎骨序列V中的排序，可以得到椎骨V sv+1与肋骨R sr+1是一个肋骨-椎骨匹配对。
以此类推，得到椎骨V sv对应的候选肋骨-椎骨匹配结果
Figure PCTCN2022077185-appb-000007
遍历椎骨集合中的每根椎骨，得到候选肋骨-椎骨匹配结果集合
Figure PCTCN2022077185-appb-000008
在示例中，针对任意一个候选肋骨-椎骨匹配结果
Figure PCTCN2022077185-appb-000009
利用下述公式(4)，确定候选肋骨-椎骨匹配结果
Figure PCTCN2022077185-appb-000010
对应的余弦距离均值
Figure PCTCN2022077185-appb-000011
Figure PCTCN2022077185-appb-000012
其中，p是候选肋骨-椎骨匹配结果
Figure PCTCN2022077185-appb-000013
中包括的肋骨-椎骨匹配对的个数，(R i,V j)是候选肋骨-椎骨匹配结果
Figure PCTCN2022077185-appb-000014
中的一个肋骨-椎骨匹配对，e i是肋骨R i对应的肋骨端点，m j是椎骨V j对应的椎骨质心，τ j是椎骨质心m j对应的切向量。
在示例中，将每个候选肋骨-椎骨匹配结果对应的余弦距离均值，确定为其对应的匹配度，进而，将匹配度符合预设匹配条件的候选肋骨-椎骨匹配结果，确定为目标肋骨-椎骨匹配结果，即，将余弦距离均值符合预设匹配条件的候选肋骨-椎骨匹配结果，确定为目标肋骨-椎骨匹配结果。
在一种可能的实现方式中，将匹配度符合预设匹配条件的候选肋骨-椎骨匹配结果，确定为目标肋骨-椎骨匹配结果，包括：根据每个候选肋骨-椎骨匹配结果对应的余弦距离均值，确定是否存在目标肋骨-椎骨匹配结果集合，其中，目标肋骨-椎骨匹配结果集合中的每个候选肋骨-椎骨匹配结果对应的余弦距离均值均大于0；在存在目标肋骨-椎骨匹配结果集合的情况下，将目标肋骨-椎骨匹配结果集合中取值最小的余弦距离均值对应的候选肋骨-椎骨匹配结果，确定为目标肋骨-椎骨匹配结果。
在一种可能的实现方式中，该图像处理方法还包括：在不存在目标肋骨-椎骨匹配结果集合的情况下，将取值最大的余弦距离均值对应的候选肋骨-椎骨匹配结果，确定为目标肋骨-椎骨匹配结果。
在示例中，根据每个候选肋骨-椎骨匹配结果对应的余弦距离均值，将候选肋骨-椎骨匹配结果集合Φ分成正余弦匹配集合Φ +，以及负余弦匹配集合Φ -。其中，正余弦匹配集合Φ +中包括的候选肋骨-椎骨匹配结果对应的余弦距离均值均大于0，负余弦匹配集合Φ -中包括的候选肋骨-椎骨匹配结果对应的余弦距离均值均小于或等于0。
在示例中，在正余弦匹配集合Φ +为非空集合的情况下，即存在目标肋骨-椎骨匹配结果集合Φ +，此时，由于目标肋骨-椎骨匹配结果集合Φ +中包括的候选肋骨-椎骨匹配结果对应的余弦距离均值均大于0，候选肋骨-椎骨匹配结果的余弦距离均值的取值越小，可以表示该候选肋骨-椎骨匹配结果中包括的肋骨-椎骨匹配对中肋骨-椎骨的距离越近，匹配度越高，因此，可以将目标肋骨-椎骨匹配结果集合Φ +中取值最小的余弦距离均值对应的候选肋骨-椎骨匹配结果，确定为目标肋骨-椎骨匹配结果
Figure PCTCN2022077185-appb-000015
在示例中，在正余弦匹配集合Φ +为空集的情况下，即不存在目标肋骨-椎骨匹配结果集合Φ +，仅存在负余弦匹配集合Φ -，由于负余弦匹配集合Φ -中包括的候选肋骨-椎骨匹配结果对应的余弦距离均值均小于或等于0，候选肋骨-椎骨匹配结果的余弦距离均值的取值越大，可以表示该候选肋骨-椎骨匹配结果中包括的肋骨-椎骨匹配对中肋骨-椎骨的距离越近，匹配度越高，因此，可以将负余弦匹配集合Φ -中取值最大的余弦距离均值对应的候选肋骨-椎骨匹配结果，确定为目标肋骨-椎骨匹配结果
Figure PCTCN2022077185-appb-000016
在示例中，可以根据下述公式(5)，确定目标肋骨-椎骨匹配结果
Figure PCTCN2022077185-appb-000017
Figure PCTCN2022077185-appb-000018
图7示出根据本公开实施例的目标肋骨-椎骨匹配结果的示意图。如图7所示，目标肋骨-椎骨匹配结果中的每个肋骨-椎骨匹配对中的肋骨端点和椎骨质心距离最近。
在示例中，在确定目标肋骨-椎骨匹配结果之后，目标肋骨-椎骨匹配结果中包括至少一个肋骨-椎骨匹配对，根据每个肋骨-椎骨匹配对中肋骨的肋骨标识，以及肋骨-椎骨在解剖学上的对应关系，确定每个肋骨-椎骨匹配对中椎骨的椎骨标识，得到至少一根椎骨的椎骨标识，进而根据椎骨集合中多根椎骨的排序，可以得到每根椎骨的椎骨标识，得到准确率较高的椎骨标记结果，有效实现对待分割椎骨图像的椎骨标记，提高椎骨标记精度。
图8示出根据本公开实施例的椎骨标记结果的示意图。如图8所示，椎骨标记结果中包括待分割椎骨图像I v中的多根椎骨以及每根椎骨对应的椎骨标识。在图8中，可以通过不同的颜色来指示每根椎骨以及每根椎骨对应的椎骨标识。例如，灰色用于指示1号椎骨，绿色用于指示2号椎骨，以此类推。
在示例中，椎骨标记结果中可以采用相关技术中的其它形式来指示每根椎骨以及每根椎骨对应的椎骨标识，本公开实施例对此不作具体限定。
上述各个方法实施例，在不违背原理逻辑的情况下，均可以彼此相互结合形成结合后的实施例。本领域技术人员可以理解，在具体实施方式的上述方法中，各步骤的具体执行顺序应当以其功能和可能的内在逻辑确定。
此外，本公开实施例还提供了图像处理装置、电子设备、计算机可读存储介质、程序，上述均可用来实现本公开实施例提供的任一种图像处理方法，相应技术方案和描述和参见方法部分的相应记载。
图9示出本公开实施例的一种图像处理装置的框图。如图9所示，装置90包括：
图像分割模块91，配置为对待分割椎骨图像进行图像分割，得到待分割椎骨图像对应的二值分割结果；
特征编码模块92，配置为对待分割椎骨图像进行特征编码，得到待分割椎骨图像中各像素点对应的像素点嵌入向量；
第一确定模块93，配置为根据二值分割结果和各像素点对应的像素点嵌入向量，确定待分割椎骨图像中椎骨区域对应的像素点嵌入向量；
聚类模块94，配置为对椎骨区域对应的像素点嵌入向量进行聚类，得到待分割椎骨图像对应的椎骨分割结果。
在一种可能的实现方式中，装置90还包括：
第一获取模块，配置为获取原始胸部扫描图像；
图像预处理模块，配置为对原始胸部扫描图像进行图像预处理，得到初始椎骨图像；
凸包分割模块，配置为对初始椎骨图像进行凸包分割，得到初始椎骨图像中的椎骨凸包区域；
裁剪模块，配置为根据椎骨凸包区域，对初始椎骨图像进行裁剪，得到待分割椎骨图像。
在一种可能的实现方式中，装置90还包括：
位置编码模块，配置为对待分割椎骨图像中的像素点进行位置编码，得到位置编码图像。
在一种可能的实现方式中，图像分割模块91，配置为：
基于位置编码图像，对待分割椎骨图像进行图像分割，得到二值分割结果。
在一种可能的实现方式中，特征编码模块92，配置为：
基于位置编码图像，对待分割椎骨图像进行特征编码，得到各像素点对应的像素点嵌入向量。
在一种可能的实现方式中，装置90还包括：
过滤模块，配置为根据椎骨凸包区域，对二值分割结果中的非椎骨区域进行过滤。
在一种可能的实现方式中，椎骨分割结果中包括待分割椎骨图像中的多根椎骨；
装置70还包括：
第二获取模块，配置为获取待分割椎骨图像中的椎骨区域对应的肋骨标记结果，其中，肋骨标记结果中包括多根肋骨以及每根肋骨对应的肋骨标识；
第二确定模块，配置为根据多根椎骨和多根肋骨，确定目标肋骨-椎骨匹配结果，其中，目标肋骨-椎骨匹配结果中包括至少一个肋骨-椎骨匹配对；
第三确定模块，配置为根据每根肋骨对应的肋骨标识，以及目标肋骨-椎骨匹配结果，确定每根椎骨对应的椎骨标识。
在一种可能的实现方式中，第二确定模块，包括：
第一确定子模块，配置为针对任意一根椎骨，确定该椎骨对应的候选肋骨-椎骨匹配结果，其中，候选肋骨-椎骨匹配结果中包括至少一个肋骨-椎骨匹配对；
第二确定子模块，配置为根据每个候选肋骨-椎骨匹配结果中包括的至少一个肋骨-椎骨匹配对的空间几何关系，确定每个候选肋骨-椎骨匹配结果对应的匹配度；
第三确定子模块，配置为将匹配度符合预设匹配条件的候选肋骨-椎骨匹配结果，确定为目标肋骨-椎骨匹配结果。
在一种可能的实现方式中，第二确定子模块，包括：
第一确定单元，配置为根据肋骨标记结果，确定每根肋骨中与椎骨距离最近的肋骨端点；
第二确定单元，配置为根据椎骨分割结果，确定每根椎骨对应的椎骨质心，以及每个椎骨质心对应的切向量；
第三确定单元，配置为根据每根肋骨对应的肋骨端点、每根椎骨对应的椎骨质心，以及每个椎骨质心对应的切向量，确定每个候选肋骨-椎骨匹配结果对应的余弦距离均值；
第四确定单元，配置为将每个候选肋骨-椎骨匹配结果对应的余弦距离均值，确定为每个候选肋骨-椎骨匹配结果对应的匹配度。
在一种可能的实现方式中，第三确定子模块，包括：
第五确定单元，配置为根据每个候选肋骨-椎骨匹配结果对应的余弦距离均值，确定是否存在目标肋骨-椎骨匹配结果集合，其中，目标肋骨-椎骨匹配结果集合中的每个候选肋骨-椎骨匹配结果对应的余弦距离均值均大于0；
第六确定单元，配置为在存在目标肋骨-椎骨匹配结果集合的情况下，将目标肋骨-椎骨匹配结果集合中取值最小的余弦距离均值对应的候选肋骨-椎骨匹配结果，确定为目标肋骨-椎骨匹配结果。
在一种可能的实现方式中，装置90还包括：
第四确定模块，配置为在不存在目标肋骨-椎骨匹配结果集合的情况下，将取值最大的余弦距离均值对应的候选肋骨-椎骨匹配结果，确定为目标肋骨-椎骨匹配结果。
在一些实施例中，本公开实施例提供的装置具有的功能或包含的模块可以用于执行上文方法实施例描述的方法，其具体实现可以参照上文方法实施例的描述。
本公开实施例还提出一种计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令被处理器执行时实现上述方法。计算机可读存储介质可以是易失性或非易失性计算机可读存储介质。
本公开实施例还提出一种电子设备，包括：处理器；用于存储处理器可执行指令的存储器；其中，所述处理器被配置为调用所述存储器存储的指令，以执行上述方法。
本公开实施例还提供了一种计算机程序产品，包括计算机可读代码，或者承载有计 算机可读代码的非易失性计算机可读存储介质，当所述计算机可读代码在电子设备的处理器中运行时，所述电子设备中的处理器执行上述方法。
电子设备可以被提供为终端、服务器或其它形态的设备。
图10示出本公开实施例的一种电子设备的框图。如图10所示，电子设备800可以是移动电话，计算机，数字广播终端，消息收发设备，游戏控制台，平板设备，医疗设备，健身设备，个人数字助理等终端。
参照图10，电子设备1000可以包括以下一个或多个组件：处理组件1002，存储器1004，电源组件1006，多媒体组件1008，音频组件1010，输入/输出(I/O)的接口1012，传感器组件1014，以及通信组件1016。
处理组件1002通常控制电子设备1000的整体操作，诸如与显示，电话呼叫，数据通信，相机操作和记录操作相关联的操作。处理组件1002可以包括一个或多个处理器1020来执行指令，以完成上述的方法的全部或部分步骤。此外，处理组件1002可以包括一个或多个模块，便于处理组件1002和其他组件之间的交互。例如，处理组件1002可以包括多媒体模块，以方便多媒体组件1008和处理组件1002之间的交互。
存储器1004被配置为存储各种类型的数据以支持在电子设备1000的操作。这些数据的示例包括用于在电子设备1000上操作的任何应用程序或方法的指令，联系人数据，电话簿数据，消息，图片，视频等。存储器1004可以由任何类型的易失性或非易失性存储设备或者它们的组合实现，如静态随机存取存储器(SRAM)，电可擦除可编程只读存储器(EEPROM)，可擦除可编程只读存储器(EPROM)，可编程只读存储器(PROM)，只读存储器(ROM)，磁存储器，快闪存储器，磁盘或光盘。
电源组件1006为电子设备1000的各种组件提供电力。电源组件1006可以包括电源管理系统，一个或多个电源，及其他与为电子设备1000生成、管理和分配电力相关联的组件。
多媒体组件1008包括在所述电子设备1000和用户之间的提供一个输出接口的屏幕。在一些实施例中，屏幕可以包括液晶显示器(LCD)和触摸面板(TP)。如果屏幕包括触摸面板，屏幕可以被实现为触摸屏，以接收来自用户的输入信号。触摸面板包括一个或多个触摸传感器以感测触摸、滑动和触摸面板上的手势。所述触摸传感器可以不仅感测触摸或滑动动作的边界，而且还检测与所述触摸或滑动操作相关的持续时间和压力。在一些实施例中，多媒体组件1008包括一个前置摄像头和/或后置摄像头。当电子设备1000处于操作模式，如拍摄模式或视频模式时，前置摄像头和/或后置摄像头可以接收外部的多媒体数据。每个前置摄像头和后置摄像头可以是一个固定的光学透镜系统或具有焦距和光学变焦能力。
音频组件1010被配置为输出和/或输入音频信号。例如，音频组件1010包括一个麦克风(MIC)，当电子设备1000处于操作模式，如呼叫模式、记录模式和语音识别模式时，麦克风被配置为接收外部音频信号。所接收的音频信号可以被进一步存储在存储器1004或经由通信组件1016发送。在一些实施例中，音频组件1010还包括一个扬声器，用于输出音频信号。
I/O接口1012为处理组件1002和外围接口模块之间提供接口，上述外围接口模块可以是键盘，点击轮，按钮等。这些按钮可包括但不限于：主页按钮、音量按钮、启动按钮和锁定按钮。
传感器组件1014包括一个或多个传感器，用于为电子设备1000提供各个方面的状态评估。例如，传感器组件1014可以检测到电子设备1000的打开/关闭状态，组件的相对定位，例如所述组件为电子设备1000的显示器和小键盘，传感器组件1014还可以检测电子设备1000或电子设备1000一个组件的位置改变，用户与电子设备1000接触的存在或不存在，电子设备1000方位或加速/减速和电子设备1000的温度变化。传感器组件1014可以包 括接近传感器，被配置用来在没有任何的物理接触时检测附近物体的存在。传感器组件1014还可以包括光传感器，如互补金属氧化物半导体(CMOS)或电荷耦合装置(CCD)图像传感器，用于在成像应用中使用。在一些实施例中，该传感器组件1014还可以包括加速度传感器，陀螺仪传感器，磁传感器，压力传感器或温度传感器。
通信组件1016被配置为便于电子设备1000和其他设备之间有线或无线方式的通信。电子设备1000可以接入基于通信标准的无线网络，如无线网络(WiFi)，第二代移动通信技术(2G)或第三代移动通信技术(3G)，或它们的组合。在一个示例性实施例中，通信组件1016经由广播信道接收来自外部广播管理系统的广播信号或广播相关信息。在一个示例性实施例中，所述通信组件1016还包括近场通信(NFC)模块，以促进短程通信。例如，在NFC模块可基于射频识别(RFID)技术，红外数据协会(IrDA)技术，超宽带(UWB)技术，蓝牙(BT)技术和其他技术来实现。
在示例中，电子设备1000可以被一个或多个应用专用集成电路(ASIC)、数字信号处理器(DSP)、数字信号处理设备(DSPD)、可编程逻辑器件(PLD)、现场可编程门阵列(FPGA)、控制器、微控制器、微处理器或其他电子元件实现，用于执行上述方法。
本公开实施例提供了一种非易失性计算机可读存储介质，例如包括计算机程序指令的存储器1004，上述计算机程序指令可由电子设备1000的处理器1020执行以完成上述方法。
图11示出本公开实施例的一种电子设备的框图。如图11所示，电子设备1100可以被提供为一服务器。参照图11，电子设备1100包括处理组件1122，其进一步包括一个或多个处理器，以及由存储器1132所代表的存储器资源，用于存储可由处理组件1122的执行的指令，例如应用程序。存储器1132中存储的应用程序可以包括一个或一个以上的每一个对应于一组指令的模块。此外，处理组件1122被配置为执行指令，以执行上述方法。
电子设备1100还可以包括一个电源组件1126被配置为执行电子设备1100的电源管理，一个有线或无线网络接口1150被配置为将电子设备1100连接到网络，和一个输入输出(I/O)接口1158。电子设备1100可以操作基于存储在存储器1132的操作系统，例如微软服务器操作系统(Windows ServerTM)，苹果公司推出的基于图形用户界面操作系统(Mac OS XTM)，多用户多进程的计算机操作系统(UnixTM),自由和开放原代码的类Unix操作系统(LinuxTM)，开放原代码的类Unix操作系统(FreeBSDTM)或类似。
本公开实施例提供了一种非易失性计算机可读存储介质，例如包括计算机程序指令的存储器1132，上述计算机程序指令可由电子设备1100的处理组件1122执行以完成上述方法。
本公开实施例提供一种计算机程序产品，该计算机程序产品承载有程序代码，所述程序代码包括的指令可用于执行上述方法实施例中的图像处理方法的步骤，具体可参见上述方法实施例。
本公开实施例可以是系统、方法和/或计算机程序产品。计算机程序产品可以包括计算机可读存储介质，其上载有用于使处理器实现本公开的各个方面的计算机可读程序指令。
计算机可读存储介质可以是可以保持和存储由指令执行设备使用的指令的有形设备。计算机可读存储介质例如可以是(但不限于)电存储设备、磁存储设备、光存储设备、电磁存储设备、半导体存储设备或者上述的任意合适的组合。计算机可读存储介质的更具体的例子(非穷举的列表)包括：便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、静态随机存取存储器(SRAM)、便携式压缩盘只读存储器(CD-ROM)、数字多功能盘(DVD)、记忆棒、软盘、机械编码设备、例如其上存储有指令的打孔卡或凹槽内凸起结构、以及上述的任意合适的组合。这里所使用的计算机可读存储介质不被解释为瞬时信号本身，诸如无线电 波或者其他自由传播的电磁波、通过波导或其他传输媒介传播的电磁波(例如，通过光纤电缆的光脉冲)、或者通过电线传输的电信号。
这里所描述的计算机可读程序指令可以从计算机可读存储介质下载到各个计算/处理设备，或者通过网络、例如因特网、局域网、广域网和/或无线网下载到外部计算机或外部存储设备。网络可以包括铜传输电缆、光纤传输、无线传输、路由器、防火墙、交换机、网关计算机和/或边缘服务器。每个计算/处理设备中的网络适配卡或者网络接口从网络接收计算机可读程序指令，并转发该计算机可读程序指令，以供存储在各个计算/处理设备中的计算机可读存储介质中。
用于执行本公开操作的计算机程序指令可以是汇编指令、指令集架构(ISA)指令、机器指令、机器相关指令、微代码、固件指令、状态设置数据、或者以一种或多种编程语言的任意组合编写的源代码或目标代码，所述编程语言包括面向对象的编程语言—诸如Smalltalk、C++等，以及常规的过程式编程语言—诸如“C”语言或类似的编程语言。计算机可读程序指令可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络—包括局域网(LAN)或广域网(WAN)—连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。在一些实施例中，通过利用计算机可读程序指令的状态信息来个性化定制电子电路，例如，可编程逻辑电路、现场可编程门阵列(FPGA)或可编程逻辑阵列(PLA)，该电子电路可以执行计算机可读程序指令，从而实现本公开的各个方面。
这里参照根据本公开实施例的方法、装置(系统)和计算机程序产品的流程图和/或框图描述了本公开的各个方面。应当理解，流程图和/或框图的每个方框以及流程图和/或框图中各方框的组合，都可以由计算机可读程序指令实现。
这些计算机可读程序指令可以提供给通用计算机、专用计算机或其它可编程数据处理装置的处理器，从而生产出一种机器，使得这些指令在通过计算机或其它可编程数据处理装置的处理器执行时，产生了实现流程图和/或框图中的一个或多个方框中规定的功能/动作的装置。也可以把这些计算机可读程序指令存储在计算机可读存储介质中，这些指令使得计算机、可编程数据处理装置和/或其他设备以特定方式工作，从而，存储有指令的计算机可读介质则包括一个制造品，其包括实现流程图和/或框图中的一个或多个方框中规定的功能/动作的各个方面的指令。
也可以把计算机可读程序指令加载到计算机、其它可编程数据处理装置、或其它设备上，使得在计算机、其它可编程数据处理装置或其它设备上执行一系列操作步骤，以产生计算机实现的过程，从而使得在计算机、其它可编程数据处理装置、或其它设备上执行的指令实现流程图和/或框图中的一个或多个方框中规定的功能/动作。
附图中的流程图和框图显示了根据本公开的多个实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段或指令的一部分，所述模块、程序段或指令的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个连续的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或动作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。
该计算机程序产品可以具体通过硬件、软件或其结合的方式实现。在一个可选实施例中，所述计算机程序产品具体体现为计算机存储介质，在另一个可选实施例中，计算 机程序产品具体体现为软件产品，例如软件开发包(Software Development Kit，SDK)等等。
以上已经描述了本公开的各实施例，上述说明是示例性的，并非穷尽性的，并且也不限于所披露的各实施例。在不偏离所说明的各实施例的范围和精神的情况下，对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。本文中所用术语的选择，旨在最好地解释各实施例的原理、实际应用或对市场中的技术的改进，或者使本技术领域的其它普通技术人员能理解本文披露的各实施例。
工业实用性
本公开实施例涉及一种图像处理方法及装置、电子设备和存储介质，所述方法包括：对待分割椎骨图像进行图像分割，得到所述待分割椎骨图像对应的二值分割结果；对所述待分割椎骨图像进行特征编码，得到所述待分割椎骨图像中各像素点对应的像素点嵌入向量；根据所述二值分割结果和所述各像素点对应的像素点嵌入向量，确定所述待分割椎骨图像中椎骨区域对应的像素点嵌入向量；对所述椎骨区域对应的像素点嵌入向量进行聚类，得到所述待分割椎骨图像对应的椎骨分割结果。本公开实施例可以降低相邻椎骨被误分的概率，提高椎骨分割精度。

Claims (14)

  1. 一种图像处理方法，应用于电子设备中，所述方法包括：
    对待分割椎骨图像进行图像分割，得到所述待分割椎骨图像对应的二值分割结果；
    对所述待分割椎骨图像进行特征编码，得到所述待分割椎骨图像中各像素点对应的像素点嵌入向量；
    根据所述二值分割结果和所述各像素点对应的像素点嵌入向量，确定所述待分割椎骨图像中椎骨区域对应的像素点嵌入向量；
    对所述椎骨区域对应的像素点嵌入向量进行聚类，得到所述待分割椎骨图像对应的椎骨分割结果。
  2. 根据权利要求1所述的方法，其中，所述方法还包括：
    获取原始胸部扫描图像；
    对所述原始胸部扫描图像进行图像预处理，得到初始椎骨图像；
    对所述初始椎骨图像进行凸包分割，得到所述初始椎骨图像中的椎骨凸包区域；
    根据所述椎骨凸包区域，对所述初始椎骨图像进行裁剪，得到所述待分割椎骨图像。
  3. 根据权利要求2所述的方法，其中，所述方法还包括：
    对所述待分割椎骨图像中的像素点进行位置编码，得到位置编码图像。
  4. 根据权利要求3所述的方法，其中，所述对待分割椎骨图像进行图像分割，得到所述待分割椎骨图像对应的二值分割结果，包括：
    基于所述位置编码图像，对所述待分割椎骨图像进行图像分割，得到所述二值分割结果。
  5. 根据权利要求3或4所述的方法，其中，所述对所述待分割椎骨图像进行特征编码，得到所述待分割椎骨图像中各像素点对应的像素点嵌入向量，包括：
    基于所述位置编码图像，对所述待分割椎骨图像进行特征编码，得到所述各像素点对应的像素点嵌入向量。
  6. 根据权利要求4或5所述的方法，其中，所述方法还包括：
    根据所述椎骨凸包区域，对所述二值分割结果中的非椎骨区域进行过滤。
  7. 根据权利要求1至6中任意一项所述的方法，其中，所述椎骨分割结果中包括待分割椎骨图像中的多根椎骨；
    所述方法还包括：
    获取所述待分割椎骨图像中的椎骨区域对应的肋骨标记结果，其中，所述肋骨标记结果中包括多根肋骨以及每根肋骨对应的肋骨标识；
    根据所述多根椎骨和所述多根肋骨，确定目标肋骨-椎骨匹配结果，其中，所述目标肋骨-椎骨匹配结果中包括至少一个肋骨-椎骨匹配对；
    根据所述每根肋骨对应的肋骨标识，以及所述目标肋骨-椎骨匹配结果，确定每根椎骨对应的椎骨标识。
  8. 根据权利要求7所述的方法，其中，所述根据所述多根椎骨和所述多根肋骨，确定目标肋骨-椎骨匹配结果，包括：
    针对任意一根椎骨，确定所述椎骨对应的候选肋骨-椎骨匹配结果，其中，所述候选肋骨-椎骨匹配结果中包括至少一个肋骨-椎骨匹配对；
    根据每个所述候选肋骨-椎骨匹配结果中包括的至少一个肋骨-椎骨匹配对的空间几何关系，确定每个所述候选肋骨-椎骨匹配结果对应的匹配度；
    将所述匹配度符合预设匹配条件的所述候选肋骨-椎骨匹配结果，确定为所述目标肋骨-椎骨匹配结果。
  9. 根据权利要求8所述的方法，其中，所述根据每个所述候选肋骨-椎骨匹配结果中包括的至少一个肋骨-椎骨匹配对的空间几何关系，确定每个所述候选肋骨-椎骨匹配结 果对应的匹配度，包括：
    根据所述肋骨标记结果，确定每根肋骨中与椎骨距离最近的肋骨端点；
    根据所述椎骨分割结果，确定每根椎骨对应的椎骨质心，以及每个椎骨质心对应的切向量；
    根据每根肋骨对应的肋骨端点、每根椎骨对应的椎骨质心，以及每个椎骨质心对应的切向量，确定每个所述候选肋骨-椎骨匹配结果对应的余弦距离均值；
    将每个所述候选肋骨-椎骨匹配结果对应的余弦距离均值，确定为每个所述候选肋骨-椎骨匹配结果对应的匹配度。
  10. 根据权利要求9所述的方法，其中，所述将所述匹配度符合预设匹配条件的所述候选肋骨-椎骨匹配结果，确定为所述目标肋骨-椎骨匹配结果，包括：
    根据每个所述候选肋骨-椎骨匹配结果对应的余弦距离均值，确定是否存在目标肋骨-椎骨匹配结果集合，其中，所述目标肋骨-椎骨匹配结果集合中的每个所述候选肋骨-椎骨匹配结果对应的余弦距离均值均大于0；
    在存在所述目标肋骨-椎骨匹配结果集合的情况下，将所述目标肋骨-椎骨匹配结果集合中取值最小的所述余弦距离均值对应的所述候选肋骨-椎骨匹配结果，确定为所述目标肋骨-椎骨匹配结果。
  11. 根据权利要求10所述的方法，其中，所述方法还包括：
    在不存在所述目标肋骨-椎骨匹配结果集合的情况下，将取值最大的所述余弦距离均值对应的所述候选肋骨-椎骨匹配结果，确定为所述目标肋骨-椎骨匹配结果。
  12. 一种图像处理装置，应用于电子设备中，包括：
    图像分割模块，配置为对待分割椎骨图像进行图像分割，得到所述待分割椎骨图像对应的二值分割结果；
    特征编码模块，配置为对所述待分割椎骨图像进行特征编码，得到所述待分割椎骨图像中各像素点对应的像素点嵌入向量；
    第一确定模块，配置为根据所述二值分割结果和所述各像素点对应的像素点嵌入向量，确定所述待分割椎骨图像中椎骨区域对应的像素点嵌入向量；
    聚类模块，配置为对所述椎骨区域对应的像素点嵌入向量进行聚类，得到所述待分割椎骨图像对应的椎骨分割结果。
  13. 一种电子设备，包括：
    处理器；
    用于存储处理器可执行指令的存储器；
    其中，所述处理器被配置为调用所述存储器存储的指令，以执行权利要求1至11中任意一项所述的方法。
  14. 一种计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令被处理器执行时实现权利要求1至11中任意一项所述的方法。
PCT/CN2022/077185 2021-09-28 2022-02-22 图像处理方法及装置、电子设备、存储介质和程序 WO2023050691A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111145348.6A CN113888548A (zh) 2021-09-28 2021-09-28 图像处理方法及装置、电子设备和存储介质
CN202111145348.6 2021-09-28

Publications (1)

Publication Number Publication Date
WO2023050691A1 true WO2023050691A1 (zh) 2023-04-06

Family

ID=79007551

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/077185 WO2023050691A1 (zh) 2021-09-28 2022-02-22 图像处理方法及装置、电子设备、存储介质和程序

Country Status (2)

Country Link
CN (1) CN113888548A (zh)
WO (1) WO2023050691A1 (zh)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113888548A (zh) * 2021-09-28 2022-01-04 上海商汤智能科技有限公司 图像处理方法及装置、电子设备和存储介质
CN113947603A (zh) * 2021-09-28 2022-01-18 上海商汤智能科技有限公司 图像处理方法及装置、电子设备和存储介质
CN115456990B (zh) * 2022-09-13 2023-05-23 北京医准智能科技有限公司 一种基于ct图像的肋骨计数方法、装置、设备及存储介质
CN115984536B (zh) * 2023-03-20 2023-06-30 慧影医疗科技(北京)股份有限公司 一种基于ct影像的图像处理方法及装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150243055A1 (en) * 2014-02-27 2015-08-27 Fujifilm Corporation Image Display Device And Method, And Medium Containing Program
CN108537779A (zh) * 2018-03-27 2018-09-14 哈尔滨理工大学 基于聚类的椎骨分割与质心检测的方法
CN109740609A (zh) * 2019-01-09 2019-05-10 银河水滴科技（北京）有限公司 一种轨距检测方法及装置
CN110765916A (zh) * 2019-10-17 2020-02-07 北京中科原动力科技有限公司 一种基于语义和实例分割的农田苗垄识别方法及系统
CN113888548A (zh) * 2021-09-28 2022-01-04 上海商汤智能科技有限公司 图像处理方法及装置、电子设备和存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150243055A1 (en) * 2014-02-27 2015-08-27 Fujifilm Corporation Image Display Device And Method, And Medium Containing Program
CN108537779A (zh) * 2018-03-27 2018-09-14 哈尔滨理工大学 基于聚类的椎骨分割与质心检测的方法
CN109740609A (zh) * 2019-01-09 2019-05-10 银河水滴科技（北京）有限公司 一种轨距检测方法及装置
CN110765916A (zh) * 2019-10-17 2020-02-07 北京中科原动力科技有限公司 一种基于语义和实例分割的农田苗垄识别方法及系统
CN113888548A (zh) * 2021-09-28 2022-01-04 上海商汤智能科技有限公司 图像处理方法及装置、电子设备和存储介质

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
NEVEN DAVY; BRABANDERE BERT DE; GEORGOULIS STAMATIOS; PROESMANS MARC; GOOL LUC VAN: "Towards End-to-End Lane Detection: an Instance Segmentation Approach", 2018 IEEE INTELLIGENT VEHICLES SYMPOSIUM (IV), IEEE, 26 June 2018 (2018-06-26), pages 286 - 291, XP033423446, DOI: 10.1109/IVS.2018.8500547 *

Also Published As

Publication number Publication date
CN113888548A (zh) 2022-01-04

Similar Documents

Publication Publication Date Title
WO2023050691A1 (zh) 图像处理方法及装置、电子设备、存储介质和程序
US20220198775A1 (en) Image processing method and apparatus, electronic device, storage medium and computer program
WO2022151755A1 (zh) 目标检测方法及装置、电子设备、存储介质、计算机程序产品和计算机程序
CN112767329B (zh) 图像处理方法及装置、电子设备
US20220180521A1 (en) Image processing method and apparatus, and electronic device, storage medium and computer program
US20210319560A1 (en) Image processing method and apparatus, and storage medium
US20220327711A1 (en) Image segmentation method and apparatus, electronic device and storage medium
US10891473B2 (en) Method and device for use in hand gesture recognition
CN111899268B (zh) 图像分割方法及装置、电子设备和存储介质
JP2022537974A (ja) ニューラルネットワーク訓練方法及び装置、電子機器並びに記憶媒体
CN113222038B (zh) 基于核磁图像的乳腺病灶分类和定位方法及装置
CN112967291B (zh) 图像处理方法及装置、电子设备和存储介质
CN110211086B (zh) 图像分割方法、装置及存储介质
CN112508918A (zh) 图像处理方法及装置、电子设备和存储介质
WO2022022350A1 (zh) 图像处理方法及装置、电子设备、存储介质和计算机程序产品
CN113034491B (zh) 一种冠脉钙化斑块检测方法及装置
CN112927239A (zh) 图像处理方法、装置、电子设备及存储介质
JP2022548453A (ja) 画像分割方法及び装置、電子デバイス並びに記憶媒体
WO2023050690A1 (zh) 图像处理方法、装置、电子设备、存储介质和程序
CN112597944A (zh) 关键点检测方法及装置、电子设备和存储介质
CN116843647A (zh) 肺野面积确定、肺发育评估方法及装置、电子设备和介质
CN113902730A (zh) 图像处理和神经网络训练方法及装置
CN112686867A (zh) 医学图像的识别方法及装置、电子设备和存储介质
CN112200820A (zh) 三维图像处理方法及装置、电子设备和存储介质
CN116993678A (zh) 肺粘连检测、分析方法及装置、电子设备和存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22874094

Country of ref document: EP

Kind code of ref document: A1