WO2022236995A1

WO2022236995A1 - Guided detection and scoring method for calcified plaque, and device and storage medium

Info

Publication number: WO2022236995A1
Application number: PCT/CN2021/112607
Authority: WO
Inventors: 朱锐; 张逸群; 鲁全茂
Original assignee: 深圳市中科微光医疗器械技术有限公司
Priority date: 2021-05-12
Filing date: 2021-08-13
Publication date: 2022-11-17
Also published as: CN115345811A

Abstract

A guided detection and scoring method for calcified plaque, and a device and a storage medium, and same are applicable to the technical field of medical imaging. The method comprises: acquiring an OCT image sequence of a target lumen (S101); inputting the OCT image sequence into a preset segmentation model for processing, and outputting and obtaining a mask image sequence of the OCT image sequence, wherein the mask image sequence is used for identifying a calcified plaque area in the OCT image sequence (S102); and calculating a size parameter of calcified plaque in the OCT image sequence according to the mask image sequence, and determining the severity score of the calcified plaque of the target lumen according to the size parameter (S103). By means of the method, the severity score of calcified plaque can be obtained by means of analyzing an image of the calcified plaque that is collected by OCT.

Description

Method, device and storage medium for detecting and scoring guided calcified plaque

This application claims priority to a Chinese patent application with application number 202110516332.5 filed at the China Patent Office on May 12, 2021, the entire contents of which are incorporated herein by reference.

technical field

The application belongs to the technical field of medical imaging, and in particular relates to a method, device and storage medium for detecting and scoring guided calcified plaques.

Background technique

With the aging of society and the rising incidence of cardiovascular diseases, vascular calcification has become an important issue in the prevention and treatment of cardiovascular diseases. Severe vascular calcification will significantly increase the difficulty and risk of stent surgery. If the doctor can know the severity of the calcified plaque in the vascular lumen in time, he can take reasonable pretreatment measures such as rotational atherectomy and excimer laser coronary atherectomy (Excimer Laser Coronary Atherectomy, ELCA). Address medical issues related to calcified plaque for better outcomes.

technical problem

Due to the low depth of OCT imaging, the calcified plaques in the deep and large areas cannot be completely displayed, which affects the identification of calcified plaques and the lack of technical problems specifically for the identification of OCT calcified plaques.

technical solution

In order to solve the above-mentioned technical problems, the technical solution adopted in the embodiment of the present application is:

In the first aspect, the present application provides a method for detecting and scoring guided calcified plaques, including: acquiring an OCT image sequence of the target lumen; inputting the OCT image sequence into a preset segmentation model for processing, and outputting the OCT image sequence to obtain the A mask image sequence of the image sequence, the mask image sequence is used to identify the calcified plaque area in the OCT image sequence; calculate the size parameter of the calcified plaque in the OCT image sequence according to the mask image sequence, and A severity score for calcified plaque in the target lumen is determined based on the size parameter.

Optionally, the preset segmentation model includes a multi-scale pyramid convolution pooling module and a U-Net model, and the multi-scale pyramid convolution pooling module includes a plurality of first downsampling layers of different scales, the The U-Net model includes a plurality of second downsampling layers, and the plurality of first downsampling layers of different scales are respectively connected to the plurality of second downsampling layers in one-to-one correspondence; the OCT image sequence is respectively input to the among the plurality of first downsampling layers and the plurality of second downsampling layers.

Optionally, the acquiring the OCT image sequence of the target lumen includes: acquiring an original OCT image sequence of the target lumen; performing image compensation on the OCT original image sequence according to an optical tomographic attenuation compensation algorithm to obtain an OCT compensated image sequence; The OCT original image sequence and the OCT compensated image sequence are superimposed to obtain an OCT image sequence of the target lumen.

Optionally, the optical tomography attenuation compensation algorithm is:

Wherein, z represents the imaging depth of the OCT original image sequence, Ii, j(z) represents the initial light intensity value of the pixel point (i, j) in the original image sequence,

is the compensated intensity value, while

is the compensation factor corresponding to Ii,j(z).

Optionally, the size parameter of the calcified plaque in the OCT image sequence includes the thickness, length and angle of the calcified plaque in the OCT image sequence.

Optionally, the mask image sequence is also used to identify the lumen region in the OCT image sequence, and the calculating the size parameter of the calcified plaque in the OCT image sequence according to the mask image sequence includes :

Calculate the central coordinates of the lumen region in the OCT image sequence according to the mask image sequence, and convert the mask image sequence into a square image sequence according to the central coordinates and a preset scanning frequency;

The thickness of the calcified plaque in the OCT image sequence is determined by multiplying the maximum number of pixels whose pixel values are continuously the first preset value in the column direction in the square image sequence multiplied by the pixel resolution;

Calculate the angles of the calcified plaques in the multi-frame square diagrams included in the square diagram sequence according to the formula a=2πr/c, and determine the maximum value of the angles of the calcified plaques in the multi-frame square diagrams as the OCT image The angle of the calcified plaque in the sequence; wherein, c represents the width of the square diagram sequence, and r represents the number of columns in the column direction of the square diagram sequence containing pixels whose pixel value is the first preset value, a represents the angle of calcified plaque;

The length of the calcified plaque in the OCT image sequence is calculated according to the formula l=p*m, wherein, p represents the number of frames in which the calcified region continuously appears in the square image sequence, and m represents the frame interval.

Optionally, the determining the severity score of the calcified plaque of the target lumen according to the size parameter includes: when the angle of the calcified plaque is within a first threshold range, determining the severity score of the target lumen The first severity score of the calcified plaque, when the angle of the calcified plaque is within the second threshold range, determine the second severity score of the calcified plaque in the target lumen; the thickness of the calcified plaque is within When within the third threshold range, determine the third severity score of the calcified plaque in the target lumen; when the length of the calcified plaque is within the fourth threshold range, determine the score of the calcified plaque in the target lumen The fourth severity score; according to the first severity score, the second severity score, the third severity score and the fourth severity score, the calcified plaque in the target lumen is obtained Severity score.

Optionally, the method further includes: training the initial segmentation model according to a preset loss function and a training set to obtain a preset segmentation model; wherein, the training set includes a plurality of OCT image sequence samples and each The mask image sequence samples corresponding to the OCT image sequence samples; the loss function is used to constrain the error between each frame image sample in the OCT image sequence and each corresponding mask image sequence sample, and also It is used to constrain the error between the continuous multi-frame image samples in the OCT image sequence and the corresponding continuous multi-frame mask image sequence samples.

In a second aspect, the present application provides a guided calcified plaque detection and scoring device, including:

an acquisition unit, configured to acquire an OCT image sequence of the target lumen;

A processing unit, configured to input the OCT image sequence into a preset segmentation model for processing, and output a mask image sequence to obtain the OCT image sequence, and the mask image sequence is used to identify calcification in the OCT image sequence plaque area;

A determining unit is configured to calculate a size parameter of the calcified plaque in the OCT image sequence according to the mask image sequence, and determine a severity score of the calcified plaque in the target lumen according to the size parameter.

In a third aspect, the present application provides a terminal device, including a memory, a processor, and a computer program stored in the memory and operable on the processor. When the processor executes the computer program, any of the first aspect or the first aspect can be realized. A method described in an optional manner.

In a fourth aspect, a computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in the first aspect or any optional mode of the first aspect is implemented.

In a fifth aspect, an embodiment of the present application provides a computer program product, which, when the computer program product is run on a terminal device, enables the terminal device to execute the method described in the first aspect or any optional manner of the first aspect.

Beneficial effect

Through the detection and scoring method of guided calcified plaques provided by this application, the optical tomographic attenuation compensation algorithm is firstly used to obtain the OCT compensation image sequence corresponding to the OCT original image sequence, which can avoid the failure of the OCT original image due to the low imaging depth. The calcified plaques in the deep and larger areas are completely displayed, so as to obtain the complete deep calcified plaques; then the preset segmentation model composed of the multi-scale pyramid convolution pooling module and the U-Net model is used to make the calcification OCT images with large size differences are input into the preset segmentation model and processed to obtain feature maps with fixed feature scales, which avoids large differences in feature scales and small-scale features in feature maps directly processed by the U-Net model. signal attenuation; finally, the size parameter of the calcified plaque in the OCT image sequence is calculated to determine the severity score of the calcified plaque in the target lumen. Therefore, the image of the calcified plaque collected by OCT can be analyzed to obtain the severity of the calcified plaque through the plaque calcification detection and scoring method provided in the present application.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the accompanying drawings that need to be used in the descriptions of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings in the following description are only for the present application For some embodiments, those of ordinary skill in the art can also obtain other drawings based on these drawings without paying creative efforts.

Fig. 1 is a schematic flowchart of an embodiment of a method for detecting calcified plaque provided by an embodiment of the present application;

Fig. 2 is a schematic diagram of an OCT calcification circle diagram provided by an embodiment of the present application;

FIG. 3 is a schematic diagram of a compensated OCT calcification circle diagram provided by an embodiment of the present application;

FIG. 4 is a schematic diagram of a network structure of a preset segmentation model provided by an embodiment of the present application;

Fig. 5 is a schematic diagram of a mask image output by a segmentation model provided by an embodiment of the present application;

Fig. 6 is a schematic diagram of the recognition result of an OCT calcification square diagram provided by an embodiment of the present application;

Fig. 7 is a schematic diagram of converting a mask image into a block diagram provided by an embodiment of the present application;

Fig. 8 is a schematic diagram of a guided calcified plaque detection and scoring device provided in the embodiment of the present application;

FIG. 9 is a schematic diagram of a terminal device provided by an embodiment of the present application.

Embodiments of the present invention

Due to the low depth of OCT imaging, the calcified plaques in the deep and large areas cannot be completely displayed, which affects the identification of calcified plaques and the lack of technical problems specifically for the identification of OCT calcified plaques. This application provides a method for detecting and scoring guided calcified plaques. First, the optical tomography attenuation compensation algorithm is used to obtain the OCT compensated calcification circle map corresponding to the OCT calcification circle map, that is, the enhanced calcification circle map; and then through the spatial multi-scale pyramid pooling Methods The U-Net image segmentation model was improved to obtain mask images; finally, the size parameters of calcified plaques were calculated according to the mask images, and the severity of calcified plaques was determined according to the size parameters. Therefore, the severity of the calcified plaque can be obtained by analyzing the image of the calcified plaque collected by OCT.

The technical solution of the present application will be described in detail below with specific embodiments. The following specific embodiments may be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

FIG. 1 is a schematic flowchart of a guided calcified plaque detection and scoring method provided by an embodiment of the present application. The subject of execution of the method may be an optical coherence tomography (optical coherence tomography, OCT) image acquisition device, or other control devices associated with the OCT image acquisition device, such as a personal computer (PC). Referring to Fig. 1, the method includes: S101, acquiring an OCT image sequence of a target lumen.

It should be noted that the OCT image sequence of the target lumen refers to multiple OCT images in the lumen of a blood vessel acquired by an OCT acquisition device. It is not difficult to understand that the OCT image sequence may be directly and continuously collected multiple frames of OCT images, or may be continuously collected multiple frames of OCT images in a video.

Wherein, acquiring the OCT image sequence of the target lumen includes: S1011, acquiring the original OCT image sequence of the target lumen.

In a possible implementation manner, the acquisition of the original OCT image sequence may be to export the acquired continuous multi-frame OCT images from the OCT image acquisition device, and then manually select the images containing calcification to form the original OCT image sequence.

Initially process each OCT original image in the above OCT original image sequence into an image of 704*704*3. It is worth noting that the initialization refers to using the initialization function to convert all the OCT images in the OCT original image sequence into images with a size of 704*704*3.

S1012. Perform image compensation on the OCT original image sequence according to an optical tomography attenuation compensation algorithm to obtain an OCT compensated image sequence.

The multiple OCT original images in the acquired OCT original image sequence use infrared light waves with a wavelength of about 1300nm as the light source, and the light emitted by the light source is divided into a sample beam and a reference beam through a beam splitter, and the reference beam and the sample at the same distance are used The optical coherence phenomenon generated after the meeting of the beam reflected waves is processed by a computer into a signal to obtain a tissue image. Among them, due to the low imaging depth of OCT, generally the imaging depth of OCT is 1-3 mm, which cannot completely display images of calcified plaques in deeper and larger areas.

In order to obtain complete images of calcified plaques in deeper and larger areas, this application uses the Attenuation-Compensated Optical Coherence Tomography (AC) algorithm to perform image compensation on the original OCT image sequence to obtain the OCT compensated image sequence.

It is worth noting that the optical tomography attenuation compensation algorithm can be calculated according to the following formula:

Among them, in the formula (1), z represents the imaging depth of the original OCT image sequence, Ii,j(z) represents the initial light intensity value of the pixel point (i, j) in the original OCT image sequence,

is the intensity value of the pixel in the OCT original image sequence after compensation for (i, j), and

is the compensation factor corresponding to Ii,j(z).

It is not difficult to understand that the OCT original image sequence contains multiple OCT original images, and the multiple OCT original images are substituted into the formula (1) to calculate the multiple OCT compensation images corresponding to the multiple OCT original images one-to-one. The OCT compensated images constitute an OCT compensated image sequence that corresponds one-to-one to the above OCT original image sequence.

S1013. Superimpose the OCT original image sequence and the OCT compensation image sequence to obtain an OCT image sequence of the target lumen.

Wherein, one OCT original image in the OCT original image sequence is shown in FIG. 2 , and the corresponding OCT compensated image is shown in FIG. 3 .

It should be noted that since the OCT image sequence of the target lumen is an image formed by superimposing the OCT original image sequence on the OCT compensation image sequence, the size of the OCT original image in the OCT original image sequence is 704*704*3, and the optical tomography The OCT compensated image sequence processed by the attenuation compensation algorithm does not change the size of the original OCT image sequence, that is, the size of the OCT compensated image sequence is also 704*704*3. Therefore, the size of the OCT images in the OCT image sequence formed by superimposing the OCT original image sequence with the OCT compensation image sequence is 704*704*6.

S102, input the OCT image sequence into a preset segmentation model for processing, and output a mask image sequence of the OCT image sequence, the mask image sequence is used to identify the calcified plaque area in the OCT image sequence .

Since the sizes of lumens and calcifications in different OCT images in the OCT image sequence vary greatly, when the OCT images with large size differences are directly input into the U-Net model for convolution, the feature scale in the feature map obtained is The difference is also large, and even small-scale characteristic signal attenuation may occur, making it difficult to extract features of lumens and plaques in OCT images.

Therefore, the preset segmentation model structure used in the embodiment of the present application is shown in Figure 4, wherein each square represents the feature map after convolution, the number under each square represents the number of feature maps, and the input image refers to The OCT image sequence to be input into the preset segmentation model, the output image refers to the mask image sequence with the numbers 0, 1 and 2 respectively marking the background, lumen and calcified plaque.

The preset segmentation model in the embodiment of this application includes the U-Net model and the multi-scale pyramid convolution pooling module. Among them, the multi-scale pyramid convolution pooling module includes multiple first down-sampling layers of different scales, and the outputs of multiple first down-sampling layers of different scales are respectively connected to multiple second down-sampling layers of the U-Net model; The OCT image sequence is input to the multi-scale pyramid convolution pooling module and the U-Net model respectively. It should be understood that multiple first downsampling layers of different scales are respectively connected to multiple second downsampling layers in one-to-one correspondence; multiple frames of OCT images contained in the OCT image sequence are respectively input to multiple first downsampling layers of different scales in one-to-one correspondence. In the sampling layer, and input the OCT image sequence into a plurality of second down-sampling layers.

After the OCT image sequence input in this way is processed by the preset segmentation model, the mask image sequence corresponding to the above OCT image sequence is output, and the mask image in the output mask image sequence corresponding to Fig. 2 and Fig. 3 is shown in Fig. 5 .

It should be noted that after normalization by the softmax function, the classification score corresponding to each pixel in the above mask image can be obtained, and the sum of the probabilities of each pixel belonging to these three categories is 1. Select the probability The largest is the classification category corresponding to the pixel. Exemplarily, 0 represents the background in the OCT calcification circle diagram, 1 represents the lumen in the OCT calcification circle diagram, and 2 represents the calcified plaque in the OCT calcification circle diagram. If one of the pixel points A corresponds to the above 3 In the first classification, the probability after normalization processing is (0.2, 0.2, 0.6), and the pixel A belongs to the third category, that is, the probability of the calcified plaque in the OCT calcification circle is the highest, then the pixel A belongs to the calcified plaque .

In addition, the training process of the preset segmentation model in the embodiment of the present application is: according to the preset loss function and training set, the initial segmentation model is trained to obtain the segmentation model.

Wherein, the training set includes a plurality of OCT image sequence samples and a mask image sequence sample corresponding to each of the OCT image sequence samples.

It is not difficult to understand that each OCT image in the above-mentioned plurality of OCT image sequence samples is marked manually (for example, an expert), and the calcification outline, the lumen outline and the background are respectively marked. The marked OCT image can be used to train the preset segmentation model in this application, and can also be used as a real result to optimize the preset segmentation model.

In the actual operation process, for one of the OCT images, after the calcification contour and lumen contour have been marked, the rest of the above OCT image (unmarked part) is the background in the image, therefore, the actual What needs to be marked is the outline of the calcification and the outline of the lumen.

Exemplarily, 0 represents the background in the OCT image, 1 represents the lumen in the OCT image, and 2 represents the calcified plaque in the OCT image. The classification category corresponding to the specific labeled data may be adjusted according to actual needs, and this application does not make any limitation thereto.

Since the acquisition and labeling of medical images are difficult, multiple OCT images in multiple OCT image sequence samples can be expanded by rotating, flipping, adding noise, translation, zooming, or cropping to increase multiple Amount of data in a sample of an OCT image sequence. The increased amount of data is also of great significance to the training of the subsequent preset segmentation model, for example, memory training data can be avoided.

In order to continuously optimize the parameters in the initial segmentation model, further reduce the error between the model prediction result and the real result, and make the prediction model prediction result in the embodiment of the present application more accurate, the following loss function is used:

Loss＝L(p _t ,g _t )+||L(p _t ,p _t-1 )-L(g _t ,g _t-1 )|

Among them, pt represents the prediction result of the t-th frame image model, gt represents the real result of the t-th frame image model; L(pt,gt) represents the error between the t-th frame image model’s prediction result and the real result; L(pt ,pt-1) represents the error between the prediction result of the t-th frame image model and the prediction result of the t-1-th frame image model; L(gt,gt-1) represents the difference between the real result of the t-th frame and the t-1-th frame The error between the true results.

It is not difficult to understand that the real result refers to the result of experts labeling the OCT image sequence samples containing calcification. When the error between the results predicted by the model and the results marked by experts gradually decreases, it means that the results predicted by the model are more accurate. Figure 6 shows the OCT image recognized by the optimized segmentation model.

The loss function is used to constrain the error between each frame image sample in the OCT image sequence and the corresponding frame mask image sequence sample, and is also used to constrain the continuous multi-frame image samples in the OCT image sequence and the corresponding continuous multi-frame mask Error between samples of the membrane image sequence.

It is worth noting that the execution subject of the above-mentioned preset segmentation model may be the same device as the terminal device running the segmentation model, or may be other computer devices, which is not limited in this application.

S103. Calculate a size parameter of the calcified plaque in the OCT image sequence according to the mask image sequence, and determine a severity score of the calcified plaque in the target lumen according to the size parameter.

Optionally, the size parameters of the calcified plaque in the embodiment of the present application include the thickness, length and angle of the calcified plaque. The size parameters of the calcified plaque can be designed according to different actual needs, which is not limited in this application.

It should be noted that the mask image sequence of the OCT image sequence is also used to identify the lumen region in the OCT image sequence. In order to facilitate the calculation, the mask image is converted into a square diagram. The conversion process is: calculate the center coordinates of the lumen area according to the edge coordinate points of the lumen area, and use the center coordinates of the lumen area and the preset scanning frequency Convert the mask image sequence of the OCT image sequence to a square map sequence. Exemplarily, FIG. 7 is a converted square diagram corresponding to the mask image in FIG. 5 . The coordinates of the center point of the lumen area are calculated by extracting nearly a hundred coordinate points from the edge of the lumen in the mask image, and taking the average value of the hundreds of coordinate points to calculate the coordinates of the center point of the lumen area.

It is worth noting that the preset scanning frequency in the embodiment of the present application refers to the scanning frequency of the OCT acquisition device, and the size of the conversion square diagram is determined by the preset scanning frequency. Wherein, the number of columns of the converted square image is determined by the horizontal resolution of the OCT acquisition device, and the number of rows of the converted square image is determined by the vertical resolution of the OCT acquisition device.

Exemplarily, when the horizontal resolution of the OCT acquisition device is 500, the number of columns corresponding to the collected OCT image is 500; when the vertical resolution of the OCT acquisition device is 700, the number of rows corresponding to the collected OCT image is 700, However, in the embodiment of the present application, the number of lines is generally 642, because the data after 642 lines carries less effective information, which can be ignored. Certainly, the converted square diagram may also be adjusted according to different actual requirements, and this application does not make any limitation thereto.

The angle, thickness and length of the calcified plaque were calculated according to the transformed square diagram.

Optionally, the thickness of the calcified plaque in the OCT image sequence is determined by multiplying the pixel resolution by the maximum number of pixels whose pixel values are continuously equal to the first preset value in the column direction in the square diagram sequence.

Wherein, the pixel resolution represents the size of a pixel, and the size of a pixel is related to the device used to collect the OCT image. The first preset value refers to the classification category used to identify the calcified plaque during the labeling process of the OCT original image sequence. Exemplarily, if the number 1 is used to represent the calcified plaque, then the first preset value is 1, and the thickness of the calcified plaque is the maximum number of pixels whose pixel values are continuously 1 in the column direction in the square map sequence multiplied by the pixel resolution Rate.

Optionally, calculate the angles of the calcified plaques in the multi-frame square diagrams included in the square diagram sequence respectively according to formula (2), and determine the maximum value of the angles of the calcified plaques in the multi-frame square diagrams as the calcified plaques in the OCT image sequence Angle. For the calculation formula of the angle of the calcified area, please refer to the following formula:

a＝2π r/c (2)

In the formula (2), c represents the width of the square diagram sequence, r represents the number of columns in the column direction of the square diagram sequence containing pixels whose pixel value is the first preset value, and a represents the angle of the calcified plaque .

Exemplarily, when the first preset value is 1, r takes the number of columns containing pixels with a pixel value of 1 in the column direction in the square map sequence, and c takes the width of the square map sequence, and puts it into the formula a=2π· In r/c, the angle of the calcified plaque contained in each square image in the square image sequence is calculated, and the maximum angle value of the calcified plaque is the angle of the calcified plaque. As another example, when the square diagram sequence contains 3 frames of square diagrams, the angle of calcified plaque in the first frame of square diagram is 30°, the angle of calcified plaque in the second frame of square diagram is 45°, and the angle of calcified plaque in the third frame of square diagram is The angle of the plaque is 60°, then the angle of the finally determined calcified plaque is 60°.

Optionally, the length of the calcified plaque in the OCT image sequence is calculated according to formula (3), and the calculation formula for the length of the calcified region is shown in the following formula:

l=p*m (3)

In formula (3), p represents the number of frames in which calcified areas appear continuously in the square image sequence, and m represents the frame spacing between each image.

It is worth noting that the overlapping parts of calcified areas appearing in continuous areas in multiple images are regarded as the same calcified area. For the same calcified area, take the angle value of the largest angle value in all images as the angle value of the calcified area, and similarly, take the largest thickness value in all images as the thickness value of the calcified area.

After calculating the angle, thickness and length of the calcified area, the severity score of the calcified plaque was obtained according to the scoring indicators of the calcified plaque shown in Table 1.

According to Table 1, the severity score of calcified plaque can be calculated correspondingly. Because the severity of calcified plaque will affect the treatment strategy of Percutaneous Trans-luminal Coronary Intervention (PCI), generally severe calcified plaque will cause malapposition and insufficiency after PCI. Therefore, the severity score is 0-4, and the higher the severity score, the more serious the severity of the calcified plaque. Doctors can take different pretreatment measures to treat the calcified plaque of the lesion according to the score of the severity score. Among them, the pretreatment measures include but are not limited to cutting balloon, calcification rotational atherectomy, shock wave therapy, laser ablation, etc.

According to clinical experience, when the severity score is 0-3, the probability of poor stent expansion is low, and pretreatment measures may not be taken; when the severity score is 4, doctors need to take pretreatment measures to treat the calcified plaque of the lesion.

Table 1

Based on the same inventive concept, as the implementation of the above method, the embodiment of the present application provides a guided calcified plaque detection and scoring device. The embodiment of the device corresponds to the aforementioned method embodiment. The details in the foregoing method embodiments are described one by one, but it should be clear that the device in this embodiment can correspondingly implement all the content in the foregoing method embodiments.

As shown in Figure 8, the present application provides a guided calcified plaque detection and scoring device, including:

An acquisition unit 801, configured to acquire an OCT image sequence of a target lumen.

The processing unit 802 is configured to input the OCT image sequence into a preset segmentation model for processing, and output a mask image sequence of the OCT image sequence, and the mask image sequence of the OCT image sequence is used to identify the OCT image sequence. Areas of calcified plaque in the image sequence.

The determination unit 803 calculates a size parameter of the calcified plaque in the OCT image sequence according to the mask image sequence, and determines a severity score of the calcified plaque in the target lumen according to the size parameter.

Optionally, the preset segmentation model includes a multi-scale pyramid convolution pooling module and a U-Net model, the multi-scale pyramid convolution pooling module includes multiple first downsampling layers of different scales, and the U-Net model includes multiple A second down-sampling layer, a plurality of first down-sampling layers of different scales are respectively connected to a plurality of second down-sampling layers in one-to-one correspondence; OCT image sequences are respectively input into a plurality of first down-sampling layers and a plurality of second down-sampling layers layer.

Optionally, obtaining an OCT image sequence of the target lumen includes:

Obtain the original OCT image sequence of the target lumen;

Perform image compensation on the OCT original image sequence according to the optical tomography attenuation compensation algorithm to obtain the OCT compensation image sequence;

The OCT original image sequence and the OCT compensated image sequence are superimposed to obtain the OCT image sequence of the target lumen.

Optionally, the optical tomography attenuation compensation algorithm is:

Among them, z represents the imaging depth of the original OCT image sequence, Ii,j(z) represents the initial light intensity value of the pixel point (i, j) in the original image sequence,

is the compensated intensity value, while

is the compensation factor corresponding to Ii,j(z).

Optionally, the mask image sequence is also used to identify the lumen area in the OCT image sequence, and the size parameters of the calcified plaque in the OCT image sequence are calculated according to the mask image sequence, including:

Calculate the center coordinates of the lumen region in the OCT image sequence according to the mask image sequence, and convert the mask image sequence into a square image sequence according to the center coordinates and the preset scanning frequency;

Determine the thickness of the calcified plaque in the OCT image sequence according to the maximum number of pixels whose pixel values are continuously the first preset value in the square image sequence in the column direction multiplied by the pixel resolution;

Calculate the angles of calcified plaques in the multi-frame square diagrams contained in the square diagram sequence according to the formula a=2πr/c, and determine the maximum value of the angles of the calcified plaques in the multi-frame square diagrams as the calcified plaques in the OCT image sequence Wherein, c represents the width of the square diagram sequence, r represents the column number of the pixel point containing the pixel value of the first preset value in the column direction of the square diagram sequence, and a represents the angle of the calcified plaque;

The length of the calcified plaque in the OCT image sequence is calculated according to the formula l=p*m, where p represents the number of frames in which the calcified region continuously appears in the square image sequence, and m represents the frame interval.

Optionally, the method also includes:

According to the preset loss function and training set, the initial segmentation model is trained to obtain the preset segmentation model;

Wherein, the training set includes a plurality of OCT image sequence samples and mask image sequence samples corresponding to each OCT image sequence sample;

The segmentation model device provided in this embodiment can execute the above-mentioned method embodiment, and its implementation principle and technical effect are similar, and will not be repeated here.

Those skilled in the art can clearly understand that for the convenience and brevity of description, only the division of the above-mentioned functional units and modules is used for illustration. In practical applications, the above-mentioned functions can be assigned to different functional units, Completion of modules means that the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment may be integrated into one processing unit, or each unit may exist separately physically, or two or more units may be integrated into one unit, and the above-mentioned integrated units may adopt hardware It can also be implemented in the form of software functional units. In addition, the specific names of each functional unit and module are only for the convenience of distinguishing each other, and are not used to limit the protection scope of the present application. For the specific working process of the units and modules in the above system, reference may be made to the corresponding process in the foregoing method embodiments, and details will not be repeated here.

Based on the same inventive concept, the embodiment of the present application also provides a terminal device. FIG. 9 is a schematic diagram of a terminal device provided in an embodiment of the present application. As shown in FIG. 9 , the terminal device provided in this embodiment includes: a memory 901 and a processor 902, the memory 901 is used to store computer programs; the processor 902 is used to When the computer program is called, the methods described in the above method embodiments are executed, such as steps S101 to S103 shown in FIG. 1 . Or, when the processor 902 executes the computer program, it realizes the functions of the modules/units in the above-mentioned device embodiments, for example, the functions of the units 801 to 803 shown in FIG. 8 .

Exemplarily, the computer program may be divided into one or more modules/units, and the one or more modules/units are stored in the memory 901 and executed by the processor 902 to complete this Application. The one or more modules/units may be a series of computer program instruction segments capable of accomplishing specific functions, and the instruction segments are used to describe the execution process of the computer program in the terminal device.

Those skilled in the art can understand that FIG. 9 is only an example of a terminal device, and does not constitute a limitation to the terminal device. It may include more or less components than those shown in the figure, or combine certain components, or different components, such as The terminal device may also include an input and output device, a network access device, a bus, and the like.

The processor 902 can be a central processing unit (Central Processing Unit, CPU), and can also be other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), Field-Programmable Gate Array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor, or the processor may be any conventional processor, or the like.

The storage 901 may be an internal storage unit of the terminal device, for example, a hard disk or a memory of the terminal device. The memory 901 may also be an external storage device of the terminal device, such as a plug-in hard disk equipped on the terminal device, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, Flash card (Flash Card), etc. Further, the memory 901 may also include both an internal storage unit of the terminal device and an external storage device. The memory 901 is used to store the computer program and other programs and data required by the terminal device. The memory 901 can also be used to temporarily store data that has been output or will be output.

The terminal device provided in this embodiment can execute the foregoing method embodiment, and its implementation principle and technical effect are similar, and details are not repeated here.

The embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method described in the foregoing method embodiment is implemented.

The embodiment of the present application further provides a computer program product, which, when the computer program product runs on a terminal device, enables the terminal device to implement the method described in the foregoing method embodiments when executed.

An embodiment of the present application further provides a chip system, including a processor, the processor is coupled to a memory, and the processor executes a computer program stored in the memory, so as to implement the method described in the above method embodiment. Wherein, the chip system may be a single chip, or a chip module composed of multiple chips.

If the above integrated units are realized in the form of software function units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the procedures in the methods of the above embodiments in the present application can be completed by instructing related hardware through computer programs, and the computer programs can be stored in a computer-readable storage medium. The computer program When executed by a processor, the steps in the above-mentioned various method embodiments can be realized. Wherein, the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file or some intermediate form. The computer-readable storage medium may at least include: any entity or device capable of carrying computer program codes to a photographing device/terminal device, a recording medium, a computer memory, a read-only memory (Read-Only Memory, ROM), a random access Memory (Random Access Memory, RAM), electrical carrier signal, telecommunication signal and software distribution medium. Such as U disk, mobile hard disk, magnetic disk or optical disk, etc. In some jurisdictions, computer readable media may not be electrical carrier signals and telecommunication signals under legislation and patent practice.

In the above-mentioned embodiments, the descriptions of each embodiment have their own emphases, and for parts that are not detailed or recorded in a certain embodiment, refer to the relevant descriptions of other embodiments.

Those skilled in the art can appreciate that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed by hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art may use different methods to implement the described functions for each specific application, but such implementation should not be regarded as exceeding the scope of the present application.

In the embodiments provided in this application, it should be understood that the disclosed device/device and method can be implemented in other ways. For example, the device/device embodiments described above are only illustrative. For example, the division of the modules or units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or Components may be combined or integrated into another system, or some features may be omitted, or not implemented. In another point, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be in electrical, mechanical or other forms.

It should be understood that when used in this specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and/or components, but does not exclude one or more other Presence or addition of features, wholes, steps, operations, elements, components and/or collections thereof.

It should also be understood that the term "and/or" used in the description of the present application and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

As used in this specification and the appended claims, the term "if" may be construed, depending on the context, as "when" or "once" or "in response to determining" or "in response to detecting ". Similarly, the phrase "if determined" or "if [the described condition or event] is detected" may be construed, depending on the context, to mean "once determined" or "in response to the determination" or "once detected [the described condition or event] ]” or “in response to detection of [described condition or event]”.

In addition, in the description of the specification and appended claims of the present application, the terms "first", "second", "third" and so on are only used to distinguish descriptions, and should not be understood as indicating or implying relative importance.

Reference to "one embodiment" or "some embodiments" or the like in the specification of the present application means that a particular feature, structure, or characteristic described in connection with the embodiment is included in one or more embodiments of the present application. Thus, appearances of the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in other embodiments," etc. in various places in this specification are not necessarily All refer to the same embodiment, but mean "one or more but not all embodiments" unless specifically stated otherwise. The terms "including", "comprising", "having" and variations thereof mean "including but not limited to", unless specifically stated otherwise.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit it; although the application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: It is still possible to modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the technical solutions of the various embodiments of the present application. scope.

Claims

A method for detecting and scoring guided calcified plaques, comprising:

Obtain an OCT image sequence of the target lumen;

Inputting the OCT image sequence into a preset segmentation model for processing, and outputting a mask image sequence of the OCT image sequence, the mask image sequence is used to identify the calcified plaque area in the OCT image sequence;

calculating a size parameter of the calcified plaque in the OCT image sequence according to the mask image sequence, and determining a severity score of the calcified plaque in the target lumen according to the size parameter.
The method according to claim 1, wherein the preset segmentation model includes a multi-scale pyramid convolution pooling module and a U-Net model, and the multi-scale pyramid convolution pooling module includes a plurality of different scales The first down-sampling layer of the U-Net model includes a plurality of second down-sampling layers, and the plurality of first down-sampling layers of different scales are respectively connected in one-to-one correspondence with the plurality of second down-sampling layers; The OCT image sequence is respectively input into the plurality of first downsampling layers and the plurality of second downsampling layers.
The method according to claim 1, wherein said obtaining the OCT image sequence of the target lumen comprises:

Obtain the original OCT image sequence of the target lumen;

performing image compensation on the OCT original image sequence according to an optical tomography attenuation compensation algorithm to obtain an OCT compensated image sequence;

The OCT original image sequence and the OCT compensated image sequence are superimposed to obtain an OCT image sequence of the target lumen.
The method according to claim 3, wherein the optical tomography attenuation compensation algorithm is:

Wherein, z represents the imaging depth of the OCT original image sequence, I i, j (z) represents the initial light intensity value of the pixel point (i, j) in the original image sequence,
is the compensated intensity value, while
is the compensation factor corresponding to I i,j (z).
The method according to claim 1, wherein the size parameter of the calcified plaque in the OCT image sequence comprises the thickness, length and angle of the calcified plaque in the OCT image sequence.
The method according to claim 5, wherein the mask image sequence is also used to identify the lumen region in the OCT image sequence, and the calculation of the OCT image sequence according to the mask image sequence The size parameters of the calcified plaque include:

Calculate the central coordinates of the lumen region in the OCT image sequence according to the mask image sequence, and convert the mask image sequence into a square image sequence according to the central coordinates and a preset scanning frequency;

Determine the thickness of the calcified plaque in the OCT image sequence according to the maximum number of pixels whose pixel values are continuously the first preset value in the column direction in the square image sequence multiplied by the pixel resolution;

Calculate the angles of the calcified plaques in the multi-frame square diagrams included in the square diagram sequence according to the formula a=2πr/c, and determine the maximum value of the angles of the calcified plaques in the multi-frame square diagrams as the OCT image The angle of the calcified plaque in the sequence; wherein, c represents the width of the square diagram sequence, and r represents the number of columns in the column direction of the square diagram sequence containing pixels whose pixel value is the first preset value, a represents the angle of calcified plaque;

The length of the calcified plaque in the OCT image sequence is calculated according to the formula l=p*m, wherein, p represents the number of frames in which the calcified region continuously appears in the square image sequence, and m represents the frame interval.
The method according to claim 5 or 6, wherein the determining the severity score of the calcified plaque in the target lumen according to the size parameter comprises:

When the angle of the calcified plaque is within a first threshold range, determine the first severity score of the calcified plaque in the target lumen, and when the angle of the calcified plaque is within a second threshold range, determine the A second severity score for calcified plaque in the target lumen;

When the thickness of the calcified plaque is within a third threshold range, determine a third severity score of the calcified plaque in the target lumen;

When the length of the calcified plaque is within a fourth threshold range, determine a fourth severity score of the calcified plaque in the target lumen;

According to the first severity score, the second severity score, the third severity score and the fourth severity score, the severity score of the calcified plaque in the target lumen is obtained.
The method of claim 1, further comprising:

According to the preset loss function and training set, the initial segmentation model is trained to obtain the preset segmentation model;

Wherein, the training set includes a plurality of OCT image sequence samples and mask image sequence samples corresponding to each of the OCT image sequence samples;

The loss function is used to constrain the error between each frame of image samples in the OCT image sequence and each corresponding frame of mask image sequence samples, and is also used to constrain the relationship between consecutive multiple frames of image samples in the OCT image sequence The error between corresponding consecutive multi-frame mask image sequence samples.
A device for detecting and scoring guided calcified plaque, characterized in that it includes:

an acquisition unit, configured to acquire an OCT image sequence of the target lumen;

A processing unit, configured to input the OCT image sequence into a preset segmentation model for processing, and output a mask image sequence to obtain the OCT image sequence, and the mask image sequence is used to identify calcification in the OCT image sequence plaque area;

A determining unit is configured to calculate a size parameter of the calcified plaque in the OCT image sequence according to the mask image sequence, and determine a severity score of the calcified plaque in the target lumen according to the size parameter.
A terminal device, comprising a memory, a processor, and a computer program stored in the memory and operable on the processor, characterized in that, when the processor executes the computer program, the following claims 1 to 1 are implemented. 8. The method described in any one.
A computer-readable storage medium storing a computer program, wherein the computer program implements the method according to any one of claims 1 to 8 when executed by a processor.
A computer program product, characterized in that, when the computer program product is run on a terminal device, the terminal device is made to execute the method according to any one of claims 1 to 8.