CN115423751A - Image processing method and device, electronic equipment and storage medium - Google Patents

Abstract

This application is applicable to the technical field of medical image processing, and provides an image processing method, device, electronic equipment, and storage medium. The method includes: during the process of acquiring an optical coherence tomography OCT image group for the object of investigation, acquiring Aiming at the angiography image group of the detection object; registering the OCT image group and each angiography image in the angiography image group to obtain registration parameters; based on each OCT image in the OCT image group, generating the light attenuation corresponding to each OCT image coefficient image group; calculate the plaque attenuation index IPA value of each light attenuation coefficient image in the light attenuation coefficient image group; based on the registration parameters, use the IPA value to mark each angiography image in the angiography image group to obtain the target blood vessel Contrast image set. Through the method provided by the embodiment of the present application, the position and conformity of vulnerable plaques can be observed clearly and intuitively, and the ability of identifying vulnerable plaques of detection objects is improved.

Description

Image processing method, device, electronic device and storage medium

This application is a divisional application of a Chinese patent application submitted to the China Patent Office on July 13, 2021, with application number 202110790290.4, and titled "Image Processing Method, Device, Electronic Equipment, and Storage Medium".

technical field

The present application belongs to the technical field of medical image processing, and in particular relates to an image processing method, device, electronic equipment and storage medium.

Background technique

Optical coherence tomography (OCT) is an imaging technique. It uses the basic principle of weak coherent light interferometer to divide the light emitted by the light source into two beams, one beam is sent to the measured tissue, also called the sample arm, and the other beam is sent to the reference mirror, also called the reference arm, and then the The measured tissue and the two beams of light signals reflected from the reference mirror are superimposed and interfered, and finally according to the light signal with the measured tissue, different intensity image gray levels are displayed, so as to image the tissue.

The existing traditional optical coherent image technology is weak in identifying vulnerable plaques, and it is difficult to improve the identification ability from the system and equipment side, which will also incur high costs; in addition, the display effect of existing OCT images is poor, and it is not It is convenient to assist medical staff in judging the load of vulnerable plaques.

Contents of the invention

Embodiments of the present application provide an image processing method, device, electronic device, and storage medium, which can solve at least part of the above problems.

In the first aspect, the embodiment of the present application provides a method for image processing, including:

In the process of acquiring the optical coherence tomography OCT image group for the exploration object, acquiring the angiography image group for the investigation object;

Registering each angiography image in the OCT image group and the angiography image group to obtain registration parameters;

Based on each OCT image in the OCT image group, generate a plaque attenuation coefficient light attenuation coefficient image group corresponding to each OCT image;

calculating the IPA value of each of the light attenuation coefficient images in the light attenuation coefficient image group;

Based on the registration parameters, each angiography image in the angiography image group is marked with the IPA value to obtain a target angiography image group.

It should be understood that by registering each angiographic image in the OCT image group and the angiographic image group, the registration parameters between the OCT image group and each angiographic image, that is, the corresponding relationship, make the light attenuation obtained by using each OCT image The coefficient image has a consistent correspondence with each angiographic image. On this basis, the IPA value obtained based on the light attenuation coefficient image is marked on each angiographic image, so that the position and compliance of the vulnerable plaque can be clearly and intuitively observed, and the ability to identify the vulnerable plaque of the detection object is improved. .

In the second aspect, the embodiment of the present application provides an image processing device, including:

An image acquisition module, configured to obtain an angiography image group for the investigation object during the process of obtaining the optical coherence tomography OCT image group for the investigation object;

An image registration module, configured to register each angiography image in the OCT image group and the angiography image group to obtain registration parameters;

An optical attenuation coefficient image generating module, configured to generate a plaque attenuation coefficient optical attenuation coefficient image group corresponding to each OCT image based on each OCT image in the OCT image group;

IPA value generating module, for calculating the IPA value of each of the light attenuation coefficient images in the light attenuation coefficient image group;

An image marking module, based on the registration parameters, uses the IPA value to mark each angiographic image in the angiographic image group to obtain a target angiographic image group.

In a third aspect, an embodiment of the present application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and operable on the processor, the computer program being executed by the processor When executed, the method steps described in the first aspect above are realized.

In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, including: the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in the above-mentioned first aspect is implemented step.

In a fifth aspect, an embodiment of the present application provides a computer program product, which, when the computer program product is run on an electronic device, causes the electronic device to execute the method steps described in the first aspect above.

It can be understood that, for the beneficial effects of the above-mentioned second aspect to the fifth aspect, reference can be made to the relevant description in the above-mentioned first aspect, and details will not be repeated here.

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 image processing method provided by an embodiment of the present application;

Fig. 2 is a schematic diagram of the corresponding relationship between OCT images and angiography images provided by an embodiment of the present application;

Fig. 3 is a schematic diagram of image processing provided by an embodiment of the present application;

Fig. 4 is a schematic flowchart of an image processing method provided by another embodiment of the present application;

FIG. 5 is a schematic flowchart of an image processing method provided by another embodiment of the present application;

Fig. 6 is a schematic diagram of image registration provided by an embodiment of the present application;

Fig. 7 is a schematic diagram of a U-shaped network of attention provided by an embodiment of the present application to detect and pull back blood vessels;

Fig. 8 is a schematic diagram of a loop generation confrontation network provided by an embodiment of the present application;

Fig. 9a is an OCT image sample provided by an embodiment of the present application;

Fig. 9b is an image sample of light attenuation coefficient provided by an embodiment of the present application;

Fig. 10 is a schematic diagram of an image processing device provided by an embodiment of the present application;

FIG. 11 is a schematic structural diagram of an electronic device provided by an embodiment of the present application.

Detailed ways

In the following description, specific details such as specific system structures and technologies are presented for the purpose of illustration rather than limitation, so as to thoroughly understand the embodiments of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

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.

Optical coherence tomography (OCT) technology is an imaging technology. It uses the basic principle of weak coherent light interferometer to divide the light emitted by the light source into two beams, one beam is sent to the measured tissue, also called the sample arm, and the other beam is sent to the reference mirror, also called the reference arm, and then the The measured tissue and the two beams of light signals reflected from the reference mirror are superimposed and interfered, and finally according to the light signal with the measured tissue, different intensity image gray levels are displayed, so as to image the tissue.

The existing traditional optical coherence imaging technology is weak in identifying vulnerable plaques. Judging vulnerable plaque requires judging the thickness of the fibrous cap, which requires manual measurements by doctors. The subjective factors of the measurer may lead to the variability of the measurement results. It is difficult to improve the recognition ability from the system and equipment side, and it will also incur high costs.

Due to the existence of dark areas such as lipids in vulnerable plaques, the borders of the fibrous caps of vulnerable plaques in OCT images are not very clear. load situation.

The embodiment of the present application provides an image processing method, based on the result of AngioCoalesceRegistration (ACR), calculates the index of plaque attenuation (IPA) value based on the image based on the light attenuation coefficient, and calculates the IPA Values are mapped into the contrast image. Since the light attenuation of the vulnerable plaque area is obvious, the angiographic image obtained by the image processing method provided in the embodiment of the present application can display the IPA value corresponding to the light attenuation image to intuitively prompt the doctor that there may be vulnerable plaques in the current image frame. The damaged plaque makes the display of IPA value more intuitive and effective.

It should be pointed out that the image processing method provided in the embodiment of the present application may be implemented by software and/hardware including but not limited to equipment for acquiring angiographic images, OCT equipment, local third-party computing equipment, and remote third-party computing equipment. The application does not limit the subjects implementing the image processing method. The third party is a device other than an OCT device and an angiographic image device.

FIG. 1 shows an image processing method provided by an embodiment of the present application. As shown in Fig. 1, the method includes steps S110 to S150. The specific implementation principle of each step is as follows:

S110. During the process of acquiring the optical coherence tomography OCT image group for the investigation object, acquire the angiography image group for the investigation object.

During the process of performing optical coherence tomography on a blood vessel of an object to be investigated, such as a coronary artery, by using an OCT device, a series of OCT images will be obtained, which are recorded as an OCT image group. During this process, angiographic imaging is performed on the probe object by an angiographic imaging device, and a series of angiographic images are acquired, which are recorded as an angiographic image group. In some embodiments, the angiographic imaging is for coronary vessels, and the obtained image is called a coronary angiography image (coronary arteriography, CAG).

It can be understood that the imaging device for acquiring the OCT image group and the angiography image group may be the same device, or may be two different devices, or may be a combination of devices with a control relationship. In some embodiments, the OCT device may perform image processing on the OCT image group and the angiography image group, and the angiography imaging device may also perform image processing on the OCT image group and the angiography image group. In some embodiments, after the OCT image group and the angiography image group can be obtained by the OCT device and the angiography imaging device, the third-party computing device obtains the OCT image group and the angiography image group through a storage medium, a communication cable, and a communication network to perform deal with.

S120. Register the OCT image group and each angiography image in the angiography image group to obtain a registration parameter.

It should be understood that, as shown in FIG. 2 , the OCT image group 21 is a series of tomographic images of blood vessels of the exploration object, that is, equivalent to cross-sectional images of blood vessels. The angiographic image 22 is a projected image of a blood vessel to be investigated. The OCT image group is registered with the angiographic image, that is, the vessel section corresponding to each OCT image 211 in the OCT image group is determined, and the position 221 of the vessel projection in the angiographic image is determined. In other words, a corresponding relationship between each OCT image in the OCT image group and the position in the angiography image is established. Register the OCT image group and each angiographic image in the angiographic image group, that is, for each angiographic image in the angiographic image group, establish the OCT image group and the angiographic image location correspondence.

S130. Based on each OCT image in the OCT image group, generate a light attenuation coefficient image group corresponding to each OCT image.

The Optical Attenuation Coefficient (OAC) of blood vessel wall (including plaque) tissue is an optical characteristic parameter of OCT images. The light attenuation coefficient of biological tissue varies with the spatial position. Therefore, tissue components such as thin fibrous caps, calcifications, and lipid-rich plaques can be quantitatively calibrated according to the light attenuation coefficient.

In some embodiments, based on the OCT image, the optical parameters of the OCT device, such as the Rayleigh length (theRayleigh length, zR), the half width of the roll-off function (the half width of the roll-off function, zW), etc., can be used to calculate Obtain the light attenuation coefficient corresponding to the tissue in each OCT image, and obtain the light attenuation coefficient image. Calculation methods include but are not limited to curve fitting (curve fitting, CF), or depth-resolved (depth-resolved, DR) model methods. Thereby, a series of light attenuation coefficient images corresponding to the OCT image group are obtained, which are recorded as the light attenuation coefficient image group. It should be pointed out that there is a one-to-one correspondence relationship between the OCT image group and the optical attenuation coefficient image group. That is to say, the registration parameters of the OCT image group and the angiography image group are consistent with the registration parameters of the light attenuation coefficient image group and the angiography image group.

S140. Calculate a plaque attenuation index IPA value of each of the light attenuation coefficient images in the light attenuation coefficient image group.

In some embodiments, the index of plaque attenuation (IPA) is the fraction of pixels in the attenuation map whose attenuation coefficient is greater than a certain threshold x. In a specific example, the score can also be multiplied by a factor of 1000. Specifically, the following formula can be used to calculate the IPA value.

Among them, x is the plaque attenuation coefficient threshold, μ _t is the attenuation coefficient, N(μ _t >x) is the total number of A-lines whose maximum attenuation value is greater than x on each A-line (A-line), and N _total represents all A-lines. number of lines.

S150. Based on the registration parameter, use the IPA value to mark each angiographic image in the angiographic image group to obtain a target angiographic image group.

In some embodiments, the IPA value is a gray value, and the IPA value is used to mark each angiographic image in the angiographic image group to obtain the target angiographic image group, including: marking the gray value in The position corresponding to the light attenuation image on each angiographic image.

In some embodiments, the corresponding relationship between the IPA value and the marking parameter is preset, such as a comparison table, or a conversion curve formula. This correspondence is used to map the IPA values to tagged parameters. The marking parameter can be an RGB color value, or a computer code corresponding to symbols such as "+", "*", "#", or a computer code corresponding to other marking symbols. The marker parameter corresponding to the IPA value is displayed at the position corresponding to the light attenuation image on each angiography image.

In some embodiments, the OCT image group is a cross-sectional image of the retracted vessel on the angiogram, that is, the OCT image in the OCT image group corresponds to the position of the retracted vessel on the angiogram. It should be understood that, due to the one-to-one correspondence between the light attenuation coefficient image and the OCT image, the IPA value obtained from each light attenuation coefficient image also corresponds to a position of the retracted blood vessel on the angiogram. Based on the registration parameters, using the IPA value to mark each angiographic image in the angiographic image group to obtain a target angiographic image group, including: according to the preset correspondence between the IPA value and the marking parameter, converting each of the IPA values into marker parameters; according to the registration parameters, at the target positions of the respective angiography images, displaying the marker parameters of the corresponding light attenuation coefficient images, and the target positions are corresponding to the light attenuation coefficient images The pixel position on the pullback path of .

In a specific example, the marker parameter can be a color value. For example, if the IPA value is 100, query the correspondence between the preset IPA value and the marking parameter (it may be a comparison table), and determine that the RGB value corresponding to the IPA value is [179, 62, 110]. According to the registration parameters, pixels corresponding to the position of the blood vessel are drawn back on the angiography image to display the RGB value. That is to say, the color rendering and display of the pull-back path on the current contrast image is realized.

It can be understood that, as shown in FIG. 3 , in the embodiment of the present application, the registration parameters of the OCT image group and each angiographic image are obtained by registering each angiographic image in the OCT image group and the angiographic image group, that is, The corresponding relationship makes the optical attenuation coefficient image obtained by using each OCT image have a consistent corresponding relationship with each angiography image. On this basis, the IPA value obtained based on the light attenuation coefficient image is marked on each angiographic image, so that the position and compliance of the vulnerable plaque can be clearly and intuitively observed, and the ability to identify the vulnerable plaque of the detection object is improved. .

On the basis of the image processing method shown in FIG. 1 above, as shown in FIG. 4, step S120 is to register each angiographic image in the OCT image group and the angiographic image group to obtain registration parameters , including step S121 to step S123. This process is also called contrast fusion registration (AngioCoalesceRegistration, ACR).

S121. Detect a pull-back path in each angiography image.

In the angiographic image, the pull-back vessel is the vessel with the guide wire, that is, the vessel to be scanned. In some occasions, pullbacks are also called retracements. The pull-back path, also known as the retraction path, is the path scanned by the optical catheter during the implementation of the OCT scan.

In some embodiments, as shown in FIG. 5 , detecting the retraction path in each angiographic image includes steps S1211 to S1214:

S1211. Using the pre-trained target detection model, detect the retracted blood vessel in each angiography image.

Wherein, the pull-back blood vessel is a blood vessel with a guide wire, that is, the blood vessel to be scanned.

In some embodiments, the target detection model may be a deep learning network model for target detection. In some embodiments, the object detection model performs an image segmentation operation on the detected pullback vessel in the angiographic image.

S1212. Detect a starting position and an ending position of a developing object in the pulling blood vessel in each angiography image.

Wherein, the developing object may be a developing ring or an optical probe. The developing ring is a metal ring set at the tip of the guide wire to increase the developing effect.

S1213. Project the starting position and the ending position to the respective angiographic images by using an iterative closest point (Iterative Closest Point, ICP) algorithm.

In some embodiments, during the process of acquiring the optical coherence tomography OCT image group for the investigation object, the angiography image group for the investigation object is acquired, and the starting position is the movement of the developing ring of the first angiography image The end position is the position where the developing ring of the last blood vessel image moves. When the positions of the developing rings of the two head and tail images are determined, the pulling path of the developing ring movement can be obtained, and then other angiographic images can be projected through the ICP algorithm. In this way, each angiography image is marked with the start position and the end position.

S1214. Using a shortest path algorithm, based on the starting position and the ending position in each angiography image, obtain a pull-back path in each angiography image.

In some embodiments, the weights matrix is established with the gray value of the pixel points of the current angiography image. Since the gray value of the blood vessel part is low, the weight is low, and the shortest path algorithm is to ensure that the weight of the path between the target points is the lowest . After the start and end positions of the developing ring are determined, the pullback path can be determined on the pullback blood vessel.

Fig. 6 is a schematic diagram of image registration provided by an embodiment of the present application. FIG. 6 shows an example of projecting the position of the development ring to other angiographic images by using the ICP algorithm by taking the position of the development ring in the last frame of angiography image 61 , that is, the end position 611 of the development ring, as an example. And after the projection operation, any frame of angiography image 62 having the starting position 612 and the end position of the developing ring 611 is an example of obtaining the pull-back path by using the shortest path method.

The registration parameters include a corresponding relationship between each OCT image in the OCT image group and a target position, and the target position is a pixel point position in each angiography image. It should be understood that the target position is a pixel point position on the pullback path in each angiography image.

By registering each angiographic image in the OCT image group and angiographic image group, the corresponding relationship between each OCT image in the OCT image group and the target position is obtained, so that the light attenuation coefficient image obtained by using each OCT image has the same relationship with the target position. consistent correspondence. On this basis, the IPA value obtained based on the light attenuation coefficient image is marked on the target position of each angiographic image, so that the position and compliance of the vulnerable plaque can be clearly and intuitively observed, and the detection of the vulnerable plaque of the detection object is improved. recognition ability.

S122. For each angiographic image in the angiographic image group, according to the frame rate of the OCT image group, perform equal interval sampling on the pull-back path in the angiographic image, to obtain each OCT image and the return path The corresponding relationship of the pixel position on the pull path.

S123. Use the corresponding relationship between each OCT image and the position of the pixel point on the pull-back path as the registration parameter.

It should be pointed out that the default OCT pullback speed remains unchanged, therefore, the distance moved per unit time is also constant. Therefore, the pull-back path in the angiography image can be sampled at equal intervals to establish a corresponding relationship between the position on the pull-back path and each OCT image.

In some embodiments, it is assumed that the pull-back path on the angiography image is 600 pixels away, and a set of OCT pull-back is 300 frames of images. The OCT device obtains one frame of OCT image per scan, and correspondingly, the position of the developing ring on the angiography image moves by 2 pixels. Therefore, one OCT image corresponds to the positions of two pixel points on each angiography image.

FIG. 6 takes a frame of OCT image 64 as an example. The registration relationship between the OCT image and the angiographic image 63 with the retraction path determined is that the frame number of the OCT image 64 corresponds to a position of the return vessel on the angiographic image 63 631. It should be noted that position 631 may contain one or more pixels, which depends on the OCT frame rate and the number of pixels contained in the pullback path.

On the basis of the image processing method shown in FIG. 5 , step S1211, using the pre-trained target detection model to detect the retracted blood vessels in each angiography image, including:

As shown in FIG. 7 , the trained target detection model is an Attention-U-net model as the target detection model.

The Attention-U-net model is trained using an angiographic image sample set of pre-marked pull-back vessels. In some embodiments, for the application of coronary angiography, a group of CAG angiography images can be prepared, and the pull-back vessels with guide wires in them can be marked by experts for training the Attention-U-net model to identify the Pull back the blood vessel.

It should be pointed out that if the Attention-U-net network is used to fine-segment the pull-back vessels in the CAG images, mainly including the left anterior descending artery (LAD), the left circumflex artery (LCX), and the right coronary artery (RCA). , when collecting training samples, try to ensure that the proportions of the CAG images of each body position are consistent, and ensure that the segmentation performance of the network in each model is close.

In some embodiments, since the amount of data in the sample set of CAG images marked by experts to pull back blood vessels is small, the amount of data samples can be expanded by rotating, flipping, adjusting contrast, and the like.

表示。Attention-U-net adds the mechanism of attention attention on the basis of U-net, which implements the attention mechanism by supervising the features of the upper level through the feature features of the next level. As shown in Figure 7, the Attention-U-net includes the process of downsampling and upsampling, and adds an attention mechanism, which is used in Figure 7

express.

The Attention-U-net model is trained using a loss function _LAUN that includes weight parameters for expanding and pulling blood vessels.

Wherein, r _ln represents the real pixel category of the category l at the nth position, and the category in the embodiment of the present application is divided into two categories: pull-back blood vessel path pixels and background pixels. And P _ln represents the corresponding prediction probability value, ω _l represents the weight of each category, when the proportion of the category in the image is larger, the weight is smaller.

It should be understood that, for the convenience of understanding the present application, the embodiment of the present application provides an example of the above formula for the loss function including the weight parameter of the enlarged and retracted blood vessel. Those skilled in the art can refer to this example, and adjust the form of the weight parameter of the enlarged and retracted blood vessel and the specific parameters of the loss function according to the actual situation.

It is understandable that the activation value is adjusted by automatically learning parameters, the activated part is limited to the area with segmentation, the activation value of the background is reduced to optimize the segmentation, and the end-to-end segmentation is realized. Attention-U-net is in complex images. The segmentation accuracy is higher, so that the trained model can automatically segment the blood vessel through which the guide wire passes, and quickly locate and pull the blood vessel.

On the basis of the image processing method shown in FIG. 1 above, as shown in FIG. 8, the embodiment of the present application also provides a cycle-generated adversarial network (CycleConsistent Generative Adversarial Networks, CycleGAN), based on the OCT image group A method for generating an optical attenuation coefficient image group corresponding to each OCT image in each OCT image. Among them, the cycle generation confrontation network is also called cycle consistency generation confrontation network.

Refer to the method for generating an optical attenuation coefficient image group corresponding to each OCT image based on each OCT image in the OCT image group provided in the foregoing embodiment. Based on the optical parameters of the OCT device, the optical hardware parameters of the OCT device are required to obtain the optical attenuation coefficient image of the OCT image through calculation. However, when the OCT equipment leaves the factory, the relevant parameters may be slightly different, which will cause systematic errors in the optical attenuation coefficient image generated by the OCT image. In order to eliminate the influence of the OCT system on the IPA calculation, the embodiment of the present application provides a method of using the CycleGAN network to synthesize the light attenuation coefficient image from the OCT image, and calculate the IPA value based on the synthesized light attenuation coefficient image.

In some embodiments of the present application, OCT images generated by multiple OCT devices are prepared for the CycleGAN network to form an OCT image sample set. Based on the OCT image sample set, the optical attenuation coefficient image sample set is obtained by calculating according to the optical parameters of each OCT device. Fig. 9a shows an OCT image sample provided by an embodiment of the present application, and Fig. 9b shows an optical attenuation coefficient image sample provided by an embodiment of the present application.

Using the OCT image sample set and light attenuation coefficient image sample set generated by multiple devices to train the CycleGAN network can improve the generalization ability of the CycleGAN network, so that the trained CycleGAN network can generate corresponding light based on the OCT image generated by any OCT device. Attenuation coefficient image.

In some embodiments, since the amount of data in the OCT image sample set and the light attenuation coefficient image sample set obtained through expert annotation is small, the OCT image sample set and the light attenuation coefficient image sample set can be rotated, flipped, and contrast adjusted. A transformation is performed to augment the data sample size.

Referring to Figure 8, the OCT image is synthesized into a light attenuation coefficient image using the CycleGAN network. The CycleGAN network mainly contains two cycles, the forward cycle and the reverse cycle.

Among them, the forward cycle mainly includes three independent CNN models. The synthesizer network is also called the generator network:

(1) Syn _IPA is a synthesizer network that converts OCT image Img _OCT into IPA image;

(2) Syn _OCT is a synthesizer network that converts the light attenuation coefficient image Syn _IPA (Img _OCT ) back to the OCT image;

(3) Dis _IPA is to distinguish the synthesized optical attenuation coefficient image Syn _IPA (Img _OCT ) from the real optical attenuation coefficient image RealIPAImg discriminator network.

The label is 0 for the synthetic light attenuation coefficient image and 1 for the real light attenuation image. The continuous learning of the discriminator network can distinguish the synthetic from the real, that is, for the synthetic, the output of the discriminator is 0, and for the real, the output is 1. However, due to the continuous training of the model, the quality of the generated light attenuation map is getting better and better, and it is closer to the real one. Finally, it is difficult for the discriminator to distinguish between synthetic and normal, and the purpose of model training is obtained.

When the network Dis _IPA tries to distinguish the synthesized optical attenuation coefficient image Syn _IPA (Img _OCT ) from the real optical attenuation coefficient image RealIPAImg, the network Syn _IPA will synthesize the OCT image as close as possible to the real optical attenuation coefficient image Syn _IPA (Img _OCT ), making the network Dis _IPA indistinguishable. In addition, the synthesized optical attenuation coefficient image Syn _IPA (Img _OCT ) also needs to be converted back to the OCT image through the network Syn _OCT , so that the original image Syn _OCT (Img _IPA ) can be reconstructed as accurately as possible.

In order to improve the stability of the training, a reverse loop is also added, using the optical attenuation coefficient image to synthesize the OCT image, and converting the synthesized OCT image back to the optical attenuation coefficient image. The reverse cycle also contains three parts, in which the two synthetic networks of the reverse cycle are common to the forward cycle, namely, the network Syn _OCT and the network Syn _IPA . In addition, the reverse loop also contains the discriminator network Dis _OCT , which is used to distinguish the synthetic OCT image Syn _OCT (Img _IPA ) from the real OCT image RealOCTImg.

The adversarial objectives of the synthesizer network and the discriminator network are reflected in the loss functions Loss _IPA and Loss _OCT as follows.

The discriminator Dis _IPA is used to judge whether the image is a real light attenuation coefficient image, when the image is a real light attenuation coefficient image, the value is 1, and when the image is a synthetic light attenuation coefficient image, the value is 0. The discriminator Dis light attenuation coefficient will minimize the loss item as much as possible, and its loss Loss _IPA is:

Loss _IPA = (1-Dis _IPA (Img _IPA )) ² +Dis _IPA (Syn _IPA (Img _OCT )) ² .

Similarly, the discriminator Dis _OCT is used to judge whether the image is a real OCT image, the real one is 1, otherwise it is 0, and its loss Loss _OCT is:

Loss _OCT = (1-Dis _OCT (Img _OCT )) ² +Dis _OCT (Syn _OCT (Img _IPA )) ² .

In addition, when calculating the cycle consistency loss Loss _cycle , considering that both the OCT image and the light attenuation coefficient image focus on the area with a large gray value, that is, the unstable plaque, but the proportion of the unstable plaque in the image is It is very small, and the weight coefficient of the area with a larger gray value is increased in the loss item.

Wherein, i represents the pixel value of the corresponding position of the real image and the synthesized image, N represents the number of pixels in the image, and in some embodiments, 704×704 is used here. Syn _OCT (Syn _IPA (Img _OCT )) means converting the synthesized optical attenuation coefficient image back to an OCT image. Similarly, Syn _IPA (Syn _OCT (Img _IPA )) represents the optical attenuation coefficient image converted from the synthesized OCT image. When the pixel value of the corresponding position in the real image is larger, it represents greater instability, and the error of the position is smaller.

和第二生成损失项

所述第一生成损失项包含与第一生成样本的像素值负相关的第一权重系数

所述第一生成样本为OCT图像样本；所述第二生成损失项包含与第二生成样本的像素值负相关的第二权重系数

所述第二生成样本为光衰减系数图像样本。因为OCT和IPA中有效的像素区域占图片的比重较小，因为如果不加入权重项，损失对OCT和IPA有效区域的权重也越小，但是通过这两项中的权重系数，有效像素占图像总像素的比例越小，这两项就越大，从而增加权重。The cycle consistency loss of the cycleGAN network provided by the embodiment of this application includes the first generation loss item

and a second generation loss term

The first generation loss term contains a first weight coefficient negatively correlated with the pixel value of the first generation sample

The first generated sample is an OCT image sample; the second generated loss item includes a second weight coefficient negatively correlated with the pixel value of the second generated sample

The second generated samples are light attenuation coefficient image samples. Because the effective pixel area in OCT and IPA accounts for a small proportion of the picture, because if the weight item is not added, the weight of the loss to the effective area of OCT and IPA is also smaller, but through the weight coefficients in these two items, the effective pixel accounts for the image The smaller the proportion of total pixels, the larger these two terms are, increasing their weight.

The total loss function Loss _total of the cycleGAN network is Loss _total = Loss _IPA + Loss _OCT + λLoss _cycle . Among them, λ is the scaling factor, which is a hyperparameter.

Refer to FIG. 8 for a specific network structure diagram, which includes a forward loop and a reverse loop. In the forward cycle, Syn _IPA synthesizes the OCT image into an optical attenuation coefficient image, Syn _OCT converts the synthesized optical attenuation coefficient image back to an OCT image close to the original image, and Dis _IPA is used to distinguish the real optical attenuation coefficient image RealIPAImg and the composite light attenuation coefficient image. In the reverse cycle, Syn _OCT is the optical attenuation coefficient image synthesis OCT image, Syn _IPA converts the synthesized OCT image back to the optical attenuation coefficient image close to the original image, and Dis _OCT is used to distinguish the real OCT image RealOCTImg from the synthetic OCT image.

It should be pointed out that when training the CycleGAN network, the training OCT image is synthesized into the light attenuation coefficient image, and the light attenuation coefficient image is synthesized into the OCT image, that is to say, both the forward cycle and the reverse cycle are trained. However, after CycleGAN is trained, only the OCT image is used to synthesize the network of the image part of the light attenuation coefficient.

It can be understood that the method of using the CycleGAN network provided in the embodiment of the present application to synthesize the light attenuation coefficient image according to the OCT image, and calculate the IPA value according to the synthesized light attenuation coefficient image can get rid of the influence of specific OCT equipment parameters, and directly pass the OCT The image synthesis of the light attenuation coefficient image reduces the system error and greatly improves the processing efficiency of the synthesis of the light attenuation coefficient image.

Corresponding to the image processing method shown in FIG. 1 above, FIG. 10 shows an image processing device M100 provided in an embodiment of the present application, including:

The image acquisition module M110 is configured to acquire an angiography image group for the investigation object during the process of obtaining the optical coherence tomography OCT image group for the investigation object.

The image registration module M120 is configured to perform registration on the OCT image group and each angiography image in the angiography image group to obtain registration parameters.

The light attenuation coefficient image generation module M130 is configured to generate a plaque attenuation coefficient light attenuation coefficient image group corresponding to each OCT image based on each OCT image in the OCT image group.

The IPA value generating module M140 is configured to calculate the IPA value of each of the light attenuation coefficient images in the light attenuation coefficient image group.

An image marking module, based on the registration parameters, uses the IPA value to mark each angiographic image in the angiographic image group to obtain a target angiographic image group.

It can be understood that the various implementation manners and combinations of implementation manners and their beneficial effects in the above embodiments are also applicable to this embodiment, and will not be repeated here.

FIG. 11 is a schematic structural diagram of an electronic device provided by an embodiment of the present application. As shown in FIG. 11 , the electronic device D10 of this embodiment includes: at least one processor D100 (only one is shown in FIG. 11 ), a processor, a memory D101, and a processor that is stored in the memory D101 and can be processed in the at least one processor. A computer program D102 running on the processor D100, when the processor D100 executes the computer program D102, implements the steps in any of the above method embodiments.

The electronic device D10 may be computing devices such as OCT equipment, angiographic imaging equipment, desktop computers, notebooks, palmtop computers, and cloud servers. The electronic device may include, but not limited to, a processor D100 and a memory D101. Those skilled in the art can understand that FIG. 11 is only an example of the electronic device D10, and does not constitute a limitation to the electronic device D10. It may include more or less components than those shown in the figure, or combine certain components, or different components. , for example, may also include input and output devices, network access devices, and so on.

The so-called processor D100 can be a central processing unit (Central Processing Unit, CPU), and the processor D100 can also be other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit , ASIC), off-the-shelf 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.

In some embodiments, the storage D101 may be an internal storage unit of the electronic device D10, such as a hard disk or a memory of the electronic device D10. The memory D101 may also be an external storage device of the electronic device D10 in other embodiments, such as a plug-in hard disk equipped on the electronic device D10, a smart memory card (Smart MediaCard, SMC), a secure digital ( Secure Digital (SD) card, flash memory card (Flash Card), etc. Further, the memory D101 may also include both an internal storage unit of the electronic device D10 and an external storage device. The memory D101 is used to store operating systems, application programs, boot loaders (BootLoader), data and other programs, such as program codes of the computer programs. The memory D101 can also be used to temporarily store data that has been output or will be output.

It should be understood that the sequence numbers of the steps in the above embodiments do not mean the order of execution, and the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiment of the present application.

It should be noted that the information interaction and execution process between the above-mentioned devices/units are based on the same concept as the method embodiment of the present application, and its specific functions and technical effects can be found in the method embodiment section. I won't repeat them 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 the functional units and modules 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.

The embodiment of the present application also provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the foregoing method embodiments can be implemented.

An embodiment of the present application provides a computer program product, which, when the computer program product runs on an electronic device, enables the electronic device to implement the steps in the foregoing method embodiments when executed.

If the integrated unit is realized in the form of a software function unit and sold or used as an independent product, it 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 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 signals, telecommunication signals, and software distribution media. 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/network device and method may be implemented in other ways. For example, the device/network 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 may be 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.

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

The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still implement the foregoing embodiments Modifications to the technical solutions described in the examples, or equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the application, and should be included in the Within the protection scope of this application.

Claims (10)

1. A method of image processing, comprising:

detecting a pull-back path in each of angiography images in an angiography image group of an inspection object; the angiographic image group is acquired in a process of performing Optical Coherence Tomography (OCT) on the probe object;

for each angiography image, sampling a pull-back path in the angiography image at equal intervals according to a frame frequency of an OCT image group corresponding to the angiography image, and obtaining a corresponding relation between each OCT image in the OCT image group and a pixel point position on the pull-back path;

and determining the corresponding relation between each OCT image and the position of the pixel point on the pull-back path as the registration parameter of the angiography image.

2. The method of claim 1, wherein said detecting a pull-back path in each angiographic image of the set of angiographic images of the object under investigation comprises:

detecting a pull-back blood vessel in each angiography image by using a target detection model obtained through pre-training;

detecting a start position and an end position of a development target in the pull-back blood vessel in the respective angiographic images;

projecting the start and end positions to the respective angiographic images using a closest point iteration algorithm;

and obtaining a pull-back path in each angiography image based on the starting position and the ending position in each angiography image by using a shortest path algorithm.

3. The method of claim 2, wherein the target detection model is an Attention-U-type network Attention-U-net model;

the Attention-U-net model is trained by adopting an angiography image sample set which is labeled with a pullback blood vessel in advance;

the Attention-U-net model is trained using a loss function that includes parameters that expand the weight of the pull-back vessel.

4. The method of claim 3, wherein the loss function comprises:

wherein L is _AUN Representing said loss function, r _ln A real pixel class representing class i at the nth position, the class including a pull-back blood vessel path pixel and a background pixel, P _ln Representing the corresponding prediction probability value, ω _l And the weight of each category is represented, and the larger the proportion of the categories on the image is, the smaller the weight is.

5. An apparatus for image processing, comprising an image registration module; the image registration module is to:

detecting a pull-back path in each of angiography images in an angiography image group of an inspection object; the angiographic image group is acquired in a process of performing Optical Coherence Tomography (OCT) on the probe object;

for each angiography image, sampling a pull-back path in the angiography image at equal intervals according to a frame frequency of an OCT image group corresponding to the angiography image, and obtaining a corresponding relation between each OCT image in the OCT image group and a pixel point position on the pull-back path;

and determining the corresponding relation between each OCT image and the position of the pixel point on the pull-back path as the registration parameter of the angiography image.

6. The apparatus of claim 5, wherein the image registration module is further to:

detecting a pull-back blood vessel in each angiography image by using a target detection model obtained through pre-training;

detecting a start position and an end position of a development target in the pull-back blood vessel in the respective angiographic images;

projecting the start and end positions to the respective angiographic images using a closest point iteration algorithm;

and obtaining a pull-back path in each angiography image based on the starting position and the ending position in each angiography image by using a shortest path algorithm.

7. The apparatus of claim 6, wherein the target detection model is an Attention-U-type network Attention-U-net model;

the Attention-U-net model is trained by adopting an angiography image sample set which is labeled with a pullback blood vessel in advance;

the Attention-U-net model is trained using a loss function including an expanded pull-back vessel weight parameter.

8. The apparatus of claim 7, wherein the loss function comprises:

wherein L is _AUN Representing said loss function, r _ln Representing a real pixel class of class i at the nth position, the class comprising a pullbackBlood vessel path pixels and background pixels, P _ln Representing the corresponding prediction probability value, ω _l And the weight of each category is represented, and the larger the proportion of the categories on the image is, the smaller the weight is.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method of any of claims 1 to 5 when executing the computer program.

10. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1 to 5.

Images