CN115908330A

CN115908330A - DSA image-based coronary artery automatic frame selection classification recommendation method and device

Info

Publication number: CN115908330A
Application number: CN202211476593.XA
Authority: CN
Inventors: 向建平; 陆徐洲; 鲁伟; 何京松
Original assignee: Arteryflow Technology Co ltd
Current assignee: Arteryflow Technology Co ltd
Priority date: 2022-11-23
Filing date: 2022-11-23
Publication date: 2023-04-04

Abstract

The invention discloses a DSA image-based coronary artery automatic frame selection classification recommendation method and device, wherein the method comprises the following steps: obtaining a DSA image group to be detected, and identifying each D I COM image in the DSA image group to be detected by adopting a pre-constructed automatic frame selection model so as to obtain a key frame image; classifying the blood vessel types of the key frame images based on a pre-constructed automatic classification model to obtain probability response values of the key frame images corresponding to the blood vessel types; arranging the key frame images in a descending order according to the probability response values and forming an image pair by every two key frame images from back to front; and scoring the image frame image pairs based on the mean value of the probability response values corresponding to the image pairs and the imaging angle difference, and recommending the image pairs with the highest scores. The method can efficiently and accurately identify the image pair suitable for reconstruction in the DSA images of the coronary artery.

Description

DSA image-based coronary artery automatic frame selection classification recommendation method and device

Technical Field

The invention relates to the technical field of artificial intelligence, in particular to a coronary artery automatic frame selection classification recommendation method and device based on DSA images.

Background

Coronary heart disease, also called ischemic heart disease, is a heart disease caused by myocardial ischemia and hypoxia due to coronary atherosclerosis. The coronary artery is the only blood vessel supplying blood to the heart, and is called coronary artery because it is shaped like a coronary artery. The blood vessel is also hardened and changed along with the whole body blood vessel, which causes blood circulation disorder of the heart, myocardial ischemia and anoxia, and is the coronary heart disease. Coronary heart disease is a common disease of middle-aged and old people, frequently occurs and seriously endangers the life of people.

Conventional approaches to diagnosing coronary heart disease include simple noninvasive electrocardiograms, coronary CTA, which can only acquire static images, invasive approaches including coronary intravascular ultrasound (IVUS) and dynamic coronary angiography. In several of these modalities, coronary angiography is considered the "gold standard" for coronary heart disease diagnosis. As the primary imaging technique for diagnosing coronary artery disease, the morphology of the coronary artery is obtained by real-time visualization of the ductal chamber during coronary angiography, while Quantitative Coronary Angiography (QCA) can also be used to provide objective quantitative measurements.

Because the reconstruction based on two-angle DSA image frame images requires a doctor to actively select two optimal images from a series of images, and a patient may have various image data, selecting reasonable data from a large amount of data requires a certain effort and judging whether the two images are suitable for reconstruction according to experience, which greatly increases the burden of the doctor.

Therefore, there is a need in the art for a new frame selection classification recommendation method to solve the above problems.

Disclosure of Invention

The invention aims to provide a DSA image-based coronary artery automatic frame selection, classification and recommendation method and device, and aims to solve the problem of how to efficiently and accurately identify an image pair suitable for reconstruction in a DSA image of a coronary artery.

In order to achieve the purpose, the invention adopts the following technical scheme:

a DSA image-based automatic frame selection and classification recommendation method for coronary arteries comprises the following steps:

obtaining a DSA image group to be detected, and identifying each D I COM image in the DSA image group to be detected by adopting a pre-constructed automatic frame selection model so as to obtain a key frame image;

classifying the blood vessel types of the key frame images based on a pre-constructed automatic classification model so as to obtain probability response values of the key frame images corresponding to the blood vessel types;

arranging the probability response values in a descending order, and forming an image pair by every two key frame images from back to front;

and scoring the image frame image pairs based on the mean value of the probability response values corresponding to the image pairs and the imaging angle difference, and recommending the image pairs with the highest scores.

In an embodiment, the automatic frame selection model is a two-classification neural network model, the automatic classification model is a four-classification neural network model, the blood vessel categories include LAD, LCX, RCA and OTHERS, and the OTHERS are blood vessel categories other than LAD, LCX and RCA.

In one embodiment, the network training step of the "automatic frame selection model" includes:

acquiring a D I COM image generated based on coronary angiography;

setting D I COM image frames in the D I COM image, in which the contrast is fully full, the contrast agent is clearly visible and the D I COM image frames are in the end diastole, as key frame images, and forming a data recording table;

traversing the data record table to select a key frame image and setting the label of the key frame image to be 1, and selecting a non-key frame image and setting the label of the non-key frame image to be 0;

and forming an automatic frame selection data set based on the key frame images and the non-key frame images and training the automatic frame selection model by adopting a machine learning algorithm.

In one embodiment, the non-key frame image includes one of the first 10 frames of the D I COM image, one of the first 5 frames of the key frame image, and one of the second 5 frames of the key frame image, which are randomly selected.

In one embodiment, the network training step of the "automatic classification model" includes:

acquiring a D I COM image generated based on coronary angiography;

reading D I COM image frames in the D I COM image, wherein the D I COM image frames are full of contrast, clear and visible in contrast agent and in the end diastole and are set as key frame images;

classifying the keyframe images into LAD, LCX, RCA and OTHERS by vessel category and assigning image labels of [1,0,0,0], [0,1,0,0], [0,0,1,0], [0,0,0,1] respectively to generate a vessel classification dataset;

training the automatic classification model based on the vessel classification dataset and using a machine learning algorithm.

In one embodiment, the method further comprises the step of adjusting the contrast of the D I COM image according to the window width level and adjusting the D I COM image to a fixed size.

In one embodiment, the step of scoring the image frame image pair based on the mean of the probability response values corresponding to the image pair and the imaging angle difference of the corresponding image pair includes:

the score is calculated according to the recommendation method shown below:

z＝a×x+(1-a)×y

wherein z is a score value, a is a preset proportionality coefficient, x is a mean value of probability response values corresponding to the image pair, and y is an angle difference corresponding to the image pair.

In one embodiment, the step of determining the a comprises:

acquiring a plurality of image pairs for reconstruction;

and sequencing according to the reconstruction quality of the image pair, and adjusting the scale coefficient by taking the sequencing as a reference so that the grading sequencing result calculated by the recommendation algorithm is consistent with the actual sequencing.

A storage device having stored therein a plurality of programs adapted to be loaded and executed by a processor to implement the DSA image-based coronary artery automated frame selection classification recommendation method described above.

A control device, comprising: a processor adapted to execute various programs; a storage device adapted to store a plurality of programs; the program is suitable for being loaded and executed by a processor to realize the DSA image-based coronary artery automatic frame selection classification recommendation method.

In summary, the DSA image-based automatic frame selection, classification and recommendation method and device for coronary artery provided by the invention have the following beneficial effects:

the invention can automatically, quickly and accurately identify the image pair suitable for reconstruction in the DSA images of the coronary artery, reduces the workload of selecting the images by a doctor and has higher clinical application value.

Drawings

FIG. 1 is a schematic flow chart of an automatic frame selection, classification and recommendation method for coronary artery based on DSA images according to the present invention;

FIG. 2 is a schematic flow chart of the network training of the automatic frame selection module of the present invention;

FIG. 3 is a flow diagram of network training of the automatic classification model of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present disclosure more clear, the technical solutions of the embodiments of the present disclosure will be described clearly and completely with reference to the drawings in the embodiments of the present disclosure, and it is obvious that the described embodiments are only a part of the embodiments of the present disclosure, not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments disclosed herein without making any creative effort, shall fall within the protection scope of the present disclosure.

The terms "first," "second," "third," "fourth," and the like in the description and in the claims of the present disclosure and in the drawings described above, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the disclosure described herein are capable of operation in sequences other than those illustrated or otherwise described herein.

It should be understood that, in various embodiments of the present disclosure, the sequence numbers of the processes do not mean the execution sequence, and the execution sequence of the processes should be determined by the functions and the inherent logic of the processes, and should not constitute any limitation on the implementation process of the embodiments of the present disclosure.

It should be understood that in the present disclosure, "including" and "having" and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

It should be understood that in the present disclosure, "plurality" means two or more. "and/or" is merely an association relationship describing an associated object, meaning that there may be three relationships, for example, and/or B, may mean: a exists alone, A and B exist simultaneously, and B exists alone. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship. "comprises A, B and C" and "comprises A, B, C" means that A, B, C all comprise, "comprises A, B or C" means that one of three A, B, C is comprised, "comprises A, B and/or C" means that any 1 or any 2 or 3 of three A, B, C are comprised.

It should be understood that in this disclosure, "B corresponding to a", "a corresponds to B", or "B corresponds to a" means that B is associated with a, from which B can be determined. Determining B from a does not mean determining B from a alone, but may also be determined from a and/or other information. And the matching of A and B means that the similarity of A and B is greater than or equal to a preset threshold value.

As used herein, if can be interpreted as "at … …" or "at … …" or "in response to a determination" or "in response to a detection", depending on the context.

The technical solution of the present disclosure is explained in detail with specific examples below. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments.

The DSA image-based coronary artery automatic frame selection and classification recommendation method and device can efficiently and accurately complete key frame image extraction, blood vessel category identification and recommendation of an optimal image pair suitable for reconstruction aiming at the DSA image of the coronary artery. The main purpose is to deal with the complicated work that when a doctor conducts coronary vessel reconstruction, the doctor needs to browse a large number of images one by one and then select a specific vessel image suitable for reconstruction.

Referring to fig. 1, fig. 1 schematically illustrates a main flow of an automatic frame selection and classification recommendation method for coronary artery based on DSA images. As shown in fig. 1, the method for automatically selecting frames and classifying and recommending coronary artery based on DSA images in this embodiment includes:

step S1: and acquiring a DSA image group to be detected, and identifying each DICOM image in the DSA image group to be detected by adopting a pre-constructed automatic frame selection model so as to acquire a key frame image.

Specifically, standard catheters may be used to catheterize through femoral or radial catheters, and digitally record coronary angiograms to generate DICOM images. And identifying each DICOM image by using the trained automatic frame selection model so as to acquire a key frame image. The automatic frame selection model is a two-classification neural network model. The keyframe image is a relatively high quality image with the contrast agent fully filled, the contrast agent clearly visible, and in end-diastole. The output of the automatic frame selection model is the probability value that each frame of DICOM image frame belongs to the key frame image, and the probability values of a plurality of frames before and after the occurrence position of the key frame image are relatively high, so the DICOM image frame with the highest probability value at the occurrence position of the key frame image can be selected as the key frame image at the position.

Step S2: and classifying the blood vessel types of the key frame images based on a pre-constructed automatic classification model so as to obtain the probability response values of the blood vessel types corresponding to the key frame images.

Specifically, all the key frame images are input into a trained automatic classification model to classify the key frame images. The automatic classification model is a four-classification neural network model. The blood vessel categories include LAD (left anterior descending branch), LCX (left circumflex), RCA (right coronary artery), and OTHERS, which are blood vessel categories other than LAD, LCX, RCA. When the key frame images are classified, the probability response values of the key frame images belonging to each blood vessel category are output by the automatic classification module, wherein the blood vessel category corresponding to the highest probability response value is the blood vessel category of the key frame images, and the highest probability response value is the probability response value of the key frame images.

And step S3: and (4) arranging the images according to the descending order of the probability response values and forming an image pair by every two key frame images from back to front.

And step S4: and scoring the image frame image pairs based on the mean value of the probability response values corresponding to the image pairs and the imaging angle difference of the corresponding image pairs, and recommending the image pairs with the highest score.

Specifically, the average of the probability response values corresponding to the image pair is an average obtained by dividing the sum of the probability response values of the two key frame images by 2. The imaging angle difference is the difference of the imaging angles of the image pairs corresponding to the two key frame images.

The score may be calculated according to the recommendation method shown in equation (1):

z＝a×x+(1-a)×y (1)

The step of determining the proportionality coefficient a comprises the following steps: acquiring a plurality of image pairs for reconstruction; and sequencing according to the reconstruction quality of the image pair, and adjusting a scale coefficient by taking the sequencing as a reference so that a grading sequencing result calculated by the recommendation method is consistent with the actual sequencing.

After the scoring result of each image pair is obtained, the image pair with the highest score is recommended to the doctor.

Referring to fig. 2, fig. 2 illustrates a network training step of an auto frame selection module. As shown in fig. 2, the network training step of the "automatic frame selection model" includes:

step S11: DICOM images generated based on coronary angiography are acquired.

Step S12: and setting DICOM image frames with full contrast, clear visible contrast agent and in end diastole in the DICOM image as key frame images, and forming a data record table.

This step may be performed by a clinician expert group reading DICOM images, reading key frame images of each image, and recording to form a datalog.

Step S13: and traversing the data record table to select the key frame image and set the label of the key frame image to be 1, and select the non-key frame image and set the label of the non-key frame image to be 0.

Specifically, all key frame images are fetched and set to a label of 1. And selecting the non-key frame image according to a certain rule. The label of the non-key frame image is set to 0. The non-key frame image comprises three parts, wherein one of the first 10 frames in the randomly selected DICOM image is an image frame with incomplete contrast agent filling. Randomly selecting one of the front 5 frames of the key frame image, wherein the part is the image frame in the diastole period in front of the key frame image. Randomly selecting one frame of the next 5 frames of the key frame image, wherein the part is the image frame of the contraction period after the key frame image. The 3 parts do not overlap with each other and do not exceed the normal sequence number range of the image sequence, and the purpose of introducing randomness is to cover the condition that the image is a non-key frame image as much as possible and avoid the limitation of subjective judgment. After the key frame image and the non-key frame image are selected, the step of adjusting the contrast of the images according to the window width and the window level and adjusting the DICOM images to be fixed in size is needed.

Step S14: and forming an automatic frame selection data set based on the key frame images and the non-key frame images and training an automatic frame selection model by adopting a machine learning algorithm.

Referring to FIG. 3, FIG. 3 illustrates the network training steps of an automatic classification model. As shown in fig. 3, the network training step of the "automatic classification model" includes:

step S21: DICOM images generated based on coronary angiography are acquired.

Step S22: and reading the DICOM image frames with full contrast, clear visible contrast agent and in the end diastole in the DICOM image to be used as key frame images.

Step S23: and dividing the key frame images into LAD, LCX, RCA and OTHERS according to the blood vessel category and respectively giving image labels to generate a blood vessel classification data set.

Specifically, the image labels corresponding to the blood vessel classes LAD, LCX, RCA, and OTHERS are [1,0,0,0], [0,1,0,0], [0,0,1,0], [0,0,0,1], respectively. Similarly, the step of resizing the image according to the window width level may be further included when generating the blood vessel classification dataset, so as to resize the image into a fixed size.

Step S24: and training an automatic classification model based on the blood vessel classification data set by adopting a machine learning algorithm.

The present implementation also provides a storage device, in which a plurality of programs are stored, the programs being suitable for being loaded and executed by a processor to implement the DSA image-based coronary artery automatic frame selection and classification recommendation method.

The present embodiment also provides a control apparatus, including: a processor adapted to execute various programs; a storage device adapted to store a plurality of programs; the program is suitable for being loaded and executed by a processor to realize the DSA image-based coronary artery automatic frame selection and classification recommendation method.

The following describes details of the DSA image-based coronary artery automatic frame selection classification recommendation method according to the present invention with reference to a specific embodiment.

1. Obtaining DSA images

A DICOM image is generated by catheterizing through femoral or radial catheters using standard catheters and digitally recording coronary angiograms. Using the anonymization tool, personal patient information in DICOM format is deleted.

2. Marking DICOM images

In order to generate key frame images and category labels of DICOM images, a clinician expert group reads an angiography sequence, the angiography sequence is completely full of contrast agent, the contrast agent is clearly visible, DICOM image frames in end diastole are used as key frame images and are recorded in a data recording list, images are placed into corresponding folders according to main blood vessel categories suitable for reconstruction while reading the images, the folder organization structures classified according to the blood vessel categories are LAD, LCX, RCA and OTHERS, the upper level takes a patient as a unit, and all DICOM images contained in the patient are classified into the four folders according to rules, wherein the specific rules are as follows: the images with clear and visible contrast agents and suitable for reconstruction of angles and blood vessel overlapping degrees are respectively placed into three folders of LAD, LCX and RCA according to the blood vessel types, and other images with low image quality and unsuitable for reconstruction and other images which are not coronary arteries, including other part images, reports or images only with guide wires, and the like are placed into an OTHERS folder. The final data available includes a list of tertiary structures for the patient, the blood vessel type, the specific image, and a data record for key frame images.

3. Data set construction

Because the training data used by the neural network is the frame image in the DICOM image sequence, the data set construction needs to be performed on the basis of the original folder and the data record table. Firstly, data for training and searching for key frame images needs to be constructed, corresponding images are extracted according to key frame sequence numbers in a data record table, images of non-key frames are selected according to a certain rule, and data in an OTHERS folder are not used. Again, data for training the classification recommendations is constructed using data from all four folders, with the OTHERS being directly selectable if it contains a single image or non-DICOM video data.

4. Model construction and training

The construction process of the model is mainly divided into three steps.

Firstly, an automatic frame selection data set is constructed and used for training an automatic frame selection model. The specific details of the automatic frame selection data set construction include that file data are traversed according to a table, each frame of image is read in and adjusted according to a window width and a window level, then corresponding images are extracted according to key frame image sequence numbers recorded in a data recording table to serve as key frame images, non-key frame images are selected and divided into three parts, the first part is an image frame with incomplete contrast agent filling, the first part is defined as randomly selecting one frame in the first 10 frames of the DICOM image, the second part is an image frame in the diastole before the key frame, the second part is defined as randomly selecting one frame in the first 5 frames from the key frame image, the third part is an image frame in the systole after the key frame image, and the third part is defined as randomly selecting one frame in the last 5 frames of the key frame image. After the key frame images and the non-key frame images are well arranged, each image is marked according to the label that the key frame images are 1 and the non-key frame images are 0, and an automatic frame selection data set is generated. The method comprises the steps of training a binary classification model on an automatic frame selection data set by using a classical neural network structure, inputting images with the size of 224 multiplied by 224, and outputting a binary classification probability value list with the sum of the two items being 1.

And secondly, constructing a blood vessel classification data set for training an automatic classification model. The construction of the blood vessel classification data set needs to add OTHERS categories on the basis of the automatic frame selection data set, and the selected images are respectively given to [1,0,0,0], [0,1,0,0], [0,0,1,0], [0,0,0,1] as image labels according to LAD, LCX, RCA and OTHERS, so as to generate the blood vessel classification data set. The method comprises the steps of utilizing a classical neural network structure to conduct four-classification model training on a blood vessel classification data set, inputting images which are unified to be 224 multiplied by 224, outputting a four-classification probability value list, and enabling the sum of four items to be 1.

And thirdly, configuring parameters of a proportionality coefficient in the image recommendation method. The data required by the parameter determination needs to be constructed by taking a patient as a unit, all image data which can be used for reconstruction of the patient needs to be screened out, each image is ranked according to the image quality, the optimal image pair is given, and then the scaling factor can be adjusted by taking the manual ranking as a reference. After the image pair is obtained, the mean value of the probability response values of the image pair and the imaging angle difference are calculated, the score is obtained by using the formula (1), and finally the image pair with the highest score is taken as the best recommended image pair. In the process, the proportion coefficient is continuously adjusted, so that the scoring ranking result calculated by the recommendation method is consistent with the actual ranking, and the proportion coefficient is determined.

5. Test flow

Giving a DICOM image of a patient, predicting each frame of each DICOM image by using a trained automatic classification model, giving a probability value of each frame, and acquiring a key frame image; classifying all the key frame images by using a trained automatic classification model, acquiring probability response values of the key frame images corresponding to the blood vessel categories, sequencing the probability response values to form image pairs, calculating scores by using a formula (1), extracting the image pairs with the highest scores, and recommending the image pairs to doctors.

And finishing the recommendation work of the optimal image pair.

In conclusion, the method combines the implementation processes of frame selection, classification and image pair recommendation, can automatically, quickly and accurately identify the image pair suitable for reconstruction in the DSA images of the coronary artery, reduces the workload of selecting the images by doctors, and has higher clinical application value.

The present invention may be methods, apparatus, systems and/or computer program products. The computer program product may include a computer-readable storage medium having computer-readable program instructions embodied therewith for carrying out aspects of the invention.

The computer readable storage medium may be a tangible device that can hold and store the instructions for use by the instruction execution device. The computer readable storage medium may be, for example, but not limited to, an electronic memory device, a magnetic memory device, an optical memory device, an electromagnetic memory device, a semiconductor memory device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a Static Random Access Memory (SRAM), a portable compact disc read-only memory (CD-ROM), a Digital Versatile Disc (DVD), a memory stick, a floppy disk, a mechanical coding device, such as punch cards or in-groove projection structures having instructions stored thereon, and any suitable combination of the foregoing. Computer-readable storage media as used herein is not to be construed as transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission medium (e.g., optical pulses through a fiber optic cable), or electrical signals transmitted through electrical wires.

The computer-readable program instructions described herein may be downloaded from a computer-readable storage medium to a respective computing/processing device, or to an external computer or external storage device via a network, such as the internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers and/or edge servers. The network adapter card or network interface in each computing/processing device receives computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in the respective computing/processing device.

The computer program instructions for carrying out operations of the present invention may be assembler instructions, instruction Set Architecture (ISA) instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider). In some embodiments, aspects of the present invention are implemented by personalizing an electronic circuit, such as a programmable logic circuit, a Field Programmable Gate Array (FPGA), or a Programmable Logic Array (PLA), with state information of computer-readable program instructions, which can execute the computer-readable program instructions.

Aspects of the present invention are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer readable program instructions.

These computer-readable program instructions may be provided to a processing unit of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processing unit of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks. These computer-readable program instructions may also be stored in a computer-readable storage medium that can direct a computer, programmable data processing apparatus, and/or other devices to function in a particular manner, such that the computer-readable medium storing the instructions comprises an article of manufacture including instructions which implement the function/act specified in the flowchart and/or block diagram block or blocks.

The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer, other programmable apparatus or other devices implement the functions/acts specified in the flowchart and/or block diagram block or blocks.

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems that perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

It is noted that, unless expressly stated otherwise, all features disclosed in this specification (including any accompanying claims, abstract and drawings) may be replaced by alternative features serving the same, equivalent or similar purpose. Thus, unless expressly stated otherwise, each feature disclosed is one example only of a generic series of equivalent or similar features. Where used, further, preferably, still further and more preferably is a brief introduction to the description of the other embodiment based on the foregoing embodiment, the combination of the contents of the further, preferably, still further or more preferably back strap with the foregoing embodiment being a complete construction of the other embodiment. Several further, preferred, still further or more preferred arrangements of the back tape of the same embodiment may be combined in any combination to form a further embodiment.

It will be appreciated by persons skilled in the art that the embodiments of the invention described above and shown in the drawings are given by way of example only and are not limiting of the invention. The objects of the invention have been fully and effectively accomplished. The functional and structural principles of the present invention have been shown and described in the examples, and any variations or modifications of the embodiments of the present invention may be made without departing from the principles.

Finally, it should be noted that: the above embodiments are only used for illustrating the technical solutions of the present disclosure, and not for limiting the same; while the present disclosure has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art will understand that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present disclosure.

Claims

1. An automatic frame selection, classification and recommendation method for coronary artery based on DSA images is characterized by comprising the following steps:

obtaining a DSA image group to be detected, and identifying each DICOM image in the DSA image group to be detected by adopting a pre-constructed automatic frame selection model so as to obtain a key frame image;

arranging the key frame images in descending order according to the probability response values and forming an image pair by every two key frame images from back to front;

and scoring the image frame image pairs based on the mean value of the probability response values corresponding to the image pairs and the imaging angle difference of the corresponding image pairs, and recommending the image pairs with the highest score.

2. The DSA image-based automatic frame selection and classification recommendation method for coronary artery according to claim 1, wherein the automatic frame selection model is a two-classification neural network model, the automatic classification model is a four-classification neural network model, the blood vessel categories include LAD, LCX, RCA and OTHERS, and the OTHERS are blood vessel categories other than LAD, LCX and RCA.

3. The method for automatically selecting frames and classifying and recommending coronary artery based on DSA image according to claim 2, wherein the network training step of the "automatic frame selection model" comprises:

acquiring a DICOM image generated based on coronary angiography;

setting DICOM image frames in which contrast is full, contrast agent is clear and visible and which are in end diastole in the DICOM image as key frame images, and forming a data record table;

4. The DSA-image-based coronary artery automatic frame selection classification recommendation method of claim 3, wherein the non-key frame images comprise one of the first 10 frames, one of the first 5 frames and one of the last 5 frames of the key frame images, which are randomly selected from the DICOM image.

5. The method for automatically selecting frames and classifying and recommending coronary arteries based on DSA images as claimed in claim 2, wherein the step of network training of the "automatic classification model" includes:

acquiring a DICOM image generated based on coronary angiography;

reading DICOM image frames which are full of contrast, clear and visible with contrast agent and in end diastole in the DICOM images and setting the DICOM image frames as key frame images;

training the automatic classification model based on the blood vessel classification dataset and using a machine learning algorithm.

6. The DSA image-based coronary artery automatic frame selection and classification recommendation method as claimed in claim 3 or 5, further comprising the step of adjusting contrast of the DICOM image according to window width window level and adjusting the DICOM image to a fixed size.

7. The method for automatically selecting frames and classifying and recommending coronary arteries based on DSA image group as claimed in claim 1, wherein the step of scoring the image frame image pairs based on the mean of the probability response values corresponding to the image pairs and the imaging angle difference of the corresponding image pairs comprises:

the score is calculated according to the recommendation method shown below:

z＝a×x+(1-a)×y

8. The DSA image-based coronary artery automatic frame selection classification recommendation method of claim 7, the step of determining the a comprising:

acquiring a plurality of image pairs for reconstruction;

9. A storage device having stored therein a plurality of programs, wherein the programs are adapted to be loaded and executed by a processor to implement the DSA image based coronary artery automated frame selection classification recommendation method of any one of claims 1-8.

10. A control device, comprising:

a processor adapted to execute various programs;

a storage device adapted to store a plurality of programs;

characterized in that the program is adapted to be loaded and executed by a processor to implement the DSA image-based coronary artery automated frame selection classification recommendation method according to any one of claims 1 to 8.