Ultrasonic image processing method and device, equipment, storage medium and program product

Abstract

The invention provides an ultrasonic image processing method and device, equipment, a storage medium and a program product, and the method comprises the steps: obtaining an ultrasonic image containing a target organ; and inputting the ultrasonic image into an image rating model to obtain an image rating of the ultrasonic image output by the image rating model. According to the scheme, the ultrasonic image is scored through two dimensions of organ integrity and image quality, so that the image rating result is more comprehensive and objective, and the quality control effect of the ultrasonic image is improved.

Description

Technical Field

[0001] This application relates to the field of image processing technology. Specifically, it relates to an ultrasonic image processing method, device, equipment, storage medium and program product. Background Art

[0002] Ultrasonic examination is a non-invasive examination method that utilizes the physical properties of ultrasonic waves. This technology is mainly based on the propagation characteristics of ultrasonic waves in various organs of the human body. When ultrasonic waves encounter different tissue interfaces, phenomena such as reflection, scattering, and refraction will occur. These phenomena are received by ultrasonic equipment and converted into images, thereby helping doctors observe and diagnose the internal organs and structures of the human body. Therefore, the image quality of ultrasonic images directly affects the accuracy of clinical diagnosis.

[0003] Currently, related technologies mainly rely on image classification networks to identify standard sections of spleen ultrasonic images. However, image classification networks are mainly used to determine whether an image belongs to a certain category, such as whether it is a standard section, but they cannot deeply analyze multiple dimensions of image quality, and the accuracy of image quality assessment is poor. Summary of the Invention

[0004] The purpose of the embodiments of this application is to provide an ultrasonic image processing method, device, equipment, storage medium and program product to solve the above problems.

[0005] In a first aspect, the embodiments of this application provide an ultrasonic image processing method, the method includes: obtaining an ultrasonic image containing a target organ; inputting the ultrasonic image into an image rating model to obtain the image rating of the ultrasonic image output by the image rating model;

[0006] Wherein, the image rating model includes: a first scoring sub-network and a second scoring sub-network respectively used to obtain an organ integrity score and an image quality score; the image rating is determined based on the organ integrity score and the image quality score.

[0007] In the implementation process of the above solution, the ultrasonic image is scored from two dimensions of organ integrity and image quality, making the image rating result more comprehensive and objective, which is beneficial to improving the quality control effect of ultrasonic images; on the other hand, the ultrasonic image is automatically rated by the image rating model, reducing the rating deviation caused by human factors, which is beneficial to improving the accuracy and objectivity of the ultrasonic image quality control rating; on the other hand, automatically rating the ultrasonic image by the image rating model is beneficial to improving the efficiency of the ultrasonic image quality control rating.

[0008] In one implementation of the first aspect, obtaining the image quality score by using the second sub-scoring network includes: inputting the ultrasound image into the spatial domain scoring module in the second sub-scoring network to obtain a first score output by the spatial domain scoring module; wherein, the first score is used to represent the probability that the ultrasound image is a high-quality image; inputting the frequency domain image of the ultrasound image into the frequency domain scoring module in the second sub-scoring network to obtain a second score output by the frequency domain scoring module; wherein, the second score is used to represent the probability that the frequency domain image is a high-quality image; inputting the first score and the second score into the quality score calculation module in the second sub-scoring network to obtain the image quality score of the ultrasound image.

[0009] In the implementation process of the above solution, by the spatial domain scoring module and the frequency domain scoring module, the image quality is scored from two perspectives of the spatial domain and the frequency domain respectively. The obtained image quality score can more comprehensively evaluate the image quality, reduce the one-sidedness and error that may be brought by a single-angle evaluation, and improve the objectivity and accuracy of the ultrasound image evaluation.

[0010] In one implementation of the first aspect, the second sub-scoring network includes a quality score calculation module, and the quality score calculation module includes a first channel and a second channel;

[0011] Obtaining the image quality score by using the second sub-scoring network includes: inputting the ultrasound image and the frequency domain image of the ultrasound image into the first channel and the second channel respectively to obtain a first score output by the first channel and a second score output by the second channel; wherein, the first score is used to represent the probability that the ultrasound image is a high-quality image; the second score is used to represent the probability that the frequency domain image is a high-quality image; inputting the first score and the second score into the quality score calculation module in the second sub-scoring network to obtain the image quality score of the ultrasound image.

[0012] In the implementation process of the above solution, the first channel and the second channel of the quality score calculation module can independently process the spatial domain image and the frequency domain image respectively, which is beneficial to improving the rating efficiency of image rating; on the other hand, through the quality score calculation module, the automation of image quality evaluation can be realized without manual intervention, saving time and labor costs, and being beneficial to improving the image processing efficiency of the above ultrasound image processing method.

[0013] In one implementation of the first aspect, obtaining the organ integrity score by using the first sub-scoring network includes: inputting the ultrasound image into the target detection module in the first sub-scoring network to obtain the detection confidence of the target organ; inputting the detection confidence into the integrity score calculation module in the first sub-scoring network to obtain the organ integrity score.

[0014] In the implementation process of the above solution, the organ integrity score of the target organ is calculated through the detection confidence obtained by the target detection module, and the organ integrity assessment is converted into specific numerical indicators, making the image quality control rating more objective and conducive to improving the accuracy and reliability of the image rating.

[0015] In one implementation manner of the first aspect, obtaining the organ integrity score by using the first scoring sub-network includes: inputting the ultrasound image into the target detection module in the first scoring sub-network to obtain the detection confidence of at least one target structure in the target organ; inputting the detection confidence of at least one target structure into the integrity score calculation module in the first scoring sub-network to obtain the organ integrity score.

[0016] In the implementation process of the above solution, the integrity score can be calculated by comprehensively considering the detection confidence of at least one target structure in the target organ, which can more comprehensively reflect the integrity of the target organ in the ultrasound image, reduce misjudgment caused by the absence or inaccurate detection of a single structure, and is conducive to improving the accuracy rate of the image rating; on the other hand, calculating the integrity score through the detection confidence of multiple target structures can achieve defect tracing for ultrasound images with incomplete or defective key structures, help obtain higher-quality ultrasound images, and is conducive to improving the quality control effect of the above ultrasound image processing method.

[0017] In one implementation manner of the first aspect, the image rating model further includes: a third scoring sub-network for obtaining the organ standardization score; the image rating is determined based on the organ integrity score, the image quality score, and the organ standardization score;

[0018] Obtaining the organ standardization score by using the third scoring sub-network includes: inputting the ultrasound image and the position information of the target organ into the cropping module in the third scoring sub-network to obtain the target organ image cropped by the cropping module in the ultrasound image; wherein, the position information of the target organ is obtained by the target detection module in the first scoring sub-network; inputting the target organ image into the standardization score calculation module in the third scoring sub-network to obtain the organ standardization score calculated by the standardization score calculation module based on the similarity between the target organ image and the standard organ image in the knowledge base.

[0019] In the implementation process of the above solution, on the one hand, jointly determining the image rating based on the organ integrity score, the image quality score, and the image standardization score is conducive to improving the objectivity and accuracy of the image rating; on the other hand, calculating the organ standardization score through the similarity between the target organ image and the standard organ image in the knowledge base enables the above ultrasonic image processing method to measure the similarity degree between the target organ and the standard organ in the ultrasonic image through specific similarity values, avoiding the deviation and inconsistency that may be brought by manual subjective judgment, and is conducive to improving the objectivity and accuracy of ultrasonic image quality control.

[0020] In one implementation manner of the first aspect, the image rating model further includes: a third scoring sub-network for obtaining the organ standardization score; the image rating is determined based on the organ integrity score, the image quality score, and the organ standardization score;

[0021] Obtaining the organ standardization score by using the third scoring sub-network includes: inputting the ultrasonic image and the position information of at least one target structure into the cropping module in the third scoring sub-network to obtain the target structure image cropped by the cropping module in the ultrasonic image; wherein, the position information of the target structure is obtained by the target detection module in the first scoring sub-network; inputting the target structure image into the standardization score calculation module in the third scoring sub-network to obtain the organ standardization score calculated by the standardization score calculation module based on the similarity between the target structure image and the standard structure image in the knowledge base.

[0022] In the implementation process of the above solution, the organ standardization score can be calculated through the similarity between the target structure in the target organ and the standard structure in the knowledge base. On the one hand, calculating the organ standardization score through the similarity between the key structure and its standard structure is conducive to improving the objectivity and accuracy of the organ standardization score; on the other hand, it can realize the defect tracing of ultrasonic images with non-standard or defective key structures, help to obtain higher-quality ultrasonic images, and thus improve the quality control effect of the above ultrasonic image processing method.

[0023] In one implementation manner of the first aspect, the target organ includes the spleen; the target structures include: the spleen capsule, the splenic hilum, and the splenic vein.

[0024] In the implementation process of the above solution, the ultrasonic image processing method can be applied to the spleen ultrasonic scanning scenario, which is conducive to improving the quality inspection efficiency and quality control effect of ultrasonic images in the spleen ultrasonic scanning scenario.

[0025] In an implementation of the first aspect, the image rating model further includes: an image rating module; the image rating module is configured to obtain the image rating of the ultrasound image by using a decision tree based on the organ integrity score, the image quality score, and the organ standardization score.

[0026] In the implementation process of the above solution, determining the rating result and the rating basis of the ultrasound image by using a decision tree based on the organ integrity score, the image quality score, and the organ standardization score is beneficial to improving the efficiency of ultrasound image quality assessment.

[0027] In an implementation of the first aspect, the method further includes:

[0028] If the image quality rating is greater than a preset quality rating threshold, output a suggestion to retain the image;

[0029] And / or, if the image quality rating is not greater than the preset quality rating threshold, output a suggestion to print the image.

[0030] In the implementation process of the above solution, the ultrasound image processing method can be applied to the ultrasound image retention scenario and the ultrasound printing scenario, which is beneficial to improving the quality inspection efficiency of ultrasound images in the ultrasound image retention scenario and the ultrasound printing scenario; on the other hand, it enables the above ultrasound image processing method to be applied to more application scenarios, which is beneficial to improving the adaptability of the above ultrasound image processing method.

[0031] In a second aspect, an embodiment of the present application provides an ultrasound image processing device, and the device includes:

[0032] An ultrasound image acquisition unit, configured to acquire an ultrasound image including a target organ;

[0033] An image rating unit, configured to input the ultrasound image into an image rating model and obtain the image rating of the ultrasound image output by the image rating model;

[0034] Wherein, the image rating model includes: a first sub-rating network and a second sub-rating network respectively configured to obtain an organ integrity score and an image quality score; the image rating is determined based on the organ integrity score and the image quality score.

[0035] In a third aspect, an embodiment of the present application provides an electronic device, including: a processor, a memory, and a communication bus, wherein the processor and the memory complete communication with each other through the communication bus; computer program instructions executable by the processor are stored in the memory, and when the computer program instructions are read and run by the processor, the method provided by the first aspect or any possible implementation of the first aspect is executed.

[0036] Fourthly, an embodiment of the present application provides a computer-readable storage medium, on which computer program instructions are stored. When the computer program instructions are read and run by a processor, the method provided by the first aspect or any possible implementation manner of the first aspect is executed.

[0037] Fifthly, an embodiment of the present application provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the method provided by the first aspect or any possible implementation manner of the first aspect is implemented.

[0038] Other features and advantages of the present application will be described in the subsequent specification, and part of them will become obvious from the specification, or can be understood by implementing the embodiments of the present application. The objectives and other advantages of the present application can be achieved and obtained through the structures specifically pointed out in the written specification, claims, and drawings. Description of the Drawings

[0039] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation of the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0040] Figure 1 It is a schematic flowchart of the ultrasonic image processing method provided by the embodiment of the present application;

[0041] Figure 2 It is a schematic architecture diagram of the image rating model provided by the embodiment of the present application;

[0042] Figure 3 It is a schematic architecture diagram of the second scoring sub-network provided by the embodiment of the present application;

[0043] Figure 4 It is a schematic architecture diagram of the image rating module provided by the embodiment of the present application;

[0044] Figure 5 It is a schematic structural diagram of the ultrasonic image processing device provided by the embodiment of the present application;

[0045] Figure 6 It is a schematic structural diagram of the electronic device provided by the embodiment of the present application. Detailed Embodiments

[0046] The technical solutions in the embodiments of the present application will be described below with reference to the accompanying drawings in the embodiments of the present application. The following embodiments are only used to illustrate the technical solutions of the present application more clearly, and thus are only examples and cannot be used to limit the protection scope of the present application.

[0047] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above drawings are intended to cover non-exclusive inclusion.

[0048] In the description of the embodiments of the present application, technical terms such as "first" and "second" are only used to distinguish different objects and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity, specific order or primary-secondary relationship of the indicated technical features. In the description of the embodiments of the present application, "a plurality of" means more than two unless otherwise specifically defined.

[0049] Referring to "embodiments" herein means that specific features, structures or characteristics described in connection with the embodiments can be included in at least one embodiment of the present application. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0050] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of associated objects, indicating that three relationships can exist, for example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " herein generally represents an "or" relationship between the associated objects before and after.

[0051] The embodiments of the present application provide an ultrasonic image processing method. This method scores ultrasonic images from two dimensions of organ integrity and image quality, making the image rating results more comprehensive and objective, which is beneficial to improving the quality control effect of ultrasonic images; on the other hand, automatically rating ultrasonic images through an image rating model reduces the rating deviation caused by human factors and is beneficial to improving the accuracy and objectivity of ultrasonic image quality control ratings; on the other hand, automatically rating ultrasonic images through an image rating model is beneficial to improving the efficiency of ultrasonic image quality control ratings.

[0052] Please refer to Figure 1Schematic flowchart of the ultrasonic image processing method provided by the embodiments of the present application. The ultrasonic image processing method provided by the embodiments of the present application can be applied to an electronic device, which can include physical devices such as a server, a PC, a tablet computer, or a smart phone, or can also be a virtual device such as a virtual machine or a container. The electronic device can be a single device, or a combination of multiple devices or a cluster of a large number of devices. The above ultrasonic image processing method can include:

[0053] Step S110: Obtain an ultrasonic image containing a target organ;

[0054] The ultrasonic image obtained in step S110 above can be an ultrasonic image collected by an ultrasonic doctor through an ultrasonic device received by the electronic device executing the above ultrasonic image processing method, or can also be an ultrasonic image obtained when a doctor performs an ultrasonic examination on a patient during a daily clinical examination.

[0055] The above target organ refers to an organ that an ultrasonic doctor or other observer wants to observe, such as the spleen, kidney, liver, etc. The ultrasonic image containing the target organ can be a set of ultrasonic images that capture the target organ that the ultrasonic doctor or other observer wants to observe.

[0056] Step S120: Input the ultrasonic image into an image rating model to obtain the image rating of the ultrasonic image output by the image rating model.

[0057] Please refer to Figure 2 , and the above image rating model will be introduced below. The image rating model 200 can include:

[0058] The first scoring sub-network 210 is used to obtain an organ integrity score based on the ultrasonic image;

[0059] The second scoring sub-network 220 is used to obtain an image quality score based on the ultrasonic image.

[0060] Optionally, the above first scoring sub-network 210 includes: a target detection module 211 and an integrity score calculation module 212, where:

[0061] The target detection module 211 is used to perform target detection on the target organ in the ultrasonic image to obtain the detection confidence of the target organ;

[0062] The integrity score calculation module 212 is used to calculate the organ integrity score based on the detection confidence of the target organ.

[0063] In the above step S120, the organ integrity score is obtained by using the first scoring sub-network 210 in the image rating model 200, including: inputting the ultrasonic image into the target detection module 211 in the first scoring sub-network 210 to obtain the detection confidence of the target organ; inputting the detection confidence into the integrity score calculation module 212 in the first scoring sub-network 210 to obtain the organ integrity score.

[0064] The above-mentioned organ integrity refers to the degree of completeness of the target organ presented in the ultrasonic image, that is, whether all parts and structures of the target organ can be clearly and comprehensively displayed without obvious missing, occlusion or deformation. Specifically, the organ integrity mainly includes the following aspects:

[0065] (1) Morphological integrity: The outline of the organ is complete without irregular morphological changes or partial loss;

[0066] (2) Structural integrity: The internal structure of the organ can be clearly displayed without being blocked by other tissues or substances;

[0067] (3) Clear boundary: The boundary between the organ and the surrounding tissues is clearly distinguishable without blurring or fusion;

[0068] (4) No obvious artifact interference: There are no artifacts in the image caused by factors such as equipment failure and patient position, which may cover part of the structure of the organ and thus affect the integrity of the organ.

[0069] The above target detection module 211 can adopt a target detection model, such as the YoLo target detection model. When training the target detection model, ultrasonic images from multiple medical centers and various ultrasonic devices can be used to train the target detection model. The ultrasonic images obtained from different medical centers and ultrasonic devices vary in image quality, resolution, imaging angle, noise level, etc., so that the training data has multi-source characteristics. Using multi-source data can enrich the diversity of training data, enabling the target detection model to access a wider range of image features and variations, which helps to improve the adaptability and generalization ability of the target detection model to different imaging conditions and the robustness of the model in the face of various complex situations.

[0070] In addition, the training images of the above target detection model can be annotated by doctors or experts with more than 5 years of experience to improve the annotation accuracy and thus improve the model training effect.

[0071] The above organ integrity score is proportional to the detection confidence. The higher the detection confidence of the target organ, the higher the organ integrity score.

[0072] The above solution calculates the organ integrity score of the target organ through the detection confidence obtained by the target detection module 211, converts the organ integrity assessment into a specific numerical index, makes the image quality control rating more objective, and is conducive to improving the accuracy and reliability of the image rating.

[0073] Optionally, the above step S120 uses the first scoring sub-network 210 to obtain the organ integrity score, including: inputting the ultrasonic image into the target detection module 211 in the first scoring sub-network 210 to obtain the detection confidence of at least one target structure in the target organ; inputting the detection confidence of at least one target structure into the integrity score calculation module 212 in the first scoring sub-network 210 to obtain the organ integrity score.

[0074] The above target structure may be a key structure of the target organ. For example, when performing an ultrasonic examination on the liver, the key structures of the liver may include the liver parenchyma, hepatic blood vessels (portal vein, hepatic artery, and hepatic vein), and bile ducts (intrahepatic bile ducts and extrahepatic bile ducts), etc. The target detection module 211 may perform target detection on the key structures of the target organ as targets, so as to obtain the detection confidence of the key structures. When the number of target structures is multiple, the integrity score calculation module 212 may calculate the overall detection confidence of the target organ by averaging or weighted averaging (the weight may be the importance of each key structure relative to the target organ) the detection confidences of the multiple target structures, and then calculate the organ integrity score based on the overall detection confidence of the organ. The organ detection confidence may be directly determined as the organ integrity score, or a mapping relationship between the organ detection confidence and the organ integrity score may be configured in advance, and the mapping relationship may be configured such that the organ detection confidence and the organ integrity score are positively correlated.

[0075] The above solution can calculate the integrity score by comprehensively considering the detection confidence of at least one target structure in the target organ, can more comprehensively reflect the integrity of the target organ in the ultrasonic image, reduce misjudgment caused by the absence or inaccurate detection of a single structure, and is conducive to improving the accuracy rate of the image rating; on the other hand, calculating the integrity score through the detection confidence of multiple target structures can realize the defect traceability of ultrasonic images with incomplete or defective key structures, help to obtain higher-quality ultrasonic images, and is conducive to improving the quality control effect of the above ultrasonic image processing method.

[0076] The following introduces the solution for the second scoring sub-network 220 to obtain the image quality score:

[0077] The above-mentioned second evaluation sub-network 220 can evaluate the image quality from two perspectives: the spatial domain (i.e., the space domain) and the frequency domain. In the spatial domain, the image quality can be evaluated from perspectives such as image sharpness, noise, and artifacts. In the frequency domain, the image quality can be evaluated from perspectives such as dynamic range and gain adjustment. The following introduces two alternative implementation manners of the above-mentioned second evaluation sub-network 220:

[0078] The first implementation manner:

[0079] Please refer to Figure 3 , the second evaluation sub-network 220 includes a spatial domain scoring module 221, a frequency domain scoring module 222, and a quality scoring calculation module 223, where:

[0080] The spatial domain scoring module 221 is used to obtain a first score of the ultrasonic image in the spatial domain based on the ultrasonic image; wherein, the first score is used to represent the probability that the ultrasonic image is a high-quality image;

[0081] The frequency domain scoring module 222 is used to obtain a second score of the ultrasonic image in the frequency domain based on the frequency domain image of the ultrasonic image; wherein, the second score is used to represent the probability that the frequency domain image is a high-quality image;

[0082] The quality scoring calculation module 223 is used to calculate the image quality score of the ultrasonic image based on the first score and the second score.

[0083] In the above implementation manner, obtaining the image quality score by using the second evaluation sub-network 220 includes:

[0084] Input the ultrasonic image into the spatial domain scoring module 221 in the second evaluation sub-network 220 to obtain the first score output by the spatial domain scoring module 221; wherein, the first score is used to represent the probability that the ultrasonic image is a high-quality image;

[0085] Input the frequency domain image of the ultrasonic image into the frequency domain scoring module 222 in the second evaluation sub-network 220 to obtain the second score output by the frequency domain scoring module 222; wherein, the second score is used to represent the probability that the frequency domain image is a high-quality image; the frequency domain image can be obtained by performing a Fourier transform on the ultrasonic image;

[0086] Input the first score and the second score into the quality scoring calculation module 223 in the second evaluation sub-network 220 to obtain the image quality score of the ultrasonic image.

[0087] Both the above-mentioned spatial domain scoring module 221 and frequency domain scoring module 222 can adopt a convolutional neural network CNN (Convolutional Neural Networks). When training the convolutional neural network, the training images can be labeled with a binary classification of good / bad. When applying the convolutional neural network to obtain the first score or the second score, the probability that the input image output by the convolutional neural network is a high-quality image can be directly used as the first score or the second score.

[0088] In addition, the above-mentioned training images can be labeled with a binary classification of image quality by doctors with more than 5 years of experience.

[0089] The above-mentioned quality score calculation module 223 can perform an averaging operation on the first score and the second score to obtain the image quality score, or can perform a weighted averaging operation on the first score and the second score to obtain the image quality score. The weights used for weighting can be the weights preset for the first score and the second score. The weights of the first score and the second score can respectively represent the importance of the first score and the second score.

[0090] The above solution scores the image quality from two perspectives of the spatial domain and the frequency domain through the spatial domain scoring module 221 and the frequency domain scoring module 222. The obtained image quality score can more comprehensively evaluate the image quality, reduce the one-sidedness and errors that may be brought by a single-angle evaluation, and improve the objectivity and accuracy of ultrasonic image evaluation.

[0091] The second implementation manner:

[0092] Please refer to Figure 2 , the second score sub-network 220 includes a quality score calculation module 223 and a quality scoring module 224. The quality scoring module 224 includes a first channel and a second channel, where:

[0093] The first channel of the quality scoring module 224 is used to obtain the first score of the ultrasonic image in the spatial domain based on the ultrasonic image;

[0094] The second channel of the quality scoring module 224 is used to obtain the second score of the ultrasonic image in the frequency domain based on the frequency domain image of the ultrasonic image;

[0095] The quality score calculation module 223 is used to calculate the image quality score of the ultrasonic image based on the first score and the second score.

[0096] Optionally, using the second score sub-network 220 to obtain the image quality score includes:

[0097] The ultrasonic image and the frequency-domain image of the ultrasonic image are respectively input into the first channel and the second channel of the quality scoring module 224, and the first score output by the first channel and the second score output by the second channel are respectively obtained; wherein, the first score is used to represent the probability that the ultrasonic image is a high-quality image; the second score is used to represent the probability that the frequency-domain image is a high-quality image.

[0098] The first score and the second score are input into the quality scoring calculation module 223 in the second sub-scoring network 220 to obtain the image quality score of the ultrasonic image.

[0099] The above-mentioned quality scoring module 224 can be a dual-channel convolutional neural network (Dual-Channel Convolutional Neural Network, DC-CNN). The characteristics of the dual-channel convolutional neural network are as follows: two independent input signal channels or parallel convolutional layers are used to process the input data simultaneously. These two channels can receive and process different types of input information at the same time. On the one hand, it can improve the data processing efficiency, and on the other hand, it can enable the network to capture more diverse feature information, which helps to improve the performance of the neural network. The input of the first channel in the above-mentioned dual-channel convolutional neural network DC-CNN can be the ultrasonic image, and the output of the first channel can be the first score of the ultrasonic image in the spatial domain. The input of the second channel in the dual-channel convolutional neural network DC-CNN can be the frequency-domain image of the ultrasonic image, and the output of the second channel can be the second score of the ultrasonic image in the frequency domain.

[0100] The training images of the above-mentioned dual-channel convolutional neural network can also be training images with good / bad binary classification labels, and the training images can also be labeled by doctors with more than 5 years of experience.

[0101] The above-mentioned dual-channel convolutional neural network DC-CNN can adopt a dual-channel convolutional neural network constructed based on the ConvNeXt network. ConvNeXt is an efficient convolutional neural network model. Its design integrates the advantages of traditional convolutional neural networks and Transformer models, aiming to provide more efficient computing performance while retaining high accuracy. The dual-channel convolutional neural network constructed based on the ConvNeXt network has strong feature extraction capabilities, and also has strong robustness and generalization capabilities.

[0102] The first channel and the second channel of the quality scoring module 224 in the above-mentioned solution can independently process the spatial-domain image and the frequency-domain image respectively, which is beneficial to improving the rating efficiency of image rating; on the other hand, through the quality scoring module, the automation of image quality assessment can be realized without manual intervention, saving time and labor costs, which is beneficial to improving the image processing efficiency of the above-mentioned ultrasonic image processing method.

[0103] It can be understood that, in addition to the organ integrity score and the image quality score, the organ standardness score can also be comprehensively considered when rating ultrasound images. The following introduces the scheme for obtaining the organ standardness score:

[0104] The above-mentioned organ standardness refers to the degree of similarity between the target organ in the ultrasound image and the standard structure of the organ. The higher the similarity between the target organ and the standard structure of the organ, the higher the organ standardness score.

[0105] Optionally, as Figure 2 shown, the image rating model 200 may further include:

[0106] A third scoring sub-network 230, configured to obtain an organ standardness score based on the ultrasound image.

[0107] When the image rating model 200 includes the third scoring sub-network 230, the image rating can be determined based on the organ integrity score, the image quality score, and the organ standardness score.

[0108] The above-mentioned third scoring sub-network 230 may include: a cropping module 231 and a standardness score calculation module 232, where:

[0109] The cropping module 231 is configured to crop the target organ image in the ultrasound image based on the position information of the target organ;

[0110] The standardness score calculation module 232 is configured to calculate the similarity between the target organ image and the standard organ image in the knowledge base; and, based on the similarity, calculate the organ standardness score.

[0111] The above step S120 uses the third scoring sub-network 230 to obtain the organ standardness score, including:

[0112] Input the ultrasound image and the position information of the target organ into the cropping module 231 in the third scoring sub-network 230 to obtain the target organ image cropped by the cropping module 231 in the ultrasound image; wherein, the position information of the target organ is obtained by the target detection module 211 in the first scoring sub-network 210;

[0113] Input the target organ image into the standardness score calculation module 232 in the third scoring sub-network 230 to obtain the organ standardness score calculated by the standardness score calculation module 232 based on the similarity between the target organ image and the standard organ image in the knowledge base.

[0114] The above-mentioned standard organ image refers to the standard image corresponding to the target organ in the knowledge base. The knowledge base can include standard images of multiple organs, and the number of standard images for each organ can be multiple. For example, at least 1000 standard images are configured for each organ in the knowledge base. The similarity between the target organ image and the standard organ image in the knowledge base can be the average similarity between the target organ image and multiple standard images corresponding to the organ. In addition, the similarity between the target organ image and the standard organ image can be calculated using cosine similarity.

[0115] On the one hand, the above solution comprehensively determines the image rating by combining the organ integrity score, image quality score, and image standardization score, which is beneficial to improving the objectivity and accuracy of the image rating. On the other hand, by calculating the organ standardization score through the similarity between the target organ image and the standard organ image in the knowledge base, the above ultrasonic image processing method can measure the similarity degree between the target organ in the ultrasonic image and the standard organ through specific similarity values, avoiding the deviation and inconsistency that may be brought by manual subjective judgment, and being beneficial to improving the objectivity and accuracy of ultrasonic image quality control.

[0116] The above-mentioned third scoring sub-network 230 can also calculate the organ standardization score through the similarity between the target structure and the standard structure. The specific solution is as follows:

[0117] The above-mentioned third scoring sub-network 230 can include: a cropping module 231 and a standardization score calculation module 232, where:

[0118] The cropping module 231 is used to crop the target structure image in the ultrasonic image based on the position information of at least one target structure in the target organ;

[0119] The standardization score calculation module 232 is used to calculate the similarity between the target structure image and the standard structure image in the knowledge base; and, based on the similarity, calculate the organ standardization score.

[0120] Optionally, the above method of obtaining the organ standardization score using the third scoring sub-network 230 includes:

[0121] Input the ultrasonic image and the position information of at least one target structure into the cropping module 231 in the third scoring sub-network 230 to obtain the target structure image cropped by the cropping module 231 in the ultrasonic image; where the position information of at least one target structure is obtained by the target detection module 211 in the first scoring sub-network 210;

[0122] Input the target structure image into the standardization score calculation module 232 in the third scoring sub-network 230 to obtain the organ standardization score calculated by the standardization score calculation module 232 based on the similarity between the target structure image and the standard structure image in the knowledge base.

[0123] The above-mentioned standard structure image refers to the standard structure image corresponding to the key structure of the target organ in the knowledge base. The knowledge base can include standard structure images of multiple organs, and each organ can correspond to multiple standard structure images of the target structure. For example, the knowledge base can include standard structure images of multiple organs such as the liver, kidney, and spleen. Among them, the liver can correspond to the standard structure image of the liver parenchyma, the standard structure image of the hepatic blood vessels, and the standard structure image of the bile duct. The number of each standard structure image can be multiple, for example, at least 1000 standard structure images are configured for each key structure.

[0124] When the number of target structures is multiple, the above organ standardness score can be obtained by taking the average of multiple similarities between the multiple target structures and the standard structures in the knowledge base. The similarity between each target structure and the standard structure in the knowledge base can be obtained by taking the average of multiple similarities between the target structure image of each target structure and multiple standard structure images in the knowledge base.

[0125] The above solution can calculate the organ standardness score through the similarity between the target structure in the target organ and the standard structure in the knowledge base. On the one hand, calculating the organ standardness score through the similarity between the key structure and its standard structure is beneficial to improving the objectivity and accuracy of the organ standardness score; on the other hand, it can realize the defect traceability of ultrasonic images with non-standard or defective key structures, which helps to obtain higher-quality ultrasonic images, and then improve the quality control effect of the above ultrasonic image processing method.

[0126] The following introduces the solution for image rating of the above image rating model based on organ integrity score, image quality score, and organ standardness score:

[0127] Optionally, please refer to Figure 2 The above image rating model 200 may further include:

[0128] An image rating module 240, configured to obtain the image rating of the ultrasonic image by using a decision tree based on the organ integrity score, the image quality score, and the organ standardness score. For example, this implementation method:

[0129] Adopt a decision tree as shown in Figure 4 to obtain the image rating of the ultrasonic image, Figure 4 The rating criteria of the decision tree shown are:

[0130] If the organ integrity score S1 is less than 0.5, the ultrasonic image is rated as unqualified;

[0131] If the organ integrity score S1 is greater than or equal to 0.5 and the image quality score S2 is less than 0.7, the ultrasonic image is rated as qualified;

[0132] If the organ integrity score S1 is greater than or equal to 0.5, the image quality score S2 is greater than or equal to 0.7, and the organ standardization score S3 is less than 0.7, the ultrasound image is rated as good;

[0133] If the organ integrity score S1 is greater than or equal to 0.5, the image quality score S2 is greater than or equal to 0.7, and the organ standardization score S3 is greater than or equal to 0.7, the ultrasound image is rated as excellent.

[0134] When using a decision tree to rate ultrasound images, in addition to the rating result, interpretable scoring bases can also be output. For example:

[0135] When rating the ultrasound image as unqualified, output the scoring basis: key structures are missing;

[0136] When rating the ultrasound image as qualified, output the scoring basis: key structures are normally displayed, but the image quality is poor or the parameter conditions are inappropriate;

[0137] When rating the ultrasound image as good, output the scoring basis: key structures are normally displayed, the image quality is good, but the key structure mapping is not standard;

[0138] When rating the ultrasound image as excellent, output the scoring basis: key structures are normally displayed, the image quality is good, and the key structure is standard.

[0139] It can be understood that in addition to the rule-based decision tree shown in Figure 4 The above image rating module 240 can also adopt other forms of decision trees such as decision trees based on machine learning such as random forests and gradient boosting trees.

[0140] In addition, it can be understood that Figure 4 shows a method for image rating based on three dimensions of organ integrity, image quality, and organ standardization. Those skilled in the art should know that when rating images based on organ integrity and image quality, a decision tree shown in Figure 4 can also be used for implementation, and the specific implementation manner is not described in detail in the embodiments of the present application.

[0141] Optionally, the above target organ includes the spleen; the target structures include: the spleen capsule, the splenic hilum, and the splenic vein.

[0142] The following introduces the specific implementation of the above ultrasound image processing method in the quality control scenario of spleen ultrasound images, mainly including the following steps:

[0143] Step 1: Obtain spleen ultrasound scanning images.

[0144] The key structures of the spleen include the spleen capsule, the splenic hilum, and the splenic vein.

[0145] Step 2: Use the first sub - scoring network 210 to obtain the organ integrity score S1.

[0146] The object detection module 211 in the first sub - scoring network 210 can adopt an object detection network based on YoLo (You Only Look Once). The training data of the object detection network can use the spleen ultrasound data collected by multiple medical centers and various ultrasound devices. The training and test data are labeled by doctors with more than 5 years of experience for the key structures of the spleen. The object detection network can output the detection coordinates and detection confidence levels of three key structures: the spleen capsule, the splenic hilum, and the splenic vein. The value range of the detection confidence level is 0 - 1. If a certain key structure is not detected, its detection confidence level is 0.

[0147] The integrity score calculation module 212 in the first sub - scoring network 210 takes the average operation of the detection confidence levels of the three key structures to obtain the organ integrity score S1.

[0148] Step 3: Use the second sub - scoring network 220 to obtain the image quality score S2 of the ultrasound image.

[0149] The quality scoring module 224 in the second sub - scoring network 220 can adopt a dual - channel convolutional neural network DC - CNN constructed based on the ConvNext network. The first channel in the dual - channel convolutional neural network is used to evaluate the ultrasound image quality in the spatial domain, and the evaluation criteria include image clarity, noise, artifacts, etc. The second channel in the dual - channel convolutional neural network is used to evaluate the ultrasound image quality in the frequency domain, and the evaluation criteria include dynamic range, gain adjustment, etc. The input image of the second channel is the frequency - domain image of the ultrasound image, and this frequency - domain image can be obtained by performing a Fourier transform on the ultrasound image.

[0150] The training and test data of the above - mentioned dual - channel convolutional neural network are labeled by doctors with more than 5 years of experience for binary classification of good / bad image quality.

[0151] The quality scoring calculation module 223 in the second sub - scoring network 220 can calculate the image quality score S2 based on the first score output by the first channel and the second score output by the second channel. The value range of the image quality score S2 is 0 - 1.

[0152] Step 4: Use the third sub - scoring network 230 to obtain the organ standardness score S3.

[0153] An expert knowledge base is pre - constructed. Through doctors with more than 5 years of experience, three key structures of the spleen are screened and labeled. 1000 standard structure diagrams with labels are selected for each structure to construct the expert knowledge base.

[0154] After the cropping module 231 in the third evaluation sub-network 230 crops the ultrasound image using the position information obtained by the target detection module 211 in the first evaluation sub-network 210, images of each target structure are obtained. Then, the cosine similarity between the target structure image and the standard structure image in the expert knowledge base is calculated using the standard score calculation module 232 in the second evaluation sub-network 230. Each target structure obtains 1000 similarity values, and the similarity matching score S3_i of each key structure can be obtained through an averaging operation. The value of i is 1, 2, and 3, representing different key structures. If the i-th key structure is not detected in the target detection module 211, then S3_i is 0. The standard score S3 of the key structures in this image is obtained through an averaging operation on S3_1, S3_2, and S3_3.

[0155] Step Five: Use the image rating module 240 as shown in Figure 4 to rate the spleen ultrasound scan image and output the rating result and the basis for the rating.

[0156] The application scenarios of the above ultrasound image processing method are introduced below:

[0157] The first application scenario: Serve as technical support for doctors during diagnosis.

[0158] During clinical diagnosis by doctors, the ultrasound image rating result obtained based on the above ultrasound image processing method can assist doctors in making clinical diagnoses, which helps doctors make diagnoses more quickly and accurately, thereby improving the treatment effect and satisfaction of patients.

[0159] The second application scenario: Provide image retention suggestions for the image acquisition operator during ultrasound imaging.

[0160] Optionally, the above ultrasound image processing method further includes: if the image quality rating is greater than a preset quality rating threshold, output a suggestion to retain the image.

[0161] In the above application scenario, after the ultrasound image acquisition operator acquires an image, the above ultrasound image processing method can quickly rate the quality of the acquired ultrasound image. If the quality rating of the ultrasound image is greater than a preset quality rating threshold, for example, the minimum requirement for the image rating is good, then a suggestion to retain the image is output to the ultrasound image acquisition operator, thereby improving the image acquisition effect of ultrasound imaging.

[0162] The third application scenario: Provide imaging suggestions for the image acquisition operator during ultrasound imaging.

[0163] Optionally, the above ultrasound image processing method further includes: if the image quality rating is not greater than a preset quality rating threshold, output an imaging suggestion.

[0164] In the above application scenario, after the ultrasound imaging operator performs imaging, the above ultrasound image processing method can quickly perform a quality rating on the ultrasound image obtained by imaging. If the quality rating of the ultrasound image is not greater than a preset quality rating threshold, a imaging suggestion is output to the ultrasound imaging operator.

[0165] The above imaging suggestion can be obtained based on the evaluation basis output by the decision tree. For example, if the rating of the ultrasound image is unqualified and its evaluation basis is "key structure is missing", then the imaging suggestion for this image is: rescan the key structure of the spleen. In addition, the above ultrasound image processing method can also determine which or which target structures have poor integrity based on the detection confidence of each target structure, so as to output an imaging suggestion to rescan the corresponding target structure.

[0166] Please refer to Figure 5 , based on the same inventive concept, an ultrasound image processing apparatus 300 is further provided in an embodiment of the present application. The apparatus includes:

[0167] An ultrasound image acquisition unit 310, configured to acquire an ultrasound image including a target organ;

[0168] An image rating unit 320, configured to input the ultrasound image into an image rating model, and obtain the image rating of the ultrasound image output by the image rating model;

[0169] Wherein, the image rating model 200 includes: a first rating sub-network 210 and a second rating sub-network 220 respectively used to obtain an organ integrity score and an image quality score; the image rating is determined based on the organ integrity score and the image quality score.

[0170] Figure 6 A schematic diagram of an electronic device provided in an embodiment of the present application. Refer to Figure 6 , the electronic device 400 includes: a processor 410, a memory 420, and a communication interface 430. These components are interconnected and communicate with each other through a communication bus 440 and / or other forms of connection mechanisms (not shown).

[0171] Among them, the memory 420 includes one or more (only one is shown in the figure), which can be, but is not limited to, random access memory (RAM), read only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), etc. The processor 410 and other possible components can access, read, and / or write data in the memory 420.

[0172] The processor 410 includes one or more (only one is shown in the figure), which can be an integrated circuit chip with signal processing capabilities. The above-mentioned processor 410 can be a general-purpose processor, including a central processing unit (CPU), a microcontroller unit (MCU), a network processor (NP), or other conventional processors; it can also be a dedicated processor, including a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0173] The communication interface 430 includes one or more (only one is shown in the figure), which can be used to communicate directly or indirectly with other devices for data interaction. For example, the communication interface 430 can be an Ethernet interface; it can be a mobile communication network interface, such as an interface for 3G, 4G, or 5G networks; it can also be other types of interfaces with data transceiver functions.

[0174] One or more computer program instructions can be stored in the memory 420, and the processor 410 can read and run these computer program instructions to implement the ultrasonic image processing method provided in the embodiments of the present application and other desired functions.

[0175] It can be understood thatFigure 6 The structure shown is only schematic, and the electronic device 400 may also include more or fewer components than those shown Figure 6 in the figure, or have a configuration different from that shown Figure 6 in the figure. Figure 6 Each component shown in the figure may be implemented by hardware, software, or a combination thereof. For example, the electronic device 400 may be a single server (or other device with computing and processing capabilities), a combination of multiple servers, a cluster of a large number of servers, etc., and may be either a physical device or a virtual device.

[0176] An embodiment of the present application also provides a computer-readable storage medium, on which computer program instructions are stored. When the computer program instructions are read and run by a processor of a computer, the ultrasonic image processing method provided by the embodiment of the present application is executed. For example, the computer-readable storage medium may be implemented as Figure 6 the memory 420 in the electronic device 400 in the figure.

[0177] An embodiment of the present application also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the above-mentioned ultrasonic image processing method is implemented.

[0178] In the embodiments provided by the present application, it should be understood that the disclosed devices and methods may be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical function division, and there may be other division methods in actual implementation. For another example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the couplings or direct couplings or communication connections shown or discussed with each other may be through some communication interfaces, and the indirect couplings or communication connections of devices or units may be electrical, mechanical, or other forms.

[0179] In addition, 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. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0180] Furthermore, in each embodiment of the present application, the functional modules may be integrated together to form an independent part, or each module may exist alone, or two or more modules may be integrated to form an independent part.

[0181] The above are only the embodiments of the present application and are not intended to limit the protection scope of the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

Claims

1. An ultrasonic image processing method, characterized in that: The method comprises: Acquiring an ultrasound image containing the target organ; Inputting the ultrasound image into an image rating model to obtain an image rating of the ultrasound image output by the image rating model; The image rating model includes: a first scoring subnetwork and a second scoring subnetwork for respectively obtaining an organ integrity score and an image quality score; and the image rating is determined based on the organ integrity score and the image quality score.

2. The ultrasonic image processing method according to claim 1, characterized in that: Obtaining the image quality score by using the second scoring subnetwork includes: Inputting the ultrasound image into a spatial domain scoring module in the second scoring subnetwork to obtain a first score output by the spatial domain scoring module; wherein the first score is used to represent the probability that the ultrasound image is a high-quality image; Inputting the frequency domain image of the ultrasound image into the frequency domain scoring module in the second scoring subnetwork to obtain a second score output by the frequency domain scoring module; wherein the second score is used to represent the probability that the frequency domain image is a high-quality image; The first score and the second score are input into a quality score calculation module in the second scoring subnetwork to obtain the image quality score of the ultrasound image.

3. The ultrasonic image processing method according to claim 1, characterized in that: The second scoring subnetwork includes a quality scoring module, and the quality scoring module includes a first channel and a second channel; Obtaining the image quality score by using the second scoring subnetwork includes: Input the ultrasound image and the frequency domain image of the ultrasound image into the first channel and the second channel respectively, and obtain a first score output by the first channel and a second score output by the second channel respectively; wherein the first score is used to represent the probability that the ultrasound image is a high-quality image; and the second score is used to represent the probability that the frequency domain image is a high-quality image; The first score and the second score are input into a quality score calculation module in the second scoring subnetwork to obtain the image quality score of the ultrasound image.

4. The ultrasonic image processing method according to claim 1, characterized in that: Obtaining the organ integrity score by using the first scoring subnetwork includes: Inputting the ultrasound image into the target detection module in the first scoring subnetwork to obtain the detection confidence of the target organ; The detection confidence is input into an integrity score calculation module in the first scoring subnetwork to obtain the organ integrity score.

5. The ultrasonic image processing method according to claim 1, characterized in that: Obtaining the organ integrity score by using the first scoring subnetwork includes: Inputting the ultrasound image into a target detection module in the first scoring subnetwork to obtain a detection confidence of at least one target structure in the target organ; The detection confidence of at least one of the target structures is input into an integrity score calculation module in the first scoring subnetwork to obtain the organ integrity score.

6. The ultrasonic image processing method according to claim 1, characterized in that: The image rating model further includes: a third rating subnetwork for obtaining an organ standardization score; the image rating is determined based on the organ integrity score, the image quality score, and the organ standardization score; Using the third scoring subnetwork to obtain the organ standardization score includes: Inputting the ultrasound image and the position information of the target organ into the cropping module in the third scoring sub-network, and obtaining the target organ image cropped by the cropping module from the ultrasound image; wherein the position information of the target organ is obtained by the target detection module in the first scoring sub-network; The target organ image is input into the standardization score calculation module in the third scoring subnetwork, and the organ standardization score calculated by the standardization score calculation module based on the similarity between the target organ image and the standard organ image in the knowledge base is obtained.

7. The ultrasonic image processing method according to claim 1, characterized in that: The image rating model further includes: a third rating subnetwork for obtaining an organ standardization score; the image rating is determined based on the organ integrity score, the image quality score, and the organ standardization score; Using the third scoring subnetwork to obtain the organ standardization score includes: Inputting the ultrasound image and the position information of at least one target structure into the cropping module in the third scoring subnetwork, and obtaining the target structure image cropped from the ultrasound image by the cropping module; wherein the position information of at least one target structure is obtained by the target detection module in the first scoring subnetwork; The target structure image is input into the standardization score calculation module in the third scoring subnetwork, and the organ standardization score calculated by the standardization score calculation module based on the similarity between the target structure image and the standard structure image in the knowledge base is obtained.

8. The ultrasonic image processing method according to claim 5 or 7, characterized in that: The target organ includes the spleen; the target structure includes: the splenic dorsal membrane, the splenic hilum and the splenic vein.

9. The ultrasonic image processing method according to any one of claims 5 to 7, characterized in that: The image rating model further includes: an image rating module; The image rating module is used to obtain the image rating of the ultrasound image using a decision tree based on the organ integrity score, the image quality score and the organ standardization score.

10. The ultrasonic image processing method according to any one of claims 1 to 7, characterized in that: The method further comprises: If the image quality rating is greater than a preset quality rating threshold, outputting a recommendation to retain the image; And / or, if the image quality rating is not greater than the preset quality rating threshold, outputting a drawing suggestion.

11. An ultrasonic image processing device, characterized in that: The device comprises: An ultrasonic image acquisition unit, used for acquiring an ultrasonic image including a target organ; an image rating unit, configured to input the ultrasound image into an image rating model, and obtain an image rating of the ultrasound image output by the image rating model; The image rating model includes: a first scoring subnetwork and a second scoring subnetwork for respectively obtaining an organ integrity score and an image quality score; and the image rating is determined based on the organ integrity score and the image quality score.

12. An electronic device, characterized in that: include: A processor, a memory and a communication bus, wherein the processor and the memory communicate with each other via the communication bus; The memory stores program instructions executable by the processor, and the processor can execute the method according to any one of claims 1 to 10 by calling the program instructions.

13. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a computer, the computer executes the method according to any one of claims 1 to 10.

14. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 10 is implemented.

Images