A cross-section quality grading method and device, an ultrasonic equipment and a storage medium

Abstract

The application discloses a section quality grading method and device, an ultrasonic equipment and a computer readable storage medium. The method comprises the following steps: constructing a classification tree according to different quality grades and different section categories of an ultrasonic image preset; obtaining a target ultrasonic image, and performing classification prediction layer by layer based on the classification tree by using a trained neural network model to obtain a target quality grade and a target section category of the target ultrasonic image; and if the target quality grade is greater than or equal to a preset section quality grade of the ultrasonic equipment, the target ultrasonic image is determined as a standard section image corresponding to the target section category. The section quality grading method provided by the application improves the accuracy of section classification and quality grade classification, and meets the demand of different level hospitals for standard section image fine quality control.

Description

Technical Field

[0001] This application relates to the field of ultrasonic image processing technology, and more specifically, to a method and apparatus for grading cross-sectional quality, an ultrasonic device, and a computer-readable storage medium. Background Technology

[0002] In related technologies, target ultrasound images are directly classified and their quality grade scores are calculated. Doctors of different qualifications or hospitals of different levels have different quality standards for standard section images. For example, 70-80 points is considered a basic passing standard section image, 80-90 points is a passing section, and 90-100 points is a standard section image. Higher-level hospitals currently require standard section image quality control to reach 90 points or above, while lower-level hospitals only require 70 points or above. In existing technologies, the quality grade score threshold for standard section images is fixed. In this approach, if hospitals of different levels all use 70 points as the quality grade score threshold, doctors in higher-level hospitals will obtain section images with scores below 90 points when using the automatic standard section image acquisition function, thus failing to meet the need for refined quality control of standard section images. If the quality grade score threshold is set to 90 points, doctors in lower-level hospitals, due to their limited qualifications and skills, will find it difficult to find section images that meet the requirements when using the automatic standard section image acquisition function.

[0003] It is evident that existing technologies do not take into account the differences in the quality standards of standard cross-sectional images among doctors of different qualifications or hospitals of different levels, and cannot adequately meet the needs of hospitals at different levels for refined quality control of standard cross-sectional images. Summary of the Invention

[0004] The purpose of this application is to provide a method, apparatus, ultrasound device, and computer-readable storage medium for grading the quality of sections, which improves the accuracy of section classification and calculation of quality grade scores, while meeting the needs of hospitals at different levels for refined quality control of standard section images.

[0005] To achieve the above objectives, this application provides a method for grading cross-sectional quality, comprising:

[0006] A classification tree is constructed based on different quality levels and different section categories of ultrasound images; wherein, the last layer of the classification tree represents different quality levels, and the other layers of nodes represent different section categories, and section categories obtained by the same classification standard are located at the same layer;

[0007] Acquire a target ultrasound image, and use a trained neural network model to perform classification prediction layer by layer based on the classification tree to obtain the target quality level and target section category of the target ultrasound image;

[0008] If the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment, then the target ultrasound image is determined as the standard section image corresponding to the target section category.

[0009] This also includes:

[0010] Obtain a training set; wherein the training set includes multiple training ultrasound images, corresponding section categories, and quality levels;

[0011] Based on the classification tree, an initial neural network model is trained using the training set to obtain a trained neural network model.

[0012] The acquisition of the target ultrasound image includes:

[0013] The data stream of the ultrasound imaging system is acquired, and each frame of the data stream is used sequentially as the target ultrasound image to predict the target section category and target quality level.

[0014] Wherein, if the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment, the target ultrasound image is determined as the standard section image corresponding to the target section category, including:

[0015] If the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment, then the target ultrasound image is determined as the candidate standard section image corresponding to the target section category;

[0016] The candidate standard cross-section image with the highest quality level under the target cross-section category is determined as the standard cross-section image corresponding to the target cross-section category.

[0017] The process of using a trained neural network model to perform layer-by-layer classification prediction based on the classification tree to obtain the target quality level and target section category of the target ultrasound image includes:

[0018] The target ultrasound image is input into the trained neural network model, and the probability of each section category in the first layer under the root node of the classification tree is predicted. The section category with the highest predicted probability in the first layer is taken as the section category predicted in the first layer.

[0019] The trained neural network model is used to classify and predict each section category under the section category predicted by the previous layer. The section category with the highest predicted probability is used as the section category predicted by the next layer, until the section category with the highest predicted probability in the penultimate layer is predicted. The section category with the highest predicted probability in the penultimate layer is determined as the target section category of the target ultrasound image.

[0020] The trained neural network model is used to classify and predict the quality levels of each section category with the highest probability in the penultimate layer, thereby obtaining the target quality level of the target ultrasound image.

[0021] Wherein, if the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment, the target ultrasound image is determined as the standard section image corresponding to the target section category, including:

[0022] If the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment, then the target ultrasound image is determined as the candidate standard section image corresponding to the target section category;

[0023] The quality grade score of the target ultrasound image is calculated based on the probability corresponding to the section category predicted by each layer and the probability corresponding to the predicted target quality grade.

[0024] The candidate standard section image with the highest quality grade score under the target section category is determined as the standard section image corresponding to the target section category.

[0025] The step of calculating the quality grade score of the target ultrasound image based on the probability corresponding to the predicted section category and the probability corresponding to the predicted target quality grade includes:

[0026] The quality grade score of the target ultrasound image is obtained by weighting the probability corresponding to the predicted section category and the probability corresponding to the predicted target quality grade.

[0027] The step of obtaining the quality grade score of the target ultrasound image by weighting the probability corresponding to the predicted section category and the probability corresponding to the predicted target quality grade from each layer includes:

[0028] Determine the minimum quality grade score and score range corresponding to the predicted target quality grade;

[0029] The product of the probability corresponding to the predicted section category and the probability corresponding to the predicted target quality level at each layer is determined as the weight.

[0030] The product of the weight and the score span, plus the sum of the lowest quality grade score, is determined as the quality grade score of the target ultrasound image.

[0031] To achieve the above objectives, this application provides a cross-section quality grading device, comprising:

[0032] The construction module is used to construct a classification tree based on different quality levels and different section categories of ultrasound images; wherein, the last layer nodes in the classification tree represent different quality levels, and the other layer nodes represent different section categories, and section categories obtained by the same classification standard are located in the same layer;

[0033] The prediction module is used to acquire the target ultrasound image and use the trained neural network model to perform classification prediction layer by layer based on the classification tree to obtain the target quality level and target section category of the target ultrasound image.

[0034] The determination module is used to determine the target ultrasound image as the standard section image corresponding to the target section category when the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment.

[0035] To achieve the above objectives, this application provides an ultrasonic device, comprising:

[0036] Memory, used to store computer programs;

[0037] A processor is used to implement the steps of the cross-section quality grading method described above when executing the computer program.

[0038] To achieve the above objectives, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the section quality grading method described above.

[0039] As can be seen from the above scheme, the section quality grading method provided in this application includes: constructing a classification tree based on different preset quality levels and different section categories of ultrasound images; wherein, the last layer node in the classification tree represents different quality levels, and other layer nodes represent different section categories, and section categories obtained by the same classification standard are located in the same layer; acquiring a target ultrasound image, and using a trained neural network model to perform classification prediction layer by layer based on the classification tree to obtain the target quality level and target section category of the target ultrasound image; if the target quality level is greater than or equal to the preset section quality level of the ultrasound device, then the target ultrasound image is determined as the standard section image corresponding to the target section category.

[0040] The section quality grading method provided in this application presets different quality levels according to actual needs, constructs a classification tree based on different quality levels and different section categories, and uses a neural network model to classify the target ultrasound image layer by layer based on this classification tree to obtain the target section category and target quality level of the target ultrasound image. Therefore, the section quality grading method provided in this application, by performing layer-by-layer classification through a classification tree, improves the accuracy of section classification and quality level classification compared to existing methods that directly classify target ultrasound images. Furthermore, in this application, the preset section quality level of the ultrasound equipment can be adjusted. When the target quality level of the target ultrasound image is greater than or equal to the preset section quality level of the ultrasound equipment, it is determined to be the standard section image corresponding to the target section category, avoiding the problem of poor flexibility with fixed quality score thresholds and meeting the needs of hospitals at different levels for refined quality control of standard section images. This application also discloses a section quality grading device, an ultrasound device, and a computer-readable storage medium, which can achieve the same technical effects.

[0041] It should be understood that the above general description and the following detailed description are merely exemplary and do not limit this application. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. The drawings are used to provide a further understanding of this disclosure and constitute a part of the specification. They are used together with the following detailed description to explain this disclosure, but do not constitute a limitation of this disclosure. In the drawings:

[0043] Figure 1 This is a flowchart illustrating a section quality grading method according to an exemplary embodiment;

[0044] Figure 2 A flowchart illustrating another method for grading cross-section quality according to an exemplary embodiment;

[0045] Figure 3 This is a structural diagram of a classification tree according to an exemplary embodiment;

[0046] Figure 4 This is a structural diagram illustrating a cross-sectional quality grading device according to an exemplary embodiment;

[0047] Figure 5 This is a structural diagram of an ultrasonic device according to an exemplary embodiment. Detailed Implementation

[0048] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Furthermore, in the embodiments of this application, "first," "second," etc., are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0049] This application discloses a method for grading the quality of cross-sections, which improves the accuracy of cross-section classification and calculation of quality grade scores, while meeting the needs of hospitals at different levels for refined quality control of standard cross-section images.

[0050] See Figure 1 A flowchart illustrating a section quality grading method according to an exemplary embodiment is shown below. Figure 1 As shown, it includes:

[0051] S101: Construct a classification tree based on different quality levels and different section categories of ultrasound images; wherein, the last layer nodes in the classification tree represent different quality levels, and the other layer nodes represent different section categories, and section categories obtained by the same classification standard are located in the same layer;

[0052] In this embodiment, different quality levels are preset according to actual needs. For example, quality levels may include below 60 points, 60-70 points, 70-80 points, 80-90 points, and above 90 points. A classification tree is constructed based on different quality levels and different facet categories. Facet categories obtained by the same classification standard are located at the same level. For example, the facet part is used as the parent category, and then facet categories are specifically divided as subcategories for the facet part, and then specific quality levels are divided for the facet categories. An example of a classification tree is shown below. Figure 2 As shown, the sections include the head section, the thoracic cavity section, the abdominal section, the four muscles and other sections. The sections under the head section include the thalamic horizontal section, etc.

[0053] S102: Acquire the target ultrasound image, and use the trained neural network model to perform classification prediction layer by layer based on the classification tree to obtain the target quality level and target section category of the target ultrasound image;

[0054] The purpose of this step is to determine the target quality level and target section category corresponding to the input target ultrasound image. As a possible implementation, acquiring the target ultrasound image includes: acquiring the data stream of the ultrasound imaging system, and sequentially using each frame of the data stream as the target ultrasound image to predict the target section category and target quality level. It is understood that the input in this embodiment can be the data stream of the ultrasound imaging system, i.e., the data stream of ultrasound images, and prediction is performed sequentially for each ultrasound image according to the input order.

[0055] In practice, the trained neural network model is used to classify and predict the input target ultrasound image layer by layer based on the classification tree, so as to obtain the target quality level and target section category of the target ultrasound image. The target quality level is the prediction result of the last layer in the classification tree, and the target section category is the prediction result of the second to last layer in the classification tree.

[0056] It should be noted that the training process of the above neural network model is as follows: A training set is obtained; the training set includes multiple training ultrasound images, corresponding section categories, and quality levels; an initial neural network model is trained using the training set based on the classification tree to obtain the trained neural network model. In specific implementation, the initial neural network is trained using the training ultrasound images and corresponding annotation information. Here, the neural network can be a common classification network, and the annotation information includes the section category and quality level of the training ultrasound image. For example, the annotation information for a training ultrasound image might be a head section, a thalamic horizontal section, and a score of 90 or higher. When annotating the quality level of a training ultrasound image, it can be based on the clarity and completeness of the structures displayed on the image. For example, for a training ultrasound image of a thalamic horizontal section, if five structures are clearly displayed, its quality level is labeled as "90 or higher"; if one structure is not clearly displayed (i.e., four structural features are clearly displayed), its quality level is labeled as "80-90," and so on.

[0057] S103: If the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment, then the target ultrasound image is determined as the standard section image corresponding to the target section category.

[0058] In this embodiment, the preset slice quality level of the ultrasound equipment can be adjusted to meet the needs of hospitals at different levels for refined quality control of standard slice images. When the target quality level of the target ultrasound image is greater than or equal to the preset slice quality level of the ultrasound equipment, it is determined to be the standard slice image corresponding to the target slice category.

[0059] In a preferred embodiment, this step may include: if the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment, then the target ultrasound image is determined as a candidate standard section image corresponding to the target section category; the candidate standard section image with the highest quality level under the target section category is determined as the standard section image corresponding to the target section category. In a specific implementation, all ultrasound images with a quality level greater than or equal to the preset section quality level under the target section category are used as candidate standard section images corresponding to that target section category. Among all candidate standard section images, the one with the highest quality level is selected as the final standard section image corresponding to the target section category and output. For example, a hospital needs an image quality level of 60-70 points, meaning the preset slice quality level is 60-70 points. Two ultrasound images obtained from consecutive scans are identified by a neural network model as belonging to the same slice category. The first ultrasound image has a quality level of 60-70 points, and the second ultrasound image has a quality level of 70-80 points, both greater than or equal to the preset slice quality level. Both are considered candidate standard slices. The second ultrasound image, with a quality level of 70-80 points, is chosen as the final standard slice image. This demonstrates that when multiple ultrasound images of the same target slice category are scanned, once the preset quality level is reached, the higher-quality ultrasound image will replace the lower-quality one.

[0060] The section quality grading method provided in this application presets different quality levels according to actual needs. A classification tree is constructed based on these different quality levels and section categories. A neural network model then performs layer-by-layer classification of the target ultrasound image based on this classification tree to obtain the target section category and target quality level of the target ultrasound image. Therefore, the section quality grading method provided in this application, by performing layer-by-layer classification through a classification tree, improves the accuracy of section classification and quality level classification compared to existing methods that directly classify target ultrasound images. Furthermore, in this application embodiment, the preset section quality level of the ultrasound equipment can be adjusted. When the target quality level of the target ultrasound image is greater than or equal to the preset section quality level of the ultrasound equipment, it is determined to be a standard section image corresponding to the target section category. This avoids the problem of poor flexibility associated with fixed quality score thresholds and meets the needs of hospitals at different levels for refined quality control of standard section images.

[0061] This application discloses a method for grading cross-sectional quality. Compared with the previous embodiment, this embodiment further explains and optimizes the technical solution. Specifically:

[0062] See Figure 2 A flowchart illustrating another section quality grading method according to an exemplary embodiment, such as... Figure 2 As shown, it includes:

[0063] S201: Construct a classification tree based on different quality levels and different section categories of ultrasound images; wherein, the last layer nodes in the classification tree represent different quality levels, and the other layer nodes represent different section categories, and section categories obtained by the same classification standard are located in the same layer;

[0064] S202: Acquire the data stream of the ultrasound imaging system, and use each frame of the data stream as a target ultrasound image to predict the target section category and target quality level;

[0065] S203: Input the target ultrasound image into the trained neural network model, predict the probability of each section category in the first layer under the root node of the classification tree, and take the section category with the highest predicted probability in the first layer as the section category predicted in the first layer.

[0066] S204: Using the trained neural network model, classify and predict each section category under the section category predicted by the previous layer, and take the section category with the highest predicted probability as the section category predicted by the next layer, until the section category with the highest probability in the second to last layer is predicted, and determine the section category with the highest probability in the second to last layer as the target section category of the target ultrasound image.

[0067] S205: Using the trained neural network model, classify and predict the quality levels of each section category with the highest probability in the penultimate layer to obtain the target quality level of the target ultrasound image;

[0068] In practical implementation, the target ultrasound image is input into a trained neural network model. The neural network model first predicts the probability of each slice category in the first layer under the root node of the classification tree, and then uses the slice category with the highest probability as the slice category predicted by the first layer. Figure 3 Taking the classification tree shown as an example, the probability corresponding to each section is first predicted. For instance, the probability of the head section is 90%, the probability of the chest section is 5%, the probability of the abdomen section is 3%, and the probability of the four muscles and other sections is 2%. The section with the highest probability is the head section, indicating that the section category predicted in the first layer is the head section. Further, the categories of each section under the predicted section category from the previous layer are classified and predicted, and the section category with the highest predicted probability is used as the section category predicted in the next layer. Similarly, using... Figure 3Taking the classification tree shown as an example, the probability of each section category in the second layer below the head section is predicted. For example, the probability of the thalamic horizontal cross section is 80%, which is greater than the probability of other section categories. This indicates that the section category predicted in the second layer is the thalamic horizontal cross section. The classification tree is traversed layer by layer in the same manner. The section category with the highest probability in the second-to-last layer is the target section category corresponding to the target ultrasound image. The quality grade with the highest probability under the section category with the highest probability in the second-to-last layer, i.e., the quality grade with the highest probability in the last layer, is the quality grade of the target ultrasound image.

[0069] S206: If the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment, then the target ultrasound image is determined as the candidate standard section image corresponding to the target section category;

[0070] S207: Calculate the quality grade score of the target ultrasound image based on the probability corresponding to the predicted section category and the probability corresponding to the predicted target quality grade of each layer;

[0071] In this step, the quality grade score of the target ultrasound image is obtained by weighting the probabilities corresponding to the predicted section categories and the predicted target quality grades at each layer. As a possible implementation, a minimum quality grade score and a score range corresponding to the predicted target quality grade are determined; the product of the probabilities corresponding to the predicted section categories and the predicted target quality grades at each layer is determined as the weight; the sum of the product of the weight and the score range, and the minimum quality grade score, is determined as the quality grade score of the target ultrasound image. Figure 3 Taking the classification tree shown as an example, if the first layer predicts the section category as head section, the second layer predicts the section category as thalamic horizontal cross section, and the third layer predicts a quality grade of 90 or above, the probability corresponding to the head section is 90%, the probability corresponding to the thalamic horizontal cross section is 80%, and the probability corresponding to 90 or above is 70%, then the lowest quality grade score corresponding to the predicted quality grade is 90, the score range is 10, the target section category of the target ultrasound image is thalamic horizontal cross section, and the quality grade score is 90 + 10 × 90% × 80% × 70% = 95.04.

[0072] S208: The candidate standard section image with the highest quality grade score under the target section category is determined as the standard section image corresponding to the target section category.

[0073] In this embodiment, for the ultrasound image data stream, the slice category and quality score of the currently input target ultrasound image are continuously calculated. The system continuously searches for the ultrasound image with the highest quality score within each target slice category, where the quality score is greater than or equal to a preset quality level. This image is then identified as the standard slice image corresponding to the target slice category and displayed in real-time on the ultrasound interface. For example, if a hospital requires an image quality score of 60-70 (i.e., the preset slice quality score is 60-70), and two ultrasound images scanned consecutively are identified by the neural network model as belonging to the same slice category, both with a quality score of 60-70, both greater than or equal to the preset slice quality level, making them both candidate standard slices. If the first ultrasound image has a quality score of 65 and the second ultrasound image has a quality score of 68, then the second ultrasound image is selected as the final standard slice image. Therefore, when multiple ultrasound images within the same target slice category are scanned, after reaching the preset quality level, the ultrasound image with the higher quality score will replace the one with the lower quality score.

[0074] Therefore, the section quality grading method provided in this application, which performs layer-by-layer classification and probability calculation using a classification tree, improves the accuracy of section classification and quality grade score calculation compared to existing methods that directly classify target ultrasound images. Furthermore, in this application, the ultrasound image with the highest quality grade score under the target section category is output, avoiding the inflexibility problem of fixed quality score thresholds and meeting the needs of hospitals at different levels for refined quality control of standard section images.

[0075] The following describes a cross-section quality grading device provided in an embodiment of this application. The cross-section quality grading device described below and the cross-section quality grading method described above can be referred to each other.

[0076] See Figure 4 A structural diagram of a cross-sectional quality grading device is shown according to an exemplary embodiment, as follows: Figure 4 As shown, it includes:

[0077] The construction module 401 is used to construct a classification tree based on different quality levels and different section categories of ultrasound images; wherein, the last layer nodes in the classification tree represent different quality levels, and the other layer nodes represent different section categories, and the section categories obtained by the same classification standard are located in the same layer;

[0078] The prediction module 402 is used to acquire the target ultrasound image and use the trained neural network model to perform classification prediction layer by layer based on the classification tree to obtain the target quality level and target section category of the target ultrasound image.

[0079] The determination module 403 is used to determine the target ultrasound image as the standard section image corresponding to the target section category when the target quality level is greater than or equal to the preset section quality level of the ultrasound device.

[0080] The section quality grading device provided in this application presets different quality levels according to actual needs. A classification tree is constructed based on these different quality levels and section categories. A neural network model then performs layer-by-layer classification of the target ultrasound image based on this classification tree to obtain the target section category and target quality level of the target ultrasound image. Therefore, the section quality grading method provided in this application, by performing layer-by-layer classification through a classification tree, improves the accuracy of section classification and quality level classification compared to existing methods that directly classify target ultrasound images. Furthermore, in this application embodiment, the preset section quality level of the ultrasound equipment can be adjusted. When the target quality level of the target ultrasound image is greater than or equal to the preset section quality level of the ultrasound equipment, it is determined to be a standard section image corresponding to the target section category. This avoids the problem of poor flexibility associated with fixed quality score thresholds and meets the needs of hospitals at different levels for refined quality control of standard section images.

[0081] Based on the above embodiments, as a preferred embodiment, it further includes:

[0082] An acquisition module is used to acquire a training set; wherein, the training set includes multiple training ultrasound images, corresponding section categories, and quality levels;

[0083] The training module is used to train an initial neural network model based on the classification tree and the training set to obtain a trained neural network model.

[0084] Based on the above embodiments, as a preferred embodiment, the prediction module 402 includes:

[0085] The acquisition unit is used to acquire the data stream of the ultrasound imaging system and sequentially use each frame of the data stream as a target ultrasound image to predict the target section category and target quality level.

[0086] The prediction unit is used to perform classification prediction layer by layer based on the classification tree using the trained neural network model to obtain the target quality level and target section category of the target ultrasound image.

[0087] Based on the above embodiments, as a preferred embodiment, the determining module 403 includes:

[0088] The first determining unit is used to determine the target ultrasound image as a candidate standard section image corresponding to the target section category when the target quality level is greater than or equal to the preset section quality level of the ultrasound device.

[0089] The second determining unit is used to determine the candidate standard section image with the highest quality level under the target section category as the standard section image corresponding to the target section category.

[0090] Based on the above embodiments, as a preferred embodiment, the prediction unit includes:

[0091] The first prediction subunit is used to input the target ultrasound image into the trained neural network model, predict the probability of each section category in the first layer under the root node of the classification tree, and take the section category with the highest predicted probability in the first layer as the section category predicted in the first layer.

[0092] The second prediction subunit is used to classify and predict each section category under the section category predicted by the previous layer using the trained neural network model, and take the section category with the highest probability as the section category predicted by the next layer, until the section category with the highest probability in the second to last layer is predicted, and the section category with the highest probability in the second to last layer is determined as the target section category of the target ultrasound image.

[0093] The third prediction subunit is used to classify and predict the quality level of the target ultrasound image by using the trained neural network model to classify each quality level under the section category with the highest probability in the penultimate layer.

[0094] Based on the above embodiments, as a preferred embodiment, the determining module 403 includes:

[0095] The first determining unit is used to determine the target ultrasound image as a candidate standard section image corresponding to the target section category when the target quality level is greater than or equal to the preset section quality level of the ultrasound device.

[0096] The calculation unit is used to calculate the quality grade score of the target ultrasound image based on the probability corresponding to the predicted section category and the probability corresponding to the predicted target quality grade of each layer.

[0097] The third determining unit is used to determine the candidate standard section image with the highest quality grade score under the target section category as the standard section image corresponding to the target section category.

[0098] Based on the above embodiments, as a preferred implementation, the calculation unit is specifically a module that performs a weighted calculation based on the probability corresponding to the predicted section category and the probability corresponding to the predicted target quality level to obtain the quality level score of the target ultrasound image.

[0099] Based on the above embodiments, as a preferred embodiment, the computing unit includes:

[0100] The first determining subunit is used to determine the minimum quality grade score and score range corresponding to the predicted target quality grade;

[0101] The second determining subunit is used to determine the weight by multiplying the probability corresponding to the predicted section category and the probability corresponding to the predicted quality level of each layer.

[0102] A calculation subunit is used to determine the quality grade score of the target ultrasound image by summing the product of the weight and the score span with the lowest quality grade score.

[0103] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0104] Based on the hardware implementation of the above program modules, and in order to implement the method of the embodiments of this application, the embodiments of this application also provide an ultrasonic device. Figure 5 This is a structural diagram illustrating an ultrasonic device according to an exemplary embodiment, such as... Figure 5 As shown, the ultrasound equipment includes:

[0105] Communication interface 1 enables information exchange with other devices, such as network devices;

[0106] Processor 2 is connected to communication interface 1 to enable information exchange with other devices and, when running a computer program, executes the cross-sectional quality grading method provided by one or more of the above-mentioned technical solutions. The computer program is stored in memory 3.

[0107] Of course, in practical applications, the various components of the ultrasonic equipment are coupled together through bus system 4. It can be understood that bus system 4 is used to achieve communication and connection between these components. In addition to the data bus, bus system 4 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in... Figure 5 The general will label all buses as Bus System 4.

[0108] The memory 3 in this embodiment is used to store various types of data to support the operation of the ultrasound device. Examples of such data include any computer program used to operate the ultrasound device.

[0109] It is understood that memory 3 can be volatile memory or non-volatile memory, or both. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), ferromagnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM); magnetic surface memory can be disk storage or magnetic tape storage. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Synchronous Static Random Access Memory (SSRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), SyncLink Dynamic Random Access Memory (SLDRAM), and Direct Rambus Random Access Memory (DRRAM).The memory 2 described in the embodiments of this application is intended to include, but is not limited to, these and any other suitable types of memory.

[0110] The methods disclosed in the embodiments of this application can be applied to processor 2, or implemented by processor 2. Processor 2 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in processor 2 or by instructions in the form of software. The processor 2 may be a general-purpose processor, DSP, or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Processor 2 can implement or execute the methods, steps and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor may be a microprocessor or any conventional processor, etc. The steps of the methods disclosed in the embodiments of this application can be directly manifested as being executed by a hardware decoding processor, or being executed by a combination of hardware and software modules in the decoding processor. The software modules may be located in a storage medium, which is located in memory 3. Processor 2 reads the program in memory 3 and completes the steps of the aforementioned method in combination with its hardware.

[0111] When processor 2 executes the program, it implements the corresponding processes in the various methods of the embodiments of this application. For the sake of brevity, these will not be described in detail here.

[0112] In an exemplary embodiment, this application also provides a storage medium, namely a computer storage medium, specifically a computer-readable storage medium, such as a memory 3 that stores a computer program, which can be executed by a processor 2 to complete the steps described in the aforementioned method. The computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface memory, optical disc, or CD-ROM.

[0113] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0114] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an ultrasound device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0115] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for grading the quality of a cross-section, characterized in that, include: A classification tree is constructed based on different quality levels and different section categories of ultrasound images; wherein, the last layer of the classification tree represents different quality levels, and the other layers of nodes represent different section categories, and section categories obtained by the same classification standard are located at the same layer; A target ultrasound image is acquired, and a trained neural network model is used to perform classification prediction layer by layer based on the classification tree to obtain the target quality level and target section category of the target ultrasound image; wherein, the target quality level is the prediction result of the last layer in the classification tree, and the target section category is the prediction result of the second to last layer in the classification tree. If the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment, then the target ultrasound image is determined as the standard section image corresponding to the target section category; wherein, the preset section quality level is set according to the refined quality control requirements of standard section images for hospitals of different levels.

2. The method for grading the cross-sectional quality according to claim 1, characterized in that, Also includes: Obtain a training set; wherein the training set includes multiple training ultrasound images, corresponding section categories, and quality levels; Based on the classification tree, an initial neural network model is trained using the training set to obtain a trained neural network model.

3. The method for grading the cross-sectional quality according to claim 1, characterized in that, The acquisition of the target ultrasound image includes: The data stream of the ultrasound imaging system is acquired, and each frame of the data stream is used sequentially as the target ultrasound image to predict the target section category and target quality level.

4. The method for grading the cross-sectional quality according to claim 1, characterized in that, If the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment, then the target ultrasound image is determined as the standard section image corresponding to the target section category, including: If the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment, then the target ultrasound image is determined as the candidate standard section image corresponding to the target section category; The candidate standard cross-section image with the highest quality level under the target cross-section category is determined as the standard cross-section image corresponding to the target cross-section category.

5. The method for grading the cross-sectional quality according to any one of claims 1 to 4, characterized in that, The trained neural network model performs classification prediction layer by layer based on the classification tree to obtain the target quality level and target section category of the target ultrasound image, including: The target ultrasound image is input into the trained neural network model, and the probability of each section category in the first layer under the root node of the classification tree is predicted. The section category with the highest predicted probability in the first layer is taken as the section category predicted in the first layer. The trained neural network model is used to classify and predict each section category under the section category predicted by the previous layer. The section category with the highest predicted probability is used as the section category predicted by the next layer, until the section category with the highest predicted probability in the penultimate layer is predicted. The section category with the highest predicted probability in the penultimate layer is determined as the target section category of the target ultrasound image. The trained neural network model is used to classify and predict the quality levels of each section category with the highest probability in the penultimate layer, thereby obtaining the target quality level of the target ultrasound image.

6. The method for grading the cross-sectional quality according to claim 5, characterized in that, If the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment, then the target ultrasound image is determined as the standard section image corresponding to the target section category, including: If the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment, then the target ultrasound image is determined as the candidate standard section image corresponding to the target section category; The quality grade score of the target ultrasound image is calculated based on the probability corresponding to the section category predicted by each layer and the probability corresponding to the predicted target quality grade. The candidate standard section image with the highest quality grade score under the target section category is determined as the standard section image corresponding to the target section category.

7. The method for grading the cross-sectional quality according to claim 6, characterized in that, The process of calculating the quality grade score of the target ultrasound image based on the probability corresponding to the section category predicted by each layer and the probability corresponding to the predicted target quality grade includes: The quality grade score of the target ultrasound image is obtained by weighting the probability corresponding to the predicted section category and the probability corresponding to the predicted target quality grade.

8. The method for grading the cross-sectional quality according to claim 7, characterized in that, The step of obtaining the quality grade score of the target ultrasound image by weighting the probability corresponding to the predicted section category and the probability corresponding to the predicted target quality grade of each layer includes: Determine the minimum quality grade score and score range corresponding to the predicted target quality grade; The product of the probability corresponding to the predicted section category and the probability corresponding to the predicted target quality level at each layer is determined as the weight. The product of the weight and the score span, plus the sum of the lowest quality grade score, is determined as the quality grade score of the target ultrasound image.

9. A cross-sectional quality grading device, characterized in that, include: The construction module is used to construct a classification tree based on different quality levels and different section categories of ultrasound images; wherein, the last layer nodes in the classification tree represent different quality levels, and the other layer nodes represent different section categories, and section categories obtained by the same classification standard are located in the same layer; The prediction module is used to acquire the target ultrasound image and use the trained neural network model to perform classification prediction layer by layer based on the classification tree to obtain the target quality level and target section category of the target ultrasound image; wherein, the target quality level is the prediction result of the last layer in the classification tree, and the target section category is the prediction result of the second to last layer in the classification tree. The determination module is used to determine the target ultrasound image as a standard section image corresponding to the target section category when the target quality level is greater than or equal to the preset section quality level of the ultrasound equipment; wherein, the preset section quality level is set according to the refined quality control requirements of standard section images for different levels of hospitals.

10. An ultrasonic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the steps of the section quality grading method as described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the section quality grading method as described in any one of claims 1 to 8.

Images