An ultrasonic image processing method and device, electronic equipment and storage medium

Abstract

This application discloses an ultrasound image processing method, apparatus, electronic device, and storage medium. The method includes: acquiring multiple types of data signals during ultrasound imaging; wherein the multiple types of data signals include at least two of the following: ultrasound echo signal after beamforming, initial wave image before digital scan conversion, and ultrasound image after digital scan conversion; performing no-load detection on each of the multiple types of data signals to obtain corresponding no-load detection results; and combining multiple no-load detection results to determine the final no-load detection result of the ultrasound image. This application performs multiple no-load detections on multiple types of data signals in multiple stages of ultrasound imaging, and combines the results of multiple no-load detections to obtain the final no-load detection result of the ultrasound image, improving the accuracy of no-load image identification, avoiding subsequent functions from making incorrect diagnoses or measurements of the no-load image, and ensuring the reliability and accuracy of ultrasound diagnostic and measurement functions.

Description

Technical Field

[0001] This application relates to the field of ultrasound technology, and more specifically, to an ultrasound image processing method, apparatus, electronic device, and storage medium. Background Technology

[0002] With the widespread application of image detection algorithms, such as those based on artificial intelligence, in the field of ultrasound, more and more algorithmic functions are being implemented. However, the problem of algorithm misidentification when the probe is unloaded is common. For example, the liver ultrasound lesion diagnosis function may misidentify diffuse nodules in unloaded images with a certain probability. Similarly, there is a certain probability that the femur will be incorrectly identified and its length calculated in unloaded images.

[0003] Therefore, how to improve the accuracy of identifying empty images is a technical problem that needs to be solved by those skilled in the art. Summary of the Invention

[0004] The purpose of this application is to provide an ultrasound image processing method, apparatus, electronic device, and computer-readable storage medium, which improves the accuracy of identifying empty images.

[0005] To achieve the above objectives, this application provides an ultrasound image processing method, comprising:

[0006] Acquire multiple types of data signals during ultrasound imaging; wherein, the multiple types of data signals include at least two of the following: ultrasound echo signal after beamforming, initial wave image before digital scan conversion, and ultrasound image after digital scan conversion;

[0007] The corresponding no-load detection results are obtained by performing no-load detection on various types of data signals respectively;

[0008] The final empty-load detection result of the ultrasound image is determined by combining multiple empty-load detection results.

[0009] The process of performing no-load detection on the ultrasonic echo signal to obtain the corresponding no-load detection result includes:

[0010] Determine whether the signal strength of the ultrasonic echo signal is greater than or equal to the intensity threshold;

[0011] If so, the no-load detection result corresponding to the ultrasonic echo signal is determined to be non-no-load;

[0012] If not, the no-load detection result corresponding to the ultrasonic echo signal is determined to be no-load.

[0013] The process of performing empty-load detection on the initial wave image to obtain the corresponding empty-load detection result includes:

[0014] Boundary detection is performed on the initial wave image, and the empty load detection result corresponding to the initial wave image is determined based on the boundary detection result.

[0015] The process includes boundary detection of the initial wave image and determining the empty-load detection result corresponding to the initial wave image based on the boundary detection result, including:

[0016] The initial wave image is subjected to boundary detection using the Canny operator, and a first detection region is extracted from the boundary detection result image; wherein, the center point of the first detection region is the center point of the boundary detection result image, the length of the first detection region is a preset proportion of the length of the boundary detection result image, and the width of the first detection region is a preset proportion of the width of the boundary detection result image.

[0017] Calculate the average pixel value of the pixels in the first detection region, and determine whether the average pixel value is greater than or equal to the pixel threshold;

[0018] If so, the empty detection result corresponding to the initial wave image is determined to be empty;

[0019] If not, the empty detection result corresponding to the initial wave image is determined to be non-empty.

[0020] The process of performing empty-load detection on the ultrasound image to obtain the corresponding empty-load detection result includes:

[0021] The ultrasound image is input into the trained empty image classification model to obtain the corresponding empty detection result.

[0022] The process of performing empty-load detection on the ultrasound image to obtain the corresponding empty-load detection result includes:

[0023] An imaging region is extracted from the ultrasound image, and an empty detection is performed on the imaging region to obtain the corresponding empty detection result.

[0024] The process of performing empty-load detection on the imaging region to obtain the corresponding empty-load detection result includes:

[0025] Multiple second detection regions are randomly selected from the imaging region; wherein the multiple second detection regions have the same shape and size, and there is no overlap between any two second detection regions;

[0026] Calculate the similarity between multiple second detection regions. If the similarity between multiple second detection regions is greater than or equal to the similarity threshold, then the empty detection result corresponding to the imaging region is determined to be empty; otherwise, the empty detection result corresponding to the imaging region is determined to be non-empty.

[0027] Specifically, the similarity between multiple second detection regions is calculated. If the similarity between multiple second detection regions is greater than or equal to a similarity threshold, the empty detection result corresponding to the imaging region is determined to be empty; otherwise, the empty detection result corresponding to the imaging region is determined to be non-empty. This includes:

[0028] Calculate the hash values ​​of multiple second detection regions respectively, and calculate the Hamming distance between the multiple hash values;

[0029] If all of the stated Hamming distances are less than or equal to the distance threshold, then the empty detection result corresponding to the imaging region is determined to be empty; otherwise, the empty detection result corresponding to the imaging region is determined to be non-empty.

[0030] The step of determining the final empty-load detection result of the ultrasound image by integrating multiple empty-load detection results includes:

[0031] If any empty detection result is empty, then the final empty detection result of the ultrasound image is determined to be an empty image;

[0032] Otherwise, the final empty detection result of the ultrasound image is determined to be a non-empty image.

[0033] The final empty-load detection result is used as the result of preprocessing the input image for the ultrasound image recognition model; correspondingly, after determining the final empty-load detection result of the ultrasound image by integrating multiple empty-load detection results, the method further includes:

[0034] If the final empty detection result of the ultrasound image is a non-empty image, then the ultrasound image is input into the ultrasound image recognition model to obtain the image recognition result corresponding to the ultrasound image; otherwise, the ultrasound image is not input into the ultrasound image recognition model.

[0035] To achieve the above objectives, this application provides an ultrasound image processing apparatus, comprising:

[0036] The acquisition module is used to acquire multiple types of data signals during ultrasound imaging; wherein, the multiple types of data signals include at least two of the following: ultrasound echo signal after beamforming, initial wave image before digital scan conversion, and ultrasound image after digital scan conversion;

[0037] The no-load detection module is used to perform no-load detection on various types of data signals to obtain the corresponding no-load detection results.

[0038] The determination module is used to determine the final empty-load detection result of the ultrasound image by integrating multiple empty-load detection results.

[0039] To achieve the above objectives, this application provides an electronic device, comprising:

[0040] Memory, used to store computer programs;

[0041] A processor is used to implement the steps of the ultrasound image processing method described above when executing the computer program.

[0042] 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 ultrasound image processing method described above.

[0043] As can be seen from the above scheme, the ultrasound image processing method provided in this application includes: acquiring multiple types of data signals during ultrasound imaging; wherein, the multiple types of data signals include at least two of the following: ultrasound echo signal after beamforming, initial wave image before digital scan conversion, and ultrasound image after digital scan conversion; performing empty detection on the multiple types of data signals respectively to obtain corresponding empty detection results; and combining multiple empty detection results to determine the final empty detection result of the ultrasound image.

[0044] The ultrasonic image processing method provided in this application performs multiple no-load detections on various types of data signals from multiple stages in the ultrasonic imaging process, and synthesizes the results of these multiple no-load detections to obtain the final no-load detection result of the ultrasonic image. This avoids misjudgments that may occur due to a single detection stage, thereby effectively improving the accuracy of no-load image identification and preventing subsequent functions from making incorrect diagnoses or measurements of the no-load image. This effectively reduces the risk of misidentification and ensures the reliability and accuracy of ultrasonic diagnostic and measurement functions. Furthermore, by performing multiple no-load detections on various types of data signals from multiple stages, this application can more comprehensively capture no-load characteristics, enabling the ultrasonic system to more stably determine the no-load state under various conditions, enhancing the system's robustness, and improving the stability and reliability of ultrasonic equipment in practical applications. This application also discloses an ultrasonic image processing device, an electronic device, and a computer-readable storage medium, which can achieve the same technical effects.

[0045] 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

[0046] 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:

[0047] Figure 1 This is a flowchart illustrating an ultrasound image processing method according to an exemplary embodiment;

[0048] Figure 2 This is a schematic diagram showing a comparison of an initial wave image under no-load and non-no-load conditions according to an exemplary embodiment;

[0049] Figure 3 This is a schematic diagram illustrating the comparison of ultrasound images under no-load and non-no-load conditions according to an exemplary embodiment;

[0050] Figure 4 This is a schematic diagram illustrating an unloaded detection method according to an exemplary embodiment;

[0051] Figure 5 This is a flowchart illustrating another ultrasound image processing method according to an exemplary embodiment;

[0052] Figure 6 This is a structural diagram of an ultrasound image processing apparatus according to an exemplary embodiment;

[0053] Figure 7 This is a structural diagram of an electronic device according to an exemplary embodiment. Detailed Implementation

[0054] 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.

[0055] This application discloses an ultrasound image processing method that improves the accuracy of identifying unloaded images.

[0056] See Figure 1A flowchart illustrating an ultrasound image processing method according to an exemplary embodiment is shown below. Figure 1 As shown, it includes:

[0057] S101: Acquire multiple types of data signals during ultrasound imaging; wherein, the multiple types of data signals include at least two of the following: ultrasound echo signal after beamforming, initial wave image before digital scan conversion, and ultrasound image after digital scan conversion;

[0058] The aforementioned data signals can be data signals from ultrasound images at various stages of the ultrasound imaging process. The ultrasound image can be an ultrasound image from various scenarios; specifically, it can be any image to be displayed, or an image to be input into an AI model or other image recognition algorithm for identification.

[0059] For example, the execution subject of this embodiment can be an ultrasound device. The ultrasound device can store data signals of each stage of the ultrasound imaging process in the storage space. When it is determined that a certain ultrasound image needs to be detected without load, the corresponding data signal is obtained from the storage space for detection without load. Alternatively, the ultrasound device can perform corresponding detection without load based on the data signal generated each time a data signal that can be detected without load is generated during the real-time imaging process.

[0060] S102: Perform no-load detection on various types of data signals to obtain the corresponding no-load detection results;

[0061] The purpose of this method is to identify empty images when the ultrasound probe is not detecting a medium or tissue. During ultrasound imaging, the data signals obtained in each stage under empty and non-empty conditions are significantly different. This embodiment utilizes the differences between empty and non-empty images from various data signal types in multiple stages to comprehensively identify empty images.

[0062] The ultrasound imaging process includes: the transmitting module generates an ultrasonic signal, which is transmitted to the ultrasound probe through a T / R switch. The ultrasound probe receives a one-dimensional ultrasonic echo signal, which is then amplified and converted into digital form by an ADC (Analog-to-Digital Converter). The digital signal is processed by beamforming, demodulation, dynamic range compression, etc., to obtain a two-dimensional initial wave image before the DSC (Digital Scan Converter). After the DSC, a two-dimensional ultrasound image is obtained, which is the ultrasound image finally displayed on the screen.

[0063] Therefore, this embodiment can acquire at least two of the following data signals: the ultrasonic echo signal after beamforming, the initial wave image before digital scanning conversion, and the ultrasonic image after digital scanning conversion. Then, at least one empty detection is performed on different types of data signals using different empty detection methods to obtain the corresponding empty detection results.

[0064] As a feasible implementation method, the ultrasonic echo signal is subjected to no-load detection to obtain the corresponding no-load detection result, including: determining whether the signal strength of the ultrasonic echo signal is greater than or equal to the strength threshold; if so, the no-load detection result corresponding to the ultrasonic echo signal is determined to be non-no-load; if not, the no-load detection result corresponding to the ultrasonic echo signal is determined to be no-load.

[0065] In practice, the presence or absence of an ultrasonic echo signal is determined by assessing signal strength. First, an intensity threshold is set, and then the intensity of the ultrasonic echo signal is measured. If the measured signal strength is greater than or equal to the intensity threshold, the corresponding no-load detection result is considered non-no-load, meaning the ultrasonic probe is in contact with the medium or tissue and is in a working state. Conversely, if the signal strength is less than the intensity threshold, it is considered no-load, meaning the ultrasonic probe is not in contact with any medium or tissue and is in an idle state. This method determines the no-load detection result through the intensity of a one-dimensional ultrasonic echo signal. The method is simple and can quickly output the no-load detection result early in the ultrasonic imaging process. If the no-load detection result at this stage is no-load, no other stages of no-load detection can be performed, effectively shortening the no-load detection processing time.

[0066] As a feasible implementation method, performing empty-load detection on the initial wave image to obtain the corresponding empty-load detection result includes: performing boundary detection on the initial wave image, and determining the empty-load detection result corresponding to the initial wave image based on the boundary detection result. That is, if the boundary detection result shows that there are obvious boundary features, the empty-load detection result corresponding to the initial wave image is determined to be empty; if the boundary detection result shows that there are no obvious boundary features, the empty-load detection result corresponding to the initial wave image is determined to be non-empty.

[0067] like Figure 2 As shown, the left side is the initial wave image under no-load conditions, and the right side is the initial wave image under non-no-load conditions. The initial wave images under no-load and non-no-load conditions differ significantly. The initial wave image under no-load conditions has a clear boundary, while the initial wave image under non-no-load conditions does not. Because the initial wave image under no-load conditions has a clear boundary, and the pixel value distribution on both sides of the boundary is different—one side is darker and the other is brighter—boundary detection is performed on the initial wave image. Based on the detected boundary features and pixel value distribution, it is determined whether the image is no-load, thus determining the no-load detection result corresponding to the initial wave image.

[0068] As a feasible implementation method, boundary detection is performed on the initial wave image, and the empty detection result corresponding to the initial wave image is determined based on the boundary detection result. This includes: performing boundary detection on the initial wave image using the Canny operator, and extracting a first detection region from the boundary detection result image; wherein the center point of the first detection region is the center point of the boundary detection result image, the length of the first detection region is a preset proportion of the length of the boundary detection result image, and the width of the first detection region is a preset proportion of the width of the boundary detection result image; calculating the pixel mean of the pixels in the first detection region, and determining whether the pixel mean is greater than or equal to a pixel threshold; if yes, the empty detection result corresponding to the initial wave image is determined to be empty; if no, the empty detection result corresponding to the initial wave image is determined to be non-empty.

[0069] In practice, the Canny operator is first used to perform boundary detection on the initial wave image, resulting in a boundary detection result image. This result is a two-dimensional image where boundary pixels are assigned a value of 1, and non-boundary pixels are assigned a value of 0. Next, a first detection region is extracted from this boundary detection result image. The center point of this region coincides with the center point of the boundary detection result image, and its length and width are preset proportions of the boundary detection result image's length and width, for example, half the length and width of the boundary detection result image. Then, the average pixel value of the pixels within the first detection region is calculated, and it is determined whether this average pixel value is greater than or equal to a set pixel threshold. If yes, the initial wave image is determined to be an empty image; otherwise, it is determined to be a non-empty image. It should be noted that 1 / 2 is only an example of a preset proportion. The preset proportion can also be other values. For example, the preset proportion can be adjusted according to the ultrasound imaging scene, target tissue type, probe type, etc., to achieve more adaptable empty detection.

[0070] In the above implementation, the no-load detection result is determined based on the boundary detection. Since there is a clear difference between no-load and non-no-load at this stage, and boundary detection can accurately identify this difference, this approach can achieve relatively accurate no-load detection.

[0071] As a feasible implementation method, obtaining corresponding empty detection results by performing empty detection on ultrasound images includes: inputting the ultrasound image into a trained empty image classification model to obtain the corresponding empty detection results. The input to the empty image classification model can be a complete ultrasound image or a partial image extracted from the ultrasound image, such as an imaging region image.

[0072] like Figure 3As shown, the left side is the ultrasound image under no-load conditions, and the right side is the ultrasound image under non-no-load conditions. There are also significant differences between the ultrasound images after digital scanning conversion under no-load and non-no-load conditions. The ultrasound image under no-load conditions may be a relatively uniform image, such as pure black, pure gray, or pure white, while the ultrasound image under non-no-load conditions shows normal tissue and is a less uniform image.

[0073] In practice, ultrasound images are input into a pre-trained empty-image classification model. This model, trained using machine learning or deep learning methods, such as a binary classification model trained with the ResNet50 (residual network) classification algorithm, can identify empty ultrasound images and non-empty ultrasound images containing tissue by utilizing the differences between empty and non-empty ultrasound images, and then output the empty detection result corresponding to the ultrasound image. In the above implementation, the pre-trained empty-image classification model enables rapid identification of ultrasound images.

[0074] As another feasible implementation method, empty detection is performed on the ultrasound image to obtain the corresponding empty detection result, including: extracting the imaging area in the ultrasound image, and performing empty detection on the imaging area to obtain the corresponding empty detection result.

[0075] In specific implementation, based on the ultrasound image after digital scanning conversion, image region matting is performed to focus on key areas of the image. This embodiment does not specifically limit the shape of the imaging region; different types of ultrasound equipment have different shapes of imaging regions, such as fan-shaped, rectangular, or circular. Empty region detection is performed on the imaging region to obtain the corresponding empty region detection result. Empty region detection can be performed by performing uniformity detection on the imaging region. If the uniformity detection result is higher than a set condition, the empty region detection result is determined to be empty; otherwise, it is not empty. Since the background area outside the imaging region is often a completely uniform black or white area, this embodiment does not perform empty region detection on the background area, but only on the imaging region where non-uniformity may exist.

[0076] As a feasible implementation method, empty detection is performed on the imaging region to obtain the corresponding empty detection result, including: randomly selecting multiple second detection regions in the imaging region; wherein, the multiple second detection regions have the same shape and size, and there is no overlap between any two second detection regions; calculating the similarity between the multiple second detection regions, if the similarity between the multiple second detection regions is greater than or equal to the similarity threshold, then the empty detection result corresponding to the imaging region is determined to be empty, otherwise the empty detection result corresponding to the imaging region is determined to be non-empty.

[0077] In practice, multiple second detection regions are first randomly selected within the imaging area. The shape and number of these second detection regions are not limited here; for example... Figure 4 As shown, three square regions are selected as the second detection regions. Then, the similarity between these second detection regions is calculated. If the similarity between all second detection regions is greater than or equal to a set similarity threshold, it indicates that these regions are highly consistent in image features and are likely uniform imaging regions under empty conditions. Therefore, the empty detection result corresponding to the imaging region is determined to be empty. Conversely, if the similarity between any two second detection regions is less than the similarity threshold, it indicates that there are different image features between the regions, which may contain actual tissue structures or other non-empty content. Therefore, the empty detection result corresponding to the imaging region is determined to be non-empty. Optionally, to ensure the accuracy of the results, each second detection region can be completely within the imaging region; that is, the second detection region does not include the boundary region of the imaging region or the surrounding region.

[0078] Optionally, multiple empty detections can be performed in the above manner. After integrating the results of these empty detections, the empty detection result corresponding to the imaging area can be determined. For example, if the number of times it is determined to be empty is greater than a set number, then the empty detection result corresponding to the imaging area is determined to be empty. The set number can be 1 time, 2 times, etc.

[0079] As a feasible implementation method, the similarity between multiple second detection regions is calculated. If the similarity between multiple second detection regions is greater than or equal to a similarity threshold, the empty detection result corresponding to the imaging region is determined to be empty; otherwise, the empty detection result corresponding to the imaging region is determined to be non-empty. This includes: calculating the hash value of multiple second detection regions respectively, and calculating the Hamming distance between multiple hash values; if the multiple Hamming distances are all less than or equal to a distance threshold, the empty detection result corresponding to the imaging region is determined to be empty; otherwise, the empty detection result corresponding to the imaging region is determined to be non-empty.

[0080] In practice, the Hamming distance between the hash values ​​of each second detection region can be used to describe the similarity between them. A larger Hamming distance indicates lower similarity, and a smaller Hamming distance indicates higher similarity. Therefore, if the Hamming distance between the hash values ​​of all second detection regions is less than or equal to the distance threshold, the empty detection result corresponding to the imaging region is determined to be empty. Conversely, if the Hamming distance between the hash values ​​of any two second detection regions is greater than the distance threshold, the empty detection result corresponding to the imaging region is determined to be non-empty.

[0081] In the above implementation, the uniformity of the imaging area is determined by comparing the image features between different detection areas within the imaging area, thereby determining the empty detection result. The empty detection object is more targeted, and therefore, the corresponding empty detection result has good reliability.

[0082] The preceding sections described various implementation methods for empty detection, each with its own advantages. In practice, at least two methods can be employed, either serially or in parallel, to fully integrate their strengths, resulting in an empty detection process that combines efficiency and accuracy. Furthermore, empty detection can be performed sequentially according to the order of the ultrasound imaging process. If a particular empty detection result is empty, subsequent empty detections can be skipped, and the final empty detection result of the ultrasound image can be directly determined as an empty image, thus improving the efficiency of empty detection. Alternatively, a unified judgment can be made after all detections have been completed.

[0083] S103: Determine the final empty-load test result of the ultrasound image by combining multiple empty-load test results.

[0084] In this step, the final empty-load detection result of the ultrasound image is determined by combining the multiple empty-load detection results obtained above. As a feasible implementation method, determining the final empty-load detection result of the ultrasound image by combining multiple empty-load detection results includes: if any empty-load detection result is empty, then the final empty-load detection result of the ultrasound image is determined to be an empty image; otherwise, the final empty-load detection result of the ultrasound image is determined to be a non-empty image.

[0085] In practice, if any one of the multiple no-load test results is determined to be no-load, the final no-load test result of the ultrasound image will be determined to be a no-load image; only when all of these test results are determined to be non-no-load images will the final no-load test result of the ultrasound image be determined to be a non-no-load image. This "one-vote veto" mechanism effectively improves the accuracy of no-load detection and avoids misjudgments that may occur due to a single test link.

[0086] In other embodiments, if at least one of the multiple empty detection results is empty, then the final empty detection result of the ultrasound image is determined to be an empty image. That is, one or more empty detection results can be combined to determine the final empty detection result, which can improve processing efficiency while ensuring the accuracy of the result.

[0087] The applicant of this application discovered that ultrasound image recognition models cannot properly identify untrained content, especially AI (Artificial Intelligence) models related to classification or semantic segmentation, which often produce misidentifications on empty images. The applicant also discovered that by using empty image detection as a preprocessing step before inputting ultrasound images into the ultrasound image recognition model, and only when the image is determined to be non-empty, can the corresponding ultrasound image be input into the AI ​​model, this method can significantly improve the accuracy of the AI ​​model in recognizing ultrasound images.

[0088] As a feasible implementation method, the final empty detection result is used as the result of preprocessing the input image of the ultrasound image recognition model; accordingly, after determining the final empty detection result of the ultrasound image by combining multiple empty detection results, the method further includes: if the final empty detection result of the ultrasound image is a non-empty image, then the ultrasound image is input into the ultrasound image recognition model to obtain the image recognition result corresponding to the ultrasound image; otherwise, the ultrasound image is not input into the ultrasound image recognition model.

[0089] In practice, only non-empty ultrasound images are input into the ultrasound image recognition model, thereby avoiding misidentification of untrained non-empty ultrasound images and improving the recognition accuracy of the ultrasound recognition model. The ultrasound image recognition model here can be an AI model used for classification or semantic segmentation.

[0090] The ultrasound image processing method provided in this application performs multiple no-load detections on various types of data signals from multiple stages in the ultrasound imaging process, and combines the results of these multiple no-load detections to obtain the final no-load detection result of the ultrasound image. This avoids misjudgments that may occur due to a single detection stage, thereby effectively improving the accuracy of no-load image identification and preventing subsequent functions from making incorrect diagnoses or measurements of the no-load image. This effectively reduces the risk of misidentification and ensures the reliability and accuracy of ultrasound diagnostic and measurement functions. Furthermore, by performing multiple no-load detections on various types of data signals from multiple stages, this application can more comprehensively capture no-load characteristics, enabling the ultrasound system to more stably determine the no-load state under various conditions, enhancing the system's robustness, and improving the stability and reliability of the ultrasound equipment in practical applications.

[0091] This application discloses an ultrasound image processing method. Compared with the previous embodiment, this embodiment further explains and optimizes the technical solution. Specifically:

[0092] See Figure 5 A flowchart illustrating another ultrasound image processing method according to an exemplary embodiment, such as... Figure 5 As shown, it includes:

[0093] For the ultrasound image to be used as the input ultrasound image recognition model, perform the following steps:

[0094] S201: During the ultrasound imaging process, the ultrasound echo signal after beamforming is acquired, and it is determined whether the signal intensity of the ultrasound echo signal is greater than or equal to the intensity threshold; if so, the first no-load detection result is determined to be non-no-load; if not, the first no-load detection result is determined to be no-load.

[0095] S202: Obtain the initial wave image before digital scanning conversion, perform boundary detection on the initial wave image using the Canny operator, and extract the first detection region from the boundary detection result image; calculate the pixel mean of the pixels in the first detection region, and determine whether the pixel mean is greater than or equal to the pixel threshold; if yes, determine that the second empty detection result is empty; if no, determine that the second empty detection result is not empty.

[0096] Wherein, the center point of the first detection region is the center point of the boundary detection result image, the length of the first detection region is a preset proportion of the length of the boundary detection result image, and the width of the first detection region is a preset proportion of the width of the boundary detection result image;

[0097] S203: Acquire the ultrasound image after digital scanning conversion of the initial wave image, and input the ultrasound image into the trained empty image classification model to determine the third empty detection result;

[0098] In this step, the ultrasound image after digital scanning conversion is acquired and input into a pre-trained empty image classification model. This model, using machine learning or deep learning methods, such as a binary classification model trained with the ResNet50 classification algorithm, can distinguish between empty ultrasound images and non-empty ultrasound images containing tissue. The model analyzes the features of the input image to determine whether the image is empty, and then outputs a third empty detection result.

[0099] The training process for an empty image classification model can include: First, collecting a large amount of ultrasound image data, including images from both empty and non-empty scenarios. These images are then divided into training and validation sets, typically in an 80% training to 20% validation ratio. Preprocessing the collected images involves operations such as image normalization, cropping, and rotation to enhance the model's generalization ability. For example, pixel values ​​can be normalized to between 0 and 1 to improve training efficiency. A suitable deep learning model architecture, such as ResNet50, is chosen for the image classification task. Pre-trained model weights can be used for initialization, which speeds up convergence and improves performance. For example, pre-trained ResNet50 model weights can be downloaded from publicly available model libraries. A suitable loss function and optimizer are selected. For classification tasks, cross-entropy loss is typically used. The optimizer Adam can be chosen, as it accelerates convergence by adjusting the learning rate. Finally, the model is trained using the training set data. In each epoch (training round), the model calculates the output through forward propagation and then calculates the gradient and updates the model parameters through backpropagation. During training, several hyperparameters can be set, such as the learning rate, batch size, and number of training rounds (epochs). For example, a batch size of 16 and 100 training rounds can be set. After each epoch, the model is validated using a validation set to evaluate its performance. The loss and accuracy on the validation set are recorded for subsequent model tuning. If the model's performance no longer improves on the validation set, training can be stopped early to prevent overfitting. During training, the model weights for each epoch, as well as the weights of the model that performed best on the validation set, can be saved. This allows the best model to be selected for subsequent empty-load detection tasks after training. Through these steps, a high-performance empty-load image classification model can be trained for empty-load detection in ultrasound images, resulting in accurate third-party empty-load detection results.

[0100] S204: Extract a fan-shaped region from the ultrasound image, randomly select multiple second detection regions from the fan-shaped region, and calculate the similarity between the multiple second detection regions; if the similarity between the multiple second detection regions is greater than or equal to the similarity threshold, then the fourth empty detection result is determined to be empty; otherwise, the fourth empty detection result is determined to be non-empty; wherein, the multiple second detection regions have the same shape and size, and there is no overlap between any two second detection regions;

[0101] The above steps are described using a fan-shaped imaging area as an example. For imaging areas of other shapes, the shape and size can be adjusted accordingly.

[0102] S205: If any one of the first, second, third, and fourth empty-load detection results is empty, then the empty-load detection result corresponding to the ultrasound image is determined to be an empty-load image; otherwise, the empty-load detection result corresponding to the ultrasound image is determined to be a non-empty-load image.

[0103] In this step, the first, second, third, and fourth empty detection results obtained earlier are combined to determine the corresponding empty detection result for the ultrasound image. This comprehensive judgment method is equivalent to setting up "multiple insurances" for the empty detection of ultrasound images. As long as one of the detection results is empty, the ultrasound image is ultimately identified as empty, thereby effectively improving the accuracy and reliability of identifying empty images.

[0104] Furthermore, only when the unloaded detection result is a non-unloaded image is the corresponding ultrasound image input into the ultrasound image recognition model for image recognition. This effectively ensures that all images recognized by the ultrasound image recognition model are valid images, thereby improving the recognition accuracy of the ultrasound image recognition model.

[0105] The following describes an ultrasonic image processing apparatus provided in an embodiment of this application. The ultrasonic image processing apparatus described below and the ultrasonic image processing method described above can be referred to each other.

[0106] See Figure 6 A structural diagram of an ultrasound image processing apparatus according to an exemplary embodiment is shown, as follows: Figure 6 As shown, it includes:

[0107] The acquisition module 100 is used to acquire multiple types of data signals during ultrasound imaging; wherein, the multiple types of data signals include at least two of the following: ultrasound echo signal after beamforming, initial wave image before digital scan conversion, and ultrasound image after digital scan conversion.

[0108] The no-load detection module 200 is used to perform no-load detection on various types of data signals to obtain the corresponding no-load detection results.

[0109] The determination module 300 is used to determine the final empty-load detection result of the ultrasound image by integrating multiple empty-load detection results.

[0110] The ultrasonic image processing apparatus provided in this application performs multiple no-load detections on various types of data signals from multiple stages in the ultrasonic imaging process, and combines the results of these multiple no-load detections to obtain the final no-load detection result of the ultrasonic image. This avoids misjudgments that may occur due to a single detection stage, thereby effectively improving the accuracy of no-load image identification and preventing subsequent functions from making incorrect diagnoses or measurements of the no-load image. This effectively reduces the risk of misidentification and ensures the reliability and accuracy of ultrasonic diagnostic and measurement functions. Furthermore, by performing multiple no-load detections on various types of data signals from multiple stages, this application can more comprehensively capture no-load characteristics, enabling the ultrasonic system to more stably determine the no-load state under various conditions. This enhances the robustness of the system and improves the stability and reliability of the ultrasonic equipment in practical applications.

[0111] Based on the above embodiments, as a preferred embodiment, the no-load detection module 200 includes:

[0112] The first no-load detection submodule is used to determine whether the signal strength of the ultrasonic echo signal is greater than or equal to the strength threshold; if so, the no-load detection result corresponding to the ultrasonic echo signal is determined to be non-no-load; if not, the no-load detection result corresponding to the ultrasonic echo signal is determined to be no-load.

[0113] Based on the above embodiments, as a preferred embodiment, the no-load detection module 200 includes:

[0114] The second empty-load detection submodule is used to perform boundary detection on the initial wave image and determine the empty-load detection result corresponding to the initial wave image based on the boundary detection result.

[0115] Based on the above embodiments, as a preferred implementation, the second no-load detection submodule includes:

[0116] The boundary detection unit is used to perform boundary detection on the initial wave image using the Canny operator and to extract a first detection region from the boundary detection result image. The center point of the first detection region is the center point of the boundary detection result image, the length of the first detection region is a preset proportion of the length of the boundary detection result image, and the width of the first detection region is a preset proportion of the width of the boundary detection result image.

[0117] The judgment unit is used to calculate the average pixel value of the pixels in the first detection area and determine whether the average pixel value is greater than or equal to the pixel threshold. If it is, the empty detection result corresponding to the initial wave image is determined to be empty; if not, the empty detection result corresponding to the initial wave image is determined to be non-empty.

[0118] Based on the above embodiments, as a preferred embodiment, the no-load detection module 200 includes:

[0119] The third empty-load detection submodule is used to input ultrasound images into the trained empty-load image classification model to obtain the corresponding empty-load detection results.

[0120] Based on the above embodiments, as a preferred embodiment, the no-load detection module 200 includes:

[0121] The fourth empty detection submodule is used to extract the imaging area from the ultrasound image and perform empty detection on the imaging area to obtain the corresponding empty detection result.

[0122] Based on the above embodiments, as a preferred implementation, the fourth no-load detection submodule includes:

[0123] The cropping unit is used to randomly crop multiple second detection regions in the imaging region; wherein the multiple second detection regions have the same shape and size, and there is no overlap between any two second detection regions;

[0124] The calculation unit is used to calculate the similarity between multiple second detection regions. If the similarity between multiple second detection regions is greater than or equal to the similarity threshold, the empty detection result corresponding to the imaging region is determined to be empty; otherwise, the empty detection result corresponding to the imaging region is determined to be non-empty.

[0125] Based on the above embodiments, as a preferred implementation, the calculation unit is specifically used to: calculate the hash values ​​of multiple second detection regions respectively, and calculate the Hamming distance between the multiple hash values; if the multiple Hamming distances are all less than or equal to the distance threshold, then the empty detection result corresponding to the imaging region is determined to be empty, otherwise the empty detection result corresponding to the imaging region is determined to be non-empty.

[0126] Based on the above embodiments, as a preferred implementation, the determining module 300 is specifically used to: if any empty detection result is empty, then determine the final empty detection result of the ultrasound image as an empty image; otherwise, determine the final empty detection result of the ultrasound image as a non-empty image.

[0127] Based on the above embodiments, as a preferred embodiment, the final empty-load detection result is used as the result of preprocessing the input image for the ultrasound image recognition model; the device further includes:

[0128] The image recognition module is used to input the ultrasound image into the ultrasound image recognition model when the final empty detection result of the ultrasound image is a non-empty image, so as to obtain the image recognition result corresponding to the ultrasound image.

[0129] 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.

[0130] 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 electronic device. Figure 7 This is a structural diagram of an electronic device according to an exemplary embodiment, such as... Figure 7 As shown, the electronic device includes:

[0131] Communication interface 1 enables information exchange with other devices, such as network devices;

[0132] Processor 2 is connected to communication interface 1 to enable information exchange with other devices and to execute the ultrasound image processing methods provided by one or more of the above-mentioned technical solutions when running computer programs. The computer program is stored in memory 3.

[0133] Of course, in practical applications, the various components in an electronic device 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 7 The general will label all buses as Bus System 4.

[0134] The memory 3 in this embodiment is used to store various types of data to support the operation of the electronic device. Examples of such data include any computer program used to operate on the electronic device.

[0135] 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 3 described in the embodiments of this application is intended to include, but is not limited to, these and any other suitable types of memory.

[0136] 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.

[0137] 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.

[0138] 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 aforementioned method steps. The computer-readable storage medium can be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface memory, optical disc, CD-ROM, etc.

[0139] 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.

[0140] 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 electronic device (which may be a personal computer, server, network device, etc.) to execute all or part of the methods of 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.

[0141] The above are merely specific embodiments 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. An ultrasound image processing method, characterized by, include: Acquire multiple types of data signals during ultrasound imaging; wherein, the multiple types of data signals include at least two of the following: ultrasound echo signal after beamforming, initial wave image before digital scan conversion, and ultrasound image after digital scan conversion; The corresponding no-load detection results are obtained by performing no-load detection on various types of data signals respectively; The final empty-load detection result of the ultrasound image is determined by combining multiple empty-load detection results.

2. The method of processing ultrasound images according to claim 1, characterized in that, The ultrasonic echo signal is subjected to no-load detection to obtain the corresponding no-load detection result, including: Determine whether the signal strength of the ultrasonic echo signal is greater than or equal to the intensity threshold; If so, the no-load detection result corresponding to the ultrasonic echo signal is determined to be non-no-load; If not, the no-load detection result corresponding to the ultrasonic echo signal is determined to be no-load.

3. The method of claim 1, wherein, The initial wave image is subjected to no-load detection to obtain the corresponding no-load detection result, including: Boundary detection is performed on the initial wave image, and the empty load detection result corresponding to the initial wave image is determined based on the boundary detection result.

4. The ultrasound image processing method according to claim 3, characterized in that, Boundary detection is performed on the initial wave image, and the empty-load detection result corresponding to the initial wave image is determined based on the boundary detection result, including: The initial wave image is subjected to boundary detection using the Canny operator, and a first detection region is extracted from the boundary detection result image; wherein, the center point of the first detection region is the center point of the boundary detection result image, the length of the first detection region is a preset proportion of the length of the boundary detection result image, and the width of the first detection region is a preset proportion of the width of the boundary detection result image. Calculate the average pixel value of the pixels in the first detection region, and determine whether the average pixel value is greater than or equal to the pixel threshold; If so, the empty detection result corresponding to the initial wave image is determined to be empty; If not, the empty detection result corresponding to the initial wave image is determined to be non-empty.

5. The method of claim 1, wherein, Performing empty-load detection on the ultrasound image to obtain the corresponding empty-load detection result includes: The ultrasound image is input into the trained empty image classification model to obtain the corresponding empty detection result.

6. The method of claim 1, wherein, Performing empty-load detection on the ultrasound image to obtain the corresponding empty-load detection result includes: An imaging region is extracted from the ultrasound image, and an empty detection is performed on the imaging region to obtain the corresponding empty detection result.

7. The method of processing ultrasound images according to claim 6, wherein, Performing no-load detection on the imaging area to obtain the corresponding no-load detection result includes: Multiple second detection regions are randomly selected from the imaging region; wherein the multiple second detection regions have the same shape and size, and there is no overlap between any two second detection regions; Calculate the similarity between multiple second detection regions. If the similarity between multiple second detection regions is greater than or equal to the similarity threshold, then the empty detection result corresponding to the imaging region is determined to be empty; otherwise, the empty detection result corresponding to the imaging region is determined to be non-empty.

8. The method of processing ultrasound images according to claim 7, characterized in that, Calculate the similarity among multiple second detection regions. If the similarity among multiple second detection regions is greater than or equal to a similarity threshold, then the empty detection result corresponding to the imaging region is determined to be empty; otherwise, the empty detection result corresponding to the imaging region is determined to be non-empty. This includes: Calculate the hash values ​​of multiple second detection regions respectively, and calculate the Hamming distance between the multiple hash values; If all of the stated Hamming distances are less than or equal to the distance threshold, then the empty detection result corresponding to the imaging region is determined to be empty; otherwise, the empty detection result corresponding to the imaging region is determined to be non-empty.

9. The method of claim 1, wherein, The process of determining the final empty-load detection result of the ultrasound image by integrating multiple empty-load detection results includes: If any empty detection result is empty, then the final empty detection result of the ultrasound image is determined to be an empty image; Otherwise, the final empty detection result of the ultrasound image is determined to be a non-empty image.

10. The method of processing ultrasound images according to any one of claims 1 to 9, characterized in that, The final empty-load detection results are used as the input images for preprocessing the ultrasound image recognition model; Accordingly, after determining the final empty detection result of the ultrasound image by integrating multiple empty detection results, the method further includes: If the final empty detection result of the ultrasound image is a non-empty image, then the ultrasound image is input into the ultrasound image recognition model to obtain the image recognition result corresponding to the ultrasound image; otherwise, the ultrasound image is not input into the ultrasound image recognition model.

11. An ultrasound image processing apparatus, characterized by, include: The acquisition module is used to acquire multiple types of data signals during ultrasound imaging; wherein, the multiple types of data signals include at least two of the following: ultrasound echo signal after beamforming, initial wave image before digital scan conversion, and ultrasound image after digital scan conversion; The no-load detection module is used to perform no-load detection on various types of data signals to obtain the corresponding no-load detection results. The determination module is used to determine the final empty-load detection result of the ultrasound image by integrating multiple empty-load detection results.

12. An electronic device, comprising: include: Memory, used to store computer programs; A processor for executing the computer program to implement the steps of the ultrasound image processing method as described in any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed, implements the steps of the ultrasound image processing method as described in any one of claims 1 to 10.

Images