A method, apparatus, device, and storage medium for enhancing liver ultrasound images.
By automatically generating detailed enhancements to liver ultrasound images using deep convolutional neural networks, the problem of time-consuming and labor-intensive manual operations in existing technologies is solved, achieving efficient and reliable image generation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- JINHUA MUNICIPAL CENT HOSPITAL
- Filing Date
- 2026-02-12
- Publication Date
- 2026-06-02
AI Technical Summary
Existing liver ultrasound imaging enhancement techniques lack specificity, resulting in high resource consumption, long generation time, and insufficient adaptability, thus affecting generation efficiency and quality.
A deep convolutional neural network is used for feature extraction and generator processing to automatically enhance liver ultrasound images. The images are then automatically generated with enhanced details by interacting with Doppler ultrasound equipment through a data transmission channel.
It reduces generation time, improves generation efficiency and image reliability, ensures image quality is not affected by human intervention, and enhances image clarity and detail.
Smart Images
Figure CN122134576A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the fields of artificial intelligence technology and imaging technology, and in particular to a method, apparatus, device and storage medium for enhancing liver ultrasound images. Background Technology
[0002] Liver ultrasound imaging is an image formed by scanning the liver with ultrasound waves. Currently, due to limitations in the ultrasound imaging principle itself and noise interference generated during the operation of Doppler ultrasound equipment, the details in current liver ultrasound images are easily blurred. Therefore, detail enhancement processing of current liver ultrasound images is urgently needed.
[0003] However, current enhancement technologies mostly adopt a generalized design, which is not specifically optimized for current liver ultrasound images, resulting in insufficient specificity. Therefore, when current enhancement technologies are applied to current liver ultrasound images, their adaptability is severely inadequate, requiring manual adjustment of parameters to achieve precise optimization of the current liver ultrasound image. However, manual operation consumes a lot of human and time resources, increasing the generation time of the enhanced liver ultrasound image and hindering the improvement of the generation efficiency. Therefore, how to generate enhanced liver ultrasound images is an urgent problem to be solved. Summary of the Invention
[0004] This application provides a method, apparatus, device, and storage medium for enhancing liver ultrasound images, in order to solve the aforementioned technical problem of how to generate current liver ultrasound images with enhanced details.
[0005] In a first aspect, embodiments of this application provide a method for enhancing liver ultrasound images, applied to an electronic device equipped with a deep convolutional neural network. The electronic device establishes a data transmission channel with a Doppler ultrasound device. The enhancement method includes: The system sends data acquisition commands to the Doppler ultrasound device through the data transmission channel, receives response data returned by the Doppler ultrasound device based on the data acquisition commands, and extracts pixel data and attribute data from the response data. The pixel data is input into the image rendering engine, which then renders and processes the pixel data to generate the current liver ultrasound image. The current liver ultrasound image is the liver ultrasound image acquired by the Doppler ultrasound device at the current acquisition time. In the current liver ultrasound image, determine the location range of the liver region, extract the local image corresponding to the liver region as the liver region image, and obtain the average value of all pixel gray values in the liver region image; Based on the average gray value of all pixels in the liver region image and the preset ultrasound image scoring model, a visual quality score of the current liver ultrasound image is generated. When the visual quality score is lower than the preset score, the texture data, gradient data, and gray value data of the current liver ultrasound image are obtained from the attribute data of the current liver ultrasound image. The feature extractor in the trained deep convolutional neural network is used to extract features from the texture data, gradient data, and grayscale data of the current liver ultrasound image, respectively, generating texture features, edge features, and grayscale features of the current liver ultrasound image. These features are then concatenated to obtain the fused features of the current liver ultrasound image. The fused features are then input into the trained deep convolutional neural network, and the generator in the trained deep convolutional neural network generates a detailed enhanced liver ultrasound image.
[0006] In one possible implementation of the first aspect, a data acquisition command is sent to a Doppler ultrasound device via a data transmission channel, and response data is received from the Doppler ultrasound device according to the data acquisition command. Pixel data and attribute data are then obtained from the response data, including: The test packet is sent to the Doppler ultrasound device through the data transmission channel. The response packet returned by the Doppler ultrasound device according to the test packet is received. The sequence number field carried by the response packet and the sequence number field carried by the test packet are obtained. The serial number of the response packet is obtained from the sequence number field carried by the response packet, and the serial number of the test packet is obtained from the sequence number field carried by the test packet. When the sequence number of the response packet and the sequence number of the test packet are the same, retrieve the sequence number from the communication log. The reception time of the response packet and the transmission time of the test packet are used to calculate the communication delay of the data transmission channel within the statistical time by subtracting the transmission time of the test packet from the reception time of the response packet. When the communication delay is less than the preset delay, a data acquisition command is sent to the Doppler ultrasound device through the data transmission channel, and the response data returned by the Doppler ultrasound device according to the data acquisition command is received. Pixel data and attribute data are obtained from the response data.
[0007] In one possible implementation of the first aspect, the location range of the liver region is determined in the current liver ultrasound image, the local image corresponding to the liver region is extracted as the liver region image, and the average value of all pixel grayscale values within the liver region image is obtained, including: The current liver ultrasound image is input into a large visual model. The large visual model determines the location range of the liver region in the current liver ultrasound image, extracts the local image corresponding to the liver region as the liver region image, and obtains the average value of the gray values of all pixels in the liver region image.
[0008] In one possible implementation of the first aspect, the step of generating a visual quality score for the current liver ultrasound image based on the average grayscale value of all pixels in the liver region image and a preset ultrasound image scoring model, and when the visual quality score is lower than the preset score, obtaining the texture data, gradient data, and grayscale data of the current liver ultrasound image from the attribute data of the current liver ultrasound image, including: Obtain the maximum value, minimum value, and average value of all pixel grayscale values in the current liver ultrasound image from the pixel data. Based on the maximum value, minimum value, average value, and average value of all pixels in the current liver ultrasound image, the visual quality score of the current liver ultrasound image is generated using a preset ultrasound image scoring model. When the visual quality score is lower than the preset score, the texture data, gradient data, and grayscale data of the current liver ultrasound image are obtained from the attribute data of the current liver ultrasound image.
[0009] In one possible implementation of the first aspect, the ultrasound image scoring model is defined as follows: ; This indicates the visual quality score of the current liver ultrasound image; a higher visual quality score means better visual quality of the current liver ultrasound image, and a lower visual quality score means worse visual quality of the current liver ultrasound image. This represents the maximum grayscale value of all pixels in the current liver ultrasound image; This represents the minimum grayscale value of all pixels in the current liver ultrasound image; This represents the average grayscale value of all pixels in the current liver ultrasound image. This represents the average grayscale value of all pixels in the liver region image, which is the local image of the liver region in the current liver ultrasound image. and The larger the ratio, the greater the brightness contrast between the liver region and the surrounding region, and the clearer the boundary of the liver region. and The larger the ratio, the smaller the brightness contrast between the liver area and the surrounding area, and the less clear the boundary of the liver area. This indicates the noise level of the current liver ultrasound image.
[0010] In one possible implementation of the first aspect, the feature extractor in the trained deep convolutional neural network extracts features from the texture data, gradient data, and grayscale data of the current liver ultrasound image, respectively, generating texture features, edge features, and grayscale features of the current liver ultrasound image. These features are then concatenated to obtain a fused feature of the current liver ultrasound image. This fused feature is then input into the trained deep convolutional neural network, and a generator in the trained deep convolutional neural network generates a detail-enhanced current liver ultrasound image. The enhancement method then includes: Create a display window to show the current liver ultrasound image with enhanced details.
[0011] In one possible implementation of the first aspect, before sending a data acquisition command to the Doppler ultrasound device via a data transmission channel, receiving response data returned by the Doppler ultrasound device according to the data acquisition command, and obtaining pixel data and attribute data from the response data, the enhancement method includes: Acquire preset liver ultrasound images, which are liver ultrasound images acquired by the Doppler ultrasound device before the current acquisition time; The feature extractor in the deep convolutional neural network is used to extract features from the texture data, gradient data, and grayscale data of the preset liver ultrasound image, respectively, to generate the texture features, edge features, and grayscale features of the preset liver ultrasound image. The texture features, edge features, and grayscale features of a preset liver ultrasound image are concatenated to obtain a fusion feature of the preset liver ultrasound image. The fusion feature and the enhanced-detail preset liver ultrasound image are combined to form a sample. Different samples are used to form a training set. The first loss value of the deep convolutional neural network on the training set is obtained through the absolute error loss function, and the second loss value of the deep convolutional neural network on the training set is obtained through the mean squared error loss function. The first loss value and the second loss value are added to generate the total loss value of the deep convolutional neural network on the training set. The training objective is to reduce the total loss value. The deep convolutional neural network is iteratively trained until the total loss value is less than the preset loss value. The training of the deep convolutional neural network is then stopped, and the trained deep convolutional neural network is saved.
[0012] Secondly, embodiments of this application provide a liver ultrasound image enhancement device, applied to an electronic device deployed with a deep convolutional neural network, wherein the electronic device establishes a data transmission channel with a Doppler ultrasound device, including: The first acquisition module is used to send a data acquisition command to the Doppler ultrasound device through the data transmission channel, receive the response data returned by the Doppler ultrasound device according to the data acquisition command, and acquire pixel data and attribute data from the response data; The input module is used to input pixel data into the image rendering engine. The image rendering engine renders the pixel data to generate the current liver ultrasound image, which is the liver ultrasound image acquired by the Doppler ultrasound device at the current acquisition time. The second acquisition module is used to determine the location range of the liver region in the current liver ultrasound image, extract the local image corresponding to the liver region as the liver region image, and obtain the average value of all pixel gray values in the liver region image. The third acquisition module is used to generate a visual quality score of the current liver ultrasound image based on the average gray value of all pixels in the liver region image and a preset ultrasound image scoring model. When the visual quality score is lower than the preset score, the texture data, gradient data, and gray value data of the current liver ultrasound image are obtained from the attribute data of the current liver ultrasound image. The generation module uses the feature extractor in the trained deep convolutional neural network to extract features from the texture data, gradient data, and grayscale data of the current liver ultrasound image, respectively, generating texture features, edge features, and grayscale features of the current liver ultrasound image. These features are then concatenated to obtain a fused feature of the current liver ultrasound image. This fused feature is then input into the trained deep convolutional neural network, and the generator within the network generates a detailed enhanced version of the current liver ultrasound image.
[0013] Thirdly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the enhanced method described in the first aspect above.
[0014] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the enhanced method described in the first aspect.
[0015] Fifthly, embodiments of this application provide a computer program product that, when run on an electronic device, causes the electronic device to execute the enhanced method described in the first aspect.
[0016] The beneficial effects of the embodiments of this application are as follows: Firstly, the feature extractor in the trained deep convolutional neural network is used to extract features from the texture data, gradient data, and grayscale data of the current liver ultrasound image, respectively, generating texture features, edge features, and grayscale features of the current liver ultrasound image. These features are then concatenated to obtain the fusion features of the current liver ultrasound image. The fusion features are then input into the trained deep convolutional neural network, and the generator in the trained deep convolutional neural network generates a detail-enhanced current liver ultrasound image. Since no manual operation is required, the generation time of the detail-enhanced current liver ultrasound image is reduced, which helps to improve the generation efficiency of the detail-enhanced current liver ultrasound image. Secondly, since the current liver ultrasound image with enhanced details is automatically generated, it is not affected by human intervention, which helps to improve the reliability of the current liver ultrasound image with enhanced details. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.
[0018] Figure 1 An application scenario diagram of the enhancement method provided in the embodiments of this application; Figure 2 This is a flowchart illustrating the enhancement method provided in an embodiment of this application; Figure 3 A flowchart of S201 provided in the embodiments of this application; Figure 4 A schematic block diagram of the enhancement device provided in the embodiments of this application; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation
[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.
[0020] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.
[0021] It should be understood that in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance. The terms "comprising," "including," "having," and their variations all mean "including but not limited to," unless otherwise specifically emphasized.
[0022] Furthermore, the technical solutions of the various embodiments can be combined with each other, but only if they are based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection claimed in this application.
[0023] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.
[0024] The enhancement method provided in this application can be applied to electronic devices that have deployed deep convolutional neural networks. These electronic devices include, but are not limited to, servers, mobile phones, tablets, and laptops. This application does not impose any restrictions on the specific type of electronic device.
[0025] Please see Figure 1 , Figure 1 The application scenario diagram of the enhancement method provided in the embodiments of this application is described in detail below: The electronic device establishes a data transmission channel with the Doppler ultrasound device, sends data acquisition commands to the Doppler ultrasound device through the data transmission channel, receives response data returned by the Doppler ultrasound device according to the data acquisition commands, and obtains pixel data and attribute data from the response data.
[0026] The electronic device establishes a data transmission channel with the Doppler ultrasound device through a preset network, which is either a WIFI network or an Ethernet network.
[0027] Doppler ultrasound equipment is an imaging device that utilizes ultrasound waves and the Doppler effect. It can detect the internal structure and the motion of internal materials simultaneously without damaging the object's appearance. Doppler ultrasound equipment can be flexibly designed as desktop or portable models, and is simple to operate and highly adaptable.
[0028] In this embodiment, the electronic device establishes a data transmission channel with the Doppler ultrasound device, sends a data acquisition command to the Doppler ultrasound device through the data transmission channel, and receives the response data returned by the Doppler ultrasound device according to the data acquisition command through the data transmission channel. This allows for rapid acquisition of response data, greatly shortening the acquisition time and improving the efficiency of response data acquisition.
[0029] Please see Figure 2 , Figure 2 This is a flowchart illustrating the enhancement method provided in this application embodiment. This method can be applied to electronic devices with deployed deep convolutional neural networks, whereby the electronic devices establish a data transmission channel with Doppler ultrasound equipment.
[0030] like Figure 2 As shown, the enhancement method provided in this application includes the following steps, detailed below: S201, send a data acquisition command to the Doppler ultrasound device through the data transmission channel, receive the response data returned by the Doppler ultrasound device according to the data acquisition command, and obtain pixel data and attribute data from the response data; The enhancement method, prior to sending a data acquisition command to the Doppler ultrasound device via the data transmission channel, receiving response data returned by the Doppler ultrasound device according to the data acquisition command, and obtaining pixel data and attribute data from the response data, includes: Acquire preset liver ultrasound images, which are liver ultrasound images acquired by the Doppler ultrasound device before the current acquisition time; The feature extractor in the deep convolutional neural network is used to extract features from the texture data, gradient data, and grayscale data of the preset liver ultrasound image, respectively, to generate the texture features, edge features, and grayscale features of the preset liver ultrasound image. The texture features, edge features, and grayscale features of a preset liver ultrasound image are concatenated to obtain a fusion feature of the preset liver ultrasound image. The fusion feature and the enhanced-detail preset liver ultrasound image are combined to form a sample. Different samples are used to form a training set. The first loss value of the deep convolutional neural network on the training set is obtained through the absolute error loss function, and the second loss value of the deep convolutional neural network on the training set is obtained through the mean squared error loss function. The first loss value and the second loss value are added to generate the total loss value of the deep convolutional neural network on the training set. The training objective is to reduce the total loss value. The deep convolutional neural network is iteratively trained until the total loss value is less than the preset loss value. The training of the deep convolutional neural network is then stopped, and the trained deep convolutional neural network is saved.
[0031] The training objective is to reduce the total loss value. The deep convolutional neural network is iteratively trained until the total loss value is less than the preset loss value. The training of the deep convolutional neural network is then stopped and the trained deep convolutional neural network is saved to ensure that the trained deep convolutional neural network has a stable detail enhancement capability in the liver ultrasound image enhancement task.
[0032] S202, input the pixel data into the image rendering engine, and render the pixel data through the image rendering engine to generate the current liver ultrasound image. The current liver ultrasound image is the liver ultrasound image acquired by the Doppler ultrasound device at the current acquisition time. Specifically, the location and extent of the liver region are determined in the current liver ultrasound image, the corresponding local image is extracted as the liver region image, and the average grayscale value of all pixels within the liver region image is obtained, including: The current liver ultrasound image is input into a large visual model. The large visual model determines the location range of the liver region in the current liver ultrasound image, extracts the local image corresponding to the liver region as the liver region image, and obtains the average value of the gray values of all pixels in the liver region image.
[0033] S203, determine the location range of the liver region in the current liver ultrasound image, extract the local image corresponding to the liver region as the liver region image, and obtain the average value of all pixel gray values in the liver region image. S204. Based on the average gray value of all pixels in the liver region image and the preset ultrasound image scoring model, generate a visual quality score for the current liver ultrasound image. When the visual quality score is lower than the preset score, obtain the texture data, gradient data, and gray value data of the current liver ultrasound image from the attribute data of the current liver ultrasound image. The ultrasound image scoring model is a model that automatically quantifies and evaluates the visual quality of current liver ultrasound images. The higher the visual quality score, the higher the detail reproduction of the current liver ultrasound image; the lower the visual quality score, the lower the detail reproduction of the current liver ultrasound image.
[0034] When the visual quality score is lower than the preset score, it means that the detail reproduction of the current liver ultrasound image does not meet the preset quality standard, and the liver ultrasound image needs to be enhanced. This is done by obtaining the texture data, gradient data, and grayscale data of the current liver ultrasound image from the attribute data of the current liver ultrasound image.
[0035] When the visual quality score is not lower than the preset score, it means that the detail restoration of the current liver ultrasound image has not reached the preset quality standard. Therefore, it is not necessary to enhance the current liver ultrasound image. The acquisition of texture data, gradient data, and grayscale data of the current liver ultrasound image from the attribute data of the current liver ultrasound image should be stopped.
[0036] Specifically, the step of generating a visual quality score for the current liver ultrasound image based on the average grayscale value of all pixels in the liver region image and a preset ultrasound image scoring model, and when the visual quality score is lower than the preset score, obtaining the texture data, gradient data, and grayscale data of the current liver ultrasound image from the attribute data of the current liver ultrasound image, including: Obtain the maximum value, minimum value, and average value of all pixel grayscale values in the current liver ultrasound image from the pixel data. Based on the maximum value, minimum value, average value, and average value of all pixels in the current liver ultrasound image, the visual quality score of the current liver ultrasound image is generated using a preset ultrasound image scoring model. When the visual quality score is lower than the preset score, the texture data, gradient data, and grayscale data of the current liver ultrasound image are obtained from the attribute data of the current liver ultrasound image.
[0037] The ultrasound image scoring model is defined as follows: ; This indicates the visual quality score of the current liver ultrasound image; a higher visual quality score means better visual quality of the current liver ultrasound image, and a lower visual quality score means worse visual quality of the current liver ultrasound image. This represents the maximum grayscale value of all pixels in the current liver ultrasound image; This represents the minimum grayscale value of all pixels in the current liver ultrasound image; This represents the average grayscale value of all pixels in the current liver ultrasound image. This represents the average grayscale value of all pixels in the liver region image, which is the local image of the liver region in the current liver ultrasound image. and The larger the ratio, the greater the brightness contrast between the liver region and the surrounding region, and the clearer the boundary of the liver region. and The larger the ratio, the smaller the brightness contrast between the liver area and the surrounding area, and the less clear the boundary of the liver area. This indicates the noise level of the current liver ultrasound image.
[0038] The noise intensity of the current liver ultrasound image can be directly obtained through the noise detection module. The noise detection module performs grayscale statistics and regional analysis on the current liver ultrasound image, calculates the grayscale standard deviation of the uniform region in the current liver ultrasound image, and uses the grayscale standard deviation of the uniform region in the current liver ultrasound image as the noise intensity of the current liver ultrasound image.
[0039] S205 uses the feature extractor in the trained deep convolutional neural network to extract features from the texture data, gradient data, and grayscale data of the current liver ultrasound image, respectively, generating texture features, edge features, and grayscale features of the current liver ultrasound image. The texture features, edge features, and grayscale features of the current liver ultrasound image are then concatenated to obtain the fused features of the current liver ultrasound image. The fused features of the current liver ultrasound image are then input into the trained deep convolutional neural network, and the generator in the trained deep convolutional neural network generates a detailed enhanced current liver ultrasound image.
[0040] The generator in the trained deep convolutional neural network generates enhanced liver ultrasound images that can automatically improve image clarity, contrast and detail without human intervention. This enhances the edge contours, internal textures and fine structural information of the liver region, suppresses noise interference and improves image blur.
[0041] For ease of explanation, the following example is provided: For example, through automatic detail enhancement processing, the boundary between the liver and surrounding tissues can be made clearer, and weak signal structures such as fine textures and blood vessel directions within the liver can be made more prominent and visible, thereby improving the observability and analytical accuracy of the images and providing higher quality and more stable and reliable image data support for subsequent medical research, medical teaching and other scenarios.
[0042] The enhancement method, wherein the feature extractor in the trained deep convolutional neural network extracts features from the texture data, gradient data, and grayscale data of the current liver ultrasound image, respectively, generating texture features, edge features, and grayscale features of the current liver ultrasound image. These features are then concatenated to obtain a fused feature of the current liver ultrasound image. This fused feature is then input into the trained deep convolutional neural network, and the generator within the network generates a detailed enhanced liver ultrasound image. Create a display window to show the current liver ultrasound image with enhanced details.
[0043] Among them, generating enhanced detail current liver ultrasound images can significantly improve the clarity and detail of current liver ultrasound images, enhance the texture, boundaries and fine structural features of liver tissue, and make current liver ultrasound images easier to identify.
[0044] For ease of explanation, the following example is provided: For example, in medical research, enhanced liver ultrasound images can help researchers observe subtle structural changes inside the liver more clearly, more accurately distinguish liver lobes, liver segments, blood vessel orientation and tissue texture features, and more systematically summarize the morphological patterns and structural characteristics of the liver, providing more reliable data for basic research related to the liver. In medical teaching settings, enhanced liver ultrasound images allow students to see the liver's true structure more clearly, making it easier to understand its location, shape, and internal distribution. This helps them quickly build an intuitive and systematic understanding, making the teaching content easier to comprehend and learning more efficient. It also helps students lay a solid professional foundation and improves the overall quality of teaching.
[0045] The beneficial effects of the embodiments of this application are as follows: Firstly, the feature extractor in the trained deep convolutional neural network is used to extract features from the texture data, gradient data, and grayscale data of the current liver ultrasound image, respectively, generating texture features, edge features, and grayscale features of the current liver ultrasound image. These features are then concatenated to obtain the fusion features of the current liver ultrasound image. The fusion features are then input into the trained deep convolutional neural network, and the generator in the trained deep convolutional neural network generates a detail-enhanced current liver ultrasound image. Since no manual operation is required, the generation time of the detail-enhanced current liver ultrasound image is reduced, which helps to improve the generation efficiency of the detail-enhanced current liver ultrasound image. Secondly, since the current liver ultrasound image with enhanced details is automatically generated, it is not affected by human intervention, which helps to improve the reliability of the current liver ultrasound image with enhanced details.
[0046] Please see Figure 3 , Figure 3 The flowchart of S201 provided for the embodiments of this application is described in detail below: S301, send a test packet to the Doppler ultrasound device through the data transmission channel, receive the response packet returned by the Doppler ultrasound device according to the test packet, obtain the serial number field carried by the response packet and the serial number field carried by the test packet, obtain the serial number of the response packet from the serial number field carried by the response packet, and obtain the serial number of the test packet from the serial number field carried by the test packet. S302, when the sequence number of the response packet and the sequence number of the test packet are the same, obtain the reception time of the response packet and the sending time of the test packet from the communication log, subtract the sending time of the test packet from the reception time of the response packet, and generate the communication delay of the data transmission channel within the statistical time. When the sequence number of the response packet is the same as that of the test packet, it indicates that there was no packet loss or out-of-order delivery during the data transmission process, thus ensuring the reliability of the data transmission channel.
[0047] When the sequence number of the response packet and the sequence number of the test packet are different, it indicates that there is packet loss or out-of-order transmission during the data transmission process. Stop obtaining the reception time of the response packet and the transmission time of the test packet from the communication log.
[0048] S303: When the communication delay is less than the preset delay, a data acquisition command is sent to the Doppler ultrasound device through the data transmission channel, and the response data returned by the Doppler ultrasound device according to the data acquisition command is received. Pixel data and attribute data are obtained from the response data.
[0049] When the communication delay is less than the preset delay, it means that the data transmission channel does not experience excessive delay during the data transmission process, thus ensuring the real-time performance of the data transmission channel.
[0050] When the communication delay is not less than the preset delay, it indicates that the data transmission channel has excessive delay during the data transmission process, and the data acquisition command is not sent to the Doppler ultrasound device through the data transmission channel.
[0051] In this embodiment, when the communication delay is less than a preset delay, a data acquisition command is sent to the Doppler ultrasound device through the data transmission channel, and the response data returned by the Doppler ultrasound device according to the data acquisition command is received. Pixel data and attribute data are obtained from the response data. This avoids transmission errors and excessive transmission delays in pixel data and attribute data, ensuring the integrity of pixel data and attribute data. This ensures the integrity of the current liver ultrasound image generated based on the pixel data, thereby improving the imaging quality and stability of the current liver ultrasound image.
[0052] For the enhancement method described in the above embodiments, please refer to [link / reference]. Figure 4 , Figure 4 This is a schematic block diagram of the enhancement device provided in the embodiments of this application. Figure 4The enhancement device 400 shown can be applied to, for example Figure 1 The application scenario diagram shows electronic devices. The following section uses electronic devices as an example to illustrate this. Figure 4 The enhancement device 400 shown will be described in detail. The enhancement device 400 may include a first acquisition module 401, an input module 402, a second acquisition module 403, a third acquisition module 404, and a generation module 405.
[0053] The first acquisition module 401 is used to send a data acquisition command to the Doppler ultrasound device through the data transmission channel, receive the response data returned by the Doppler ultrasound device according to the data acquisition command, and acquire pixel data and attribute data from the response data. The input module 402 is used to input pixel data into the image rendering engine, and to render the pixel data through the image rendering engine to generate the current liver ultrasound image. The current liver ultrasound image is the liver ultrasound image acquired by the Doppler ultrasound device at the current acquisition time. The second acquisition module 403 is used to determine the location range of the liver region in the current liver ultrasound image, extract the local image corresponding to the liver region as the liver region image, and obtain the average value of all pixel gray values in the liver region image. The third acquisition module 404 is used to generate a visual quality score of the current liver ultrasound image based on the average gray value of all pixels in the liver region image and a preset ultrasound image scoring model. When the visual quality score is lower than the preset score, the texture data, gradient data, and gray value data of the current liver ultrasound image are obtained from the attribute data of the current liver ultrasound image. The generation module 405 is used to extract features from the texture data, gradient data, and grayscale data of the current liver ultrasound image using the feature extractor in the trained deep convolutional neural network. This generates texture features, edge features, and grayscale features of the current liver ultrasound image. The texture features, edge features, and grayscale features of the current liver ultrasound image are then concatenated to obtain the fused features of the current liver ultrasound image. The fused features of the current liver ultrasound image are then input into the trained deep convolutional neural network, and the generator in the trained deep convolutional neural network generates a detailed enhanced current liver ultrasound image.
[0054] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.
[0055] The beneficial effects of the embodiments of this application are as follows: Firstly, the feature extractor in the trained deep convolutional neural network is used to extract features from the texture data, gradient data, and grayscale data of the current liver ultrasound image, respectively, generating texture features, edge features, and grayscale features of the current liver ultrasound image. These features are then concatenated to obtain the fusion features of the current liver ultrasound image. The fusion features are then input into the trained deep convolutional neural network, and the generator in the trained deep convolutional neural network generates a detail-enhanced current liver ultrasound image. Since no manual operation is required, the generation time of the detail-enhanced current liver ultrasound image is reduced, which helps to improve the generation efficiency of the detail-enhanced current liver ultrasound image. Secondly, since the current liver ultrasound image with enhanced details is automatically generated, it is not affected by human intervention, which helps to improve the reliability of the current liver ultrasound image with enhanced details.
[0056] Please see Figure 5 , Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.
[0057] like Figure 5 As shown, Figure 5 The electronic device includes: at least one processor 20, a memory 21, and a computer program 22 stored in the memory 21 and executable on the at least one processor 20, wherein the processor 20 executes the computer program 22 to implement the steps in any of the above method embodiments.
[0058] The electronic device may include, but is not limited to, processor 20 and memory 21. Those skilled in the art will understand that... Figure 5 This is merely an example of an electronic device and does not constitute a limitation on electronic devices. It may include more or fewer components than shown in the illustration, or combinations of certain components, or different components. For example, it may also include input / output devices, network access devices, etc.
[0059] The processor 20 is used to run a computer program 22 stored in the memory 21, and performs the following steps when executing the computer program 22: The system sends data acquisition commands to the Doppler ultrasound device through the data transmission channel, receives response data returned by the Doppler ultrasound device based on the data acquisition commands, and extracts pixel data and attribute data from the response data. The pixel data is input into the image rendering engine, which then renders and processes the pixel data to generate the current liver ultrasound image. The current liver ultrasound image is the liver ultrasound image acquired by the Doppler ultrasound device at the current acquisition time. In the current liver ultrasound image, determine the location range of the liver region, extract the local image corresponding to the liver region as the liver region image, and obtain the average value of all pixel gray values in the liver region image; Based on the average gray value of all pixels in the liver region image and the preset ultrasound image scoring model, a visual quality score of the current liver ultrasound image is generated. When the visual quality score is lower than the preset score, the texture data, gradient data, and gray value data of the current liver ultrasound image are obtained from the attribute data of the current liver ultrasound image. The feature extractor in the trained deep convolutional neural network is used to extract features from the texture data, gradient data, and grayscale data of the current liver ultrasound image, respectively, generating texture features, edge features, and grayscale features of the current liver ultrasound image. These features are then concatenated to obtain the fused features of the current liver ultrasound image. The fused features are then input into the trained deep convolutional neural network, and the generator in the trained deep convolutional neural network generates a detailed enhanced liver ultrasound image.
[0060] The processor 20 may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors, field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.
[0061] In some embodiments, the memory 21 may be an internal storage unit of the electronic device, such as a hard disk or memory of the electronic device. In other embodiments, the memory 21 may also be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on the electronic device.
[0062] Furthermore, the memory 21 may include both internal storage units and external storage devices of the electronic device. The memory 21 is used to store the operating system, applications, boot loader, data, and other programs, such as the program code of the computer program. The memory 21 can also be used to temporarily store data that has been output or will be output.
[0063] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.
[0064] This application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.
[0065] The computer-readable storage medium may also be an external storage device of the enhancement device or electronic device, such as a plug-in hard drive, smart media card (SMC), secure digital (SD) card, flash card, or non-transitory computer-readable storage medium equipped on the enhancement device or electronic device.
[0066] Since the computer program stored in the computer-readable storage medium can execute any of the liver ultrasound image enhancement methods provided in the embodiments of this application, the computer-readable storage medium can achieve the beneficial effects that any of the liver ultrasound image enhancement methods provided in the embodiments of this application can achieve, as detailed in the preceding embodiments, and will not be repeated here.
[0067] This application provides a computer program product that, when run on an electronic device, causes the electronic device to perform the aforementioned enhancement method.
[0068] When a computer program is loaded into an electronic device, it can perform the following steps: The system sends data acquisition commands to the Doppler ultrasound device through the data transmission channel, receives response data returned by the Doppler ultrasound device based on the data acquisition commands, and extracts pixel data and attribute data from the response data. The pixel data is input into the image rendering engine, which then renders and processes the pixel data to generate the current liver ultrasound image. The current liver ultrasound image is the liver ultrasound image acquired by the Doppler ultrasound device at the current acquisition time. In the current liver ultrasound image, determine the location range of the liver region, extract the local image corresponding to the liver region as the liver region image, and obtain the average value of all pixel gray values in the liver region image; Based on the average gray value of all pixels in the liver region image and the preset ultrasound image scoring model, a visual quality score of the current liver ultrasound image is generated. When the visual quality score is lower than the preset score, the texture data, gradient data, and gray value data of the current liver ultrasound image are obtained from the attribute data of the current liver ultrasound image. The feature extractor in the trained deep convolutional neural network is used to extract features from the texture data, gradient data, and grayscale data of the current liver ultrasound image, respectively, generating texture features, edge features, and grayscale features of the current liver ultrasound image. These features are then concatenated to obtain the fused features of the current liver ultrasound image. The fused features are then input into the trained deep convolutional neural network, and the generator in the trained deep convolutional neural network generates a detailed enhanced liver ultrasound image.
[0069] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium.
[0070] Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium includes: an entity or device for carrying computer program code to an electronic device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium.
[0071] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.
[0072] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.
Claims
1. A method for enhancing liver ultrasound images, characterized in that, An enhancement method is applied to an electronic device equipped with a deep convolutional neural network, wherein the electronic device establishes a data transmission channel with a Doppler ultrasound device, and the enhancement method includes: The system sends data acquisition commands to the Doppler ultrasound device through the data transmission channel, receives response data returned by the Doppler ultrasound device based on the data acquisition commands, and extracts pixel data and attribute data from the response data. The pixel data is input into the image rendering engine, which then renders and processes the pixel data to generate the current liver ultrasound image. The current liver ultrasound image is the liver ultrasound image acquired by the Doppler ultrasound device at the current acquisition time. In the current liver ultrasound image, determine the location range of the liver region, extract the local image corresponding to the liver region as the liver region image, and obtain the average value of all pixel gray values in the liver region image; Based on the average gray value of all pixels in the liver region image and the preset ultrasound image scoring model, a visual quality score of the current liver ultrasound image is generated. When the visual quality score is lower than the preset score, the texture data, gradient data, and gray value data of the current liver ultrasound image are obtained from the attribute data of the current liver ultrasound image. The feature extractor in the trained deep convolutional neural network is used to extract features from the texture data, gradient data, and grayscale data of the current liver ultrasound image, respectively, generating texture features, edge features, and grayscale features of the current liver ultrasound image. These features are then concatenated to obtain the fused features of the current liver ultrasound image. The fused features are then input into the trained deep convolutional neural network, and the generator in the trained deep convolutional neural network generates a detailed enhanced liver ultrasound image.
2. The enhancement method according to claim 1, characterized in that, The system sends data acquisition commands to the Doppler ultrasound device via a data transmission channel, receives response data from the Doppler ultrasound device based on the data acquisition commands, and extracts pixel data and attribute data from the response data, including: The test packet is sent to the Doppler ultrasound device through the data transmission channel. The response packet returned by the Doppler ultrasound device according to the test packet is received. The sequence number field carried by the response packet and the sequence number field carried by the test packet are obtained. The serial number of the response packet is obtained from the sequence number field carried by the response packet, and the serial number of the test packet is obtained from the sequence number field carried by the test packet. When the sequence number of the response packet and the sequence number of the test packet are the same, retrieve the sequence number from the communication log. The reception time of the response packet and the transmission time of the test packet are used to calculate the communication delay of the data transmission channel within the statistical time by subtracting the transmission time of the test packet from the reception time of the response packet. When the communication delay is less than the preset delay, a data acquisition command is sent to the Doppler ultrasound device through the data transmission channel, and the response data returned by the Doppler ultrasound device according to the data acquisition command is received. Pixel data and attribute data are obtained from the response data.
3. The enhancement method according to claim 1, characterized in that, The location and extent of the liver region are determined in the current liver ultrasound image. The corresponding local image of the liver region is extracted as the liver region image, and the average gray value of all pixels within the liver region image is obtained, including: The current liver ultrasound image is input into a large visual model. The large visual model determines the location range of the liver region in the current liver ultrasound image, extracts the local image corresponding to the liver region as the liver region image, and obtains the average value of the gray values of all pixels in the liver region image.
4. The enhancement method according to claim 1, characterized in that, The process involves generating a visual quality score for the current liver ultrasound image based on the average grayscale value of all pixels within the liver region image and a preset ultrasound image scoring model. When the visual quality score is lower than the preset score, the texture data, gradient data, and grayscale data of the current liver ultrasound image are obtained from the attribute data of the current liver ultrasound image, including: Obtain the maximum value, minimum value, and average value of all pixel grayscale values in the current liver ultrasound image from the pixel data. Based on the maximum value, minimum value, average value, and average value of all pixels in the current liver ultrasound image, the visual quality score of the current liver ultrasound image is generated using a preset ultrasound image scoring model. When the visual quality score is lower than the preset score, the texture data, gradient data, and grayscale data of the current liver ultrasound image are obtained from the attribute data of the current liver ultrasound image.
5. The enhancement method according to claim 1, characterized in that, The ultrasound image scoring model is defined as follows: ； This indicates the visual quality score of the current liver ultrasound image; a higher visual quality score means better visual quality of the current liver ultrasound image, and a lower visual quality score means worse visual quality of the current liver ultrasound image. This represents the maximum grayscale value of all pixels in the current liver ultrasound image; This represents the minimum grayscale value of all pixels in the current liver ultrasound image; This represents the average grayscale value of all pixels in the current liver ultrasound image. This represents the average grayscale value of all pixels in the liver region image, which is the local image of the liver region in the current liver ultrasound image. and The larger the ratio, the greater the brightness contrast between the liver region and the surrounding region, and the clearer the boundary of the liver region. and The larger the ratio, the smaller the brightness contrast between the liver area and the surrounding area, and the less clear the boundary of the liver area. This indicates the noise level of the current liver ultrasound image.
6. The enhancement method according to claim 1, characterized in that, In the feature extractor of the trained deep convolutional neural network, features are extracted from the texture data, gradient data, and grayscale data of the current liver ultrasound image, respectively, generating texture features, edge features, and grayscale features of the current liver ultrasound image. These features are then concatenated to obtain a fused feature of the current liver ultrasound image. This fused feature is then input into the trained deep convolutional neural network. After the generator in the trained deep convolutional neural network generates a detailed enhanced liver ultrasound image, the enhancement method includes: Create a display window to show the current liver ultrasound image with enhanced details.
7. The enhancement method according to claim 1, characterized in that, Before sending a data acquisition command to the Doppler ultrasound device via the data transmission channel, receiving response data returned by the Doppler ultrasound device according to the data acquisition command, and obtaining pixel data and attribute data from the response data, the enhancement method includes: Acquire preset liver ultrasound images, which are liver ultrasound images acquired by the Doppler ultrasound device before the current acquisition time; The feature extractor in the deep convolutional neural network is used to extract features from the texture data, gradient data, and grayscale data of the preset liver ultrasound image, respectively, to generate the texture features, edge features, and grayscale features of the preset liver ultrasound image. The texture features, edge features, and grayscale features of a preset liver ultrasound image are concatenated to obtain a fusion feature of the preset liver ultrasound image. The fusion feature and the enhanced-detail preset liver ultrasound image are combined to form a sample. Different samples are used to form a training set. The first loss value of the deep convolutional neural network on the training set is obtained through the absolute error loss function, and the second loss value of the deep convolutional neural network on the training set is obtained through the mean squared error loss function. The first loss value and the second loss value are added to generate the total loss value of the deep convolutional neural network on the training set. The training objective is to reduce the total loss value. The deep convolutional neural network is iteratively trained until the total loss value is less than the preset loss value. The training of the deep convolutional neural network is then stopped, and the trained deep convolutional neural network is saved.
8. A device for enhancing liver ultrasound images, characterized in that, Applied to electronic devices with deployed deep convolutional neural networks, the electronic devices establish a data transmission channel with Doppler ultrasound equipment, including: The first acquisition module is used to send a data acquisition command to the Doppler ultrasound device through the data transmission channel, receive the response data returned by the Doppler ultrasound device according to the data acquisition command, and acquire pixel data and attribute data from the response data; The input module is used to input pixel data into the image rendering engine. The image rendering engine renders the pixel data to generate the current liver ultrasound image, which is the liver ultrasound image acquired by the Doppler ultrasound device at the current acquisition time. The second acquisition module is used to determine the location range of the liver region in the current liver ultrasound image, extract the local image corresponding to the liver region as the liver region image, and obtain the average value of all pixel gray values in the liver region image. The third acquisition module is used to generate a visual quality score of the current liver ultrasound image based on the average gray value of all pixels in the liver region image and a preset ultrasound image scoring model. When the visual quality score is lower than the preset score, the texture data, gradient data, and gray value data of the current liver ultrasound image are obtained from the attribute data of the current liver ultrasound image. The generation module uses the feature extractor in the trained deep convolutional neural network to extract features from the texture data, gradient data, and grayscale data of the current liver ultrasound image, respectively, generating texture features, edge features, and grayscale features of the current liver ultrasound image. These features are then concatenated to obtain a fused feature of the current liver ultrasound image. This fused feature is then input into the trained deep convolutional neural network, and the generator within the network generates a detailed enhanced version of the current liver ultrasound image.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the enhanced method as described in any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the enhanced method as described in any one of claims 1 to 7.