Remote Ultrasound Image Transmission System

Abstract

The present invention relates to the field of image transmission technology, specifically to a remote ultrasound image transmission system, which includes an ultrasound image feature extraction module, a bandwidth dynamic adaptation module, a real-time annotation information encoding module, an HDR image processing module, a network status monitoring module, an image and annotation synchronous transmission module, and a client decoding and display module. In the present invention, ultrasound images are analyzed by a deep learning algorithm, the network bandwidth is dynamically adapted, a differential coding algorithm is used to optimize the transmission of annotation information, local contrast enhancement and HDR processing are used to improve the image visual effect, the synchronous transmission of images and annotations is guaranteed, the quality and efficiency of remote medical image transmission are improved, the precise processing of diagnosis-related key areas ensures the accurate display of important details, improves the accuracy of diagnosis, and the dynamic adjustment strategy of network bandwidth effectively responds to network fluctuations, ensures the stability of transmission, and significantly improves the accessibility and quality of medical services.

Description

Technical Field

[0001] The present invention relates to the field of image transmission technology, and in particular to a remote ultrasonic image transmission system. Background Art

[0002] The field of image transmission technology aims to effectively transmit image and video data from one place to another. Key technologies include image compression, encoding, transmission, decoding, and final display or storage. Currently, image transmission technology is essential for a variety of applications, such as online video streaming services, remote monitoring, remote medical diagnosis, and personal communications. With the development of cloud computing and Internet of Things technologies, image transmission technology is also evolving to meet a wider range of application needs.

[0003] Among them, the remote ultrasound image transmission system is a technical application specific to the medical field. Its main purpose is to achieve remote acquisition, transmission and reception of ultrasound image data, so that medical professionals can view and analyze ultrasound images in real time at different locations, which greatly promotes the development of telemedicine and diagnosis, and significantly improves the accessibility of medical services. It also provides a platform for communication and collaboration for medical professionals, allowing them to share image data and conduct remote consultations, thereby providing patients with more accurate and comprehensive diagnostic opinions. It aims to ensure the real-time and high-quality of image data through efficient image transmission technology, thereby achieving the purpose of improving the efficiency and quality of medical services.

[0004] Existing technologies do not fully highlight key diagnostic information when processing high-quality medical images, which may cause key details to be lost during compression and transmission. In addition, fixed network bandwidth occupancy and failure to dynamically adjust to real-time network status make image transmission vulnerable when network conditions are poor, reducing the real-time and reliability of remote diagnosis, limiting the efficiency and coverage of remote medical services, especially in the context of complex network environments or limited resources, affecting the quality of medical services. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose a remote ultrasonic image transmission system.

[0006] In order to achieve the above-mentioned object, the present invention adopts the following technical scheme: the remote ultrasound image transmission system includes an ultrasound image feature extraction module, a bandwidth dynamic adaptation module, a real-time annotation information encoding module, an HDR image processing module, a network status monitoring module, an image and annotation synchronous transmission module and a client decoding and display module, the ultrasound image feature extraction module uses a deep learning algorithm to analyze ultrasound images, extract key features, including texture complexity and edge information, and generate feature extraction results;

[0007] The bandwidth dynamic adaptation module evaluates the network bandwidth demand of the current image based on the feature extraction result, compares it with the real-time network status, and dynamically adjusts the transmission strategy, including adjusting the transmission route, to obtain the bandwidth adaptation strategy;

[0008] The real-time annotation information encoding module uses a differential encoding algorithm to encode the annotation and labeling information of ultrasound images, optimize the storage and transmission of annotation data, reduce the bandwidth occupied by annotation information, and maintain the integrity and readability of the annotation content to generate encoded annotation information;

[0009] The HDR image processing module reveals the key areas of the image based on the feature extraction results, adopts a local contrast enhancement algorithm, applies high dynamic range processing to the key areas, improves the visual effect and highlights the diagnosis-related details, and obtains the HDR-processed image;

[0010] The network status monitoring module combines the bandwidth adaptation strategy to monitor the transmission network in real time, including network bandwidth and delay, and generates network status records;

[0011] The image and annotation synchronous transmission module adopts a synchronous transmission strategy based on the HDR processed image, the encoded annotation information and the network status record to ensure the synchronization and integrity of the image and annotation data and generate a synchronous transmission data packet;

[0012] The client decoding and display module receives the synchronously transmitted data packets, decodes and processes them, and accurately displays the HDR images and annotation information on the terminal device to obtain the remote ultrasound transmission images.

[0013] As a further solution of the present invention, the feature extraction result includes a texture complexity score, an edge clarity index and a key shape identifier, the bandwidth adaptation strategy specifically includes a selected network path, an adjusted data transmission priority and the amount of bandwidth allocated to each transmission task, the encoded annotation information includes a differential change descriptor, a compressed annotation symbol set and a reduced representation of the annotation text, the HDR-processed image includes a dynamic range extension value, a local detail enhancement index and a color depth adjustment record, the network status record includes a real-time network delay index, the bandwidth consumption of the current active connection and a network quality score, the synchronously transmitted data packet specifically includes an image data block of time synchronization information, an annotation data sequence and a transmission sequence number, and the remote ultrasound transmission image specifically includes a decoded and restored transmission image and a reconstructed annotation content layout.

[0014] As a further solution of the present invention, the ultrasound image feature extraction module includes an image analysis submodule, a feature scoring submodule, and a feature synthesis submodule;

[0015] The image analysis submodule uses a deep learning algorithm to scan the ultrasound image layer by layer, identify the texture complexity and edge information in the image, and generate preliminary feature analysis results;

[0016] The feature scoring submodule quantitatively scores texture complexity and edge information based on the preliminary feature analysis results to obtain feature scoring results;

[0017] The feature integration submodule integrates all scores in the feature scoring results, evaluates the importance of texture complexity and edge clarity, and generates feature extraction results.

[0018] As a further solution of the present invention, the bandwidth dynamic adaptation module includes an image demand assessment submodule, a network status analysis submodule, and a transmission strategy adjustment submodule;

[0019] The image demand assessment submodule assesses the network bandwidth demand of the current ultrasound image based on the feature extraction result, and generates an image bandwidth demand assessment result by combining the influence of texture complexity and edge information on bandwidth consumption;

[0020] The network status analysis submodule analyzes the capacity and stability of the current network based on the image bandwidth demand assessment result to obtain a network status analysis result;

[0021] The transmission strategy adjustment submodule dynamically adjusts the transmission strategy of ultrasound images in combination with the network status analysis results, including selecting the optimal transmission route and adjusting the data priority during transmission, matching the current network status, ensuring the efficiency and quality of image data transmission, and generating a bandwidth adaptation strategy.

[0022] As a further solution of the present invention, the real-time annotation information encoding module includes an annotation content extraction submodule, a differential encoding processing submodule, and an annotation information compression submodule;

[0023] The annotation content extraction submodule extracts the annotation and annotation information to be encoded from the ultrasound image, including text descriptions and graphic marks, ensures the integrity and accuracy of the annotation content, and generates extracted annotation content;

[0024] The differential coding processing submodule applies a differential coding algorithm to the extracted annotation content, records the changes between consecutive annotations, reduces redundant information, improves coding efficiency, and obtains a differential coding processing result;

[0025] The differential encoding algorithm is based on the formula:

[0026] ;

[0027] Calculate the difference between the continuous annotation contents to obtain the differential encoding processing result, where: Indicates The difference between the content of the annotation and the previous annotation, It is the first The value of the annotation content, Indicates The value of the annotation content, is the adjustment coefficient, adjust The degree of influence Indicates Additional attribute values ​​of the annotation content, is the adjustment coefficient, adjust The degree of influence Indicates -1 additional attribute value of the comment content, is the adjustment coefficient, adjust The degree of influence Representative The conditional coefficient of the annotation content, is the global adjustment factor, is a constant;

[0028] The annotation information compression submodule generates encoded annotation information by applying symbol compression technology to optimize storage format and transmission size based on the differential encoding processing result, reducing the bandwidth occupied by the annotation information while maintaining the readability of the content.

[0029] As a further solution of the present invention, the HDR image processing module includes a key area identification submodule, a local contrast enhancement submodule, and an HDR rendering submodule;

[0030] The key area identification submodule analyzes the ultrasound image based on the feature extraction results, identifies key areas related to diagnosis, including abnormal structures of tumors and cysts, and generates key area identification records;

[0031] The local contrast enhancement submodule processes the key areas based on the key area identification records using a local contrast enhancement algorithm to optimize their visibility and diagnostic accuracy in the image, thereby obtaining a local enhanced image result;

[0032] The local contrast enhancement algorithm is based on the formula:

[0033] ;

[0034] Calculate the enhanced pixel value and generate a local enhanced image result with high diagnostic value, where: For the Row and The new pixel value of the pixel in the column after local contrast enhancement processing, represents the global enhancement coefficient, Represents the cumulative distribution function value of the current pixel in its local window, represents the minimum value of the cumulative distribution function within the local window, represents the maximum value of the cumulative distribution function within the local window, represents the local detail enhancement factor, Representation and Pixels , Related local detail enhancement, represents the target area weight coefficient, represents the weight distribution of the target area in the local window, represents the global saturation adjustment factor, Represents the saturation of the overall image. represents the local brightness adjustment factor, Represents the average brightness of the local window.

[0035] The HDR rendering submodule performs high dynamic range processing on the local enhanced image results, optimizes the overall visual effect of the image, reveals diagnosis-related details, and generates an HDR-processed image.

[0036] As a further solution of the present invention, the network status monitoring module includes a bandwidth monitoring submodule, a delay measurement submodule, and a network status comprehensive evaluation submodule;

[0037] The bandwidth monitoring submodule monitors the current bandwidth usage of the transmission network in real time based on the bandwidth adaptation strategy, including the total bandwidth, the used bandwidth and the remaining available bandwidth, and generates a bandwidth usage record;

[0038] The delay measurement submodule measures the delay of network transmission, including round-trip time and data packet transmission delay, to obtain a network delay measurement result;

[0039] The network status evaluation submodule combines the bandwidth usage record and the network delay measurement result to evaluate the overall status of the current network, including network stability and transmission efficiency, and generates a network status record.

[0040] As a further solution of the present invention, the image and annotation synchronous transmission module includes a data packaging submodule, a synchronization strategy customization submodule, and a transmission efficiency optimization submodule;

[0041] The data packaging submodule performs data packaging based on the HDR processed image and the encoded annotation information, ensuring that each data packet includes complete image data and corresponding annotation information, and generates preliminary packaged data;

[0042] The synchronization strategy customization submodule designs a synchronization transmission strategy based on the initially packaged data and network status records, including the sending timing of the data packet, the retransmission mechanism and the data integrity verification, to ensure the synchronization and integrity of the data during the network transmission process, and obtain a synchronization transmission strategy solution;

[0043] The transmission efficiency optimization submodule applies a synchronous transmission strategy scheme, adjusts transmission parameters, including window size and data packet segmentation method, optimizes data transmission efficiency in the network, and generates synchronously transmitted data packets.

[0044] As a further solution of the present invention, the client decoding and display module includes a data decoding submodule, an image rendering submodule, and an annotation display submodule;

[0045] The data decoding submodule receives the synchronously transmitted data packets, performs preliminary decoding processing, including decoding of HDR image data and annotation information, and generates decoded image data and decoded annotation data;

[0046] The image rendering submodule performs high-quality rendering processing on the decoded image data, adjusts the brightness, contrast and color balance, optimizes the display effect, and obtains rendered image display data;

[0047] The annotation display submodule combines the decoded annotation data with the rendered image display data to ensure that the annotation content is accurately aligned with the corresponding image part, and generates a remote ultrasound transmission image.

[0048] Compared with the prior art, the advantages and positive effects of the present invention are:

[0049] In the present invention, ultrasound images are analyzed by deep learning algorithms, network bandwidth is dynamically adapted, differential coding algorithms are used to optimize annotation information transmission, local contrast enhancement and HDR processing are used to improve image visual effects, synchronous transmission of images and annotations is guaranteed, the quality and efficiency of remote medical image transmission are improved, precise processing of diagnosis-related key areas ensures accurate display of important details, and improves diagnostic accuracy. The dynamic adjustment strategy of network bandwidth effectively responds to network fluctuations and ensures transmission stability, making remote diagnosis support more accurate and rapid, and significantly improving the accessibility and quality of medical services. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a system flow chart of the present invention;

[0051] Figure 2 It is a schematic diagram of the system framework of the present invention;

[0052] Figure 3 This is a flow chart of the ultrasonic image feature extraction module of the present invention;

[0053] Figure 4 This is a flow chart of the bandwidth dynamic adaptation module of the present invention;

[0054] Figure 5 It is a flow chart of the real-time annotation information encoding module of the present invention;

[0055] Figure 6 This is a flow chart of the HDR image processing module of the present invention;

[0056] Figure 7 It is a flow chart of the network status monitoring module of the present invention;

[0057] Figure 8 It is a flow chart of the image and annotation synchronous transmission module of the present invention;

[0058] Fig. 9 This is a flow chart of the client decoding and display module of the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0060] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0061] Example 1: Please refer to Figure 1 and Figure 2 ,The remote ultrasound image transmission system includes an ultrasound image feature extraction module, ,bandwidth dynamic adaptation module, real-time annotation information encoding module, ,HDR image processing module, network status monitoring module, image and annotation ,synchronization transmission module and client decoding and display module. ,The ultrasound image feature extraction module uses deep learning ,algorithms to analyze ultrasound images, extract key features, including texture complexity and edge information, and generate ,feature extraction results;

[0062] The bandwidth dynamic adaptation module evaluates the network bandwidth demand of the current image based on the feature extraction results, compares it with the real-time network status, and dynamically adjusts the transmission strategy, including adjusting the transmission route, to obtain the bandwidth adaptation strategy;

[0063] The real-time annotation information encoding module uses a differential encoding algorithm to encode the annotation and labeling information of ultrasound images, optimize the storage and transmission of annotation data, reduce the bandwidth occupied by annotation information, and maintain the integrity and readability of the annotation content to generate encoded annotation information;

[0064] The HDR image processing module reveals the key areas of the image based on the feature extraction results, uses a local contrast enhancement algorithm, applies high dynamic range processing to the key areas, improves the visual effect and highlights the diagnosis-related details, and obtains the HDR-processed image;

[0065] The network status monitoring module combines bandwidth adaptation strategies to monitor the transmission network in real time, including network bandwidth and latency, and generates network status records;

[0066] The image and annotation synchronous transmission module adopts a synchronous transmission strategy based on the HDR-processed image, the encoded annotation information and the network status record to ensure the synchronization and integrity of the image and annotation data and generate synchronous transmission data packets;

[0067] The client decoding and display module receives the synchronously transmitted data packets, decodes and processes them, and accurately displays the HDR images and annotation information on the terminal device to obtain the remote ultrasound transmission images.

[0068] The feature extraction results include texture complexity score, edge clarity index and key shape identifier. The bandwidth adaptation strategy specifically includes the selected network path, the adjusted data transmission priority and the amount of bandwidth allocated to each transmission task. The encoded annotation information includes differential change descriptor, compressed annotation symbol set and reduced representation of annotation text. The HDR-processed image includes dynamic range extension value, local detail enhancement index and color depth adjustment record. The network status record includes real-time network delay index, bandwidth consumption of the current active connection and network quality score. The synchronously transmitted data packet specifically includes image data block of time synchronization information, annotation data sequence and transmission sequence number. The remote ultrasound transmission image specifically includes the decoded and restored transmission image and the reconstructed annotation content layout.

[0069] See also Figure 2 and Figure 3 ,The ultrasound image feature extraction module includes an image analysis submodule, a feature scoring submodule, and a feature synthesis submodule;

[0070] The image analysis submodule uses a deep learning algorithm to scan ultrasound images layer by layer, identify the texture complexity and edge information in the image, and generate preliminary feature analysis results;

[0071] In the image analysis submodule, based on ultrasound image data, a convolutional neural network (CNN) was constructed using a deep learning algorithm and the PyTorch framework. The network structure included five convolutional layers (Conv2D) to identify texture and edge information in the image. Each convolutional layer was connected to a ReLU activation function to increase the nonlinearity of the network. Each convolutional layer was followed by a maximum pooling layer (MaxPooling2D) to reduce the feature dimension and retain the most important features. Finally, a fully connected layer (Dense) was used to output the texture complexity and edge clarity scores of each region. The learning rate was set to 0.001, and the mean square error (MSE) was selected as the loss function. The Adam optimizer was used to update the network parameters and generate preliminary feature analysis results.

[0072] The feature scoring submodule quantitatively scores the texture complexity and edge information based on the preliminary feature analysis results to obtain the feature scoring results;

[0073] In the feature scoring submodule, based on the preliminary feature analysis results, a scoring algorithm is adopted, specifically a linear weighted model. Weights are set for texture complexity and edge clarity respectively. The texture complexity weight is 0.6, and the edge clarity weight is 0.4. The comprehensive feature score of each ultrasound image area is calculated by weighted sum, and the scoring results are standardized in the range of 0 to 100. The NumPy library is used to perform array operations and normalization to generate feature scoring results.

[0074] The feature integration submodule integrates all scores in the feature scoring results, evaluates the importance of texture complexity and edge clarity, and generates feature extraction results;

[0075] In the feature integration submodule, the feature scoring results are integrated, and the decision tree algorithm is adopted. The model is built using the scikit-learn library to comprehensively evaluate the scores of texture complexity and edge clarity. The depth of the decision tree is set to 5. The purpose is to rank the importance of each area of ​​the ultrasound image according to its feature score and identify key imaging areas in diagnosis, such as tumors or other lesion areas. This process not only considers a single feature, but reveals the overall picture of the image through comprehensive scoring and generates feature extraction results.

[0076] See also Figure 2 and Figure 4 ,The bandwidth dynamic adaptation module includes an image demand assessment submodule, a network status analysis submodule, and a transmission strategy adjustment submodule;

[0077] The image demand assessment submodule evaluates the network bandwidth demand of the current ultrasound image based on the feature extraction results, and generates the image bandwidth demand assessment result by combining the influence of texture complexity and edge information on bandwidth consumption;

[0078] In the image demand assessment submodule, the ultrasound imaging network bandwidth demand assessment is performed based on the feature extraction results. The bandwidth demand assessment script written in Python is used to estimate the required bandwidth by calculating the information entropy of the image, taking advantage of the impact of texture complexity and edge information on bandwidth consumption. Images with high information entropy require higher bandwidth to ensure transmission quality. The information entropy calculation formula is set in the script, and the probability distribution of image pixel values ​​is used for calculation. The expected bandwidth demand value is adjusted according to the result to generate the image bandwidth demand assessment result.

[0079] The network status analysis submodule analyzes the capacity and stability of the current network based on the image bandwidth demand assessment results to obtain the network status analysis results;

[0080] In the network status analysis submodule, based on the image bandwidth demand assessment results, network monitoring tools such as Wireshark are used to analyze the capacity and stability of the current network. By capturing network data packets, the average transmission rate and packet loss rate are calculated. Wireshark sets filtering conditions to monitor only data packets related to ultrasound image transmission, analyzes the timestamp and size of the data packets, and automatically calculates the actual bandwidth usage and delay time under the current network conditions to obtain the network status analysis results.

[0081] The transmission strategy adjustment submodule dynamically adjusts the transmission strategy of ultrasound images based on the results of network status analysis, including selecting the optimal transmission route and adjusting the data priority during transmission, matching the current network status, ensuring the efficiency and quality of image data transmission, and generating a bandwidth adaptation strategy;

[0082] In the transmission strategy adjustment submodule, an adaptive flow control algorithm is used to dynamically adjust the transmission strategy of ultrasound images in combination with the results of network status analysis. The algorithm is written in a Python script and adapts to network changes by modifying the TCP window size and adjusting the queue management strategy. The script sets the logic for dynamic adjustment based on real-time network delay and packet loss rate to ensure that the sending rate is reduced when the network status deteriorates to reduce packet loss and delay, match the current network status, optimize the data transmission path, select the route with the smallest delay, and generate a bandwidth adaptation strategy.

[0083] See also Figure 2 and Figure 5 ,The real-time annotation information encoding module includes an annotation content extraction submodule, a differential encoding processing submodule, and an annotation information compression submodule;

[0084] The annotation content extraction submodule extracts the annotation and annotation information to be encoded from the ultrasound image, including text descriptions and graphic marks, ensures the integrity and accuracy of the annotation content, and generates extracted annotation content;

[0085] In the annotation content extraction submodule, Python scripts are used to extract text and graphic marks, the OpenCV library is used to process ultrasound images, the position of annotation information is located through image recognition technology, and the TesseractOCR library is used to perform optical character recognition on the text in the image. When extracting graphic marks, the Canny edge detection algorithm is used to identify the graphic contour. The parameters include the low threshold and the high threshold are set to 100 and 200 respectively to ensure the accurate extraction of graphic marks and generate the extracted annotation content.

[0086] The differential coding processing submodule applies the differential coding algorithm to the extracted annotation content, records the changes between consecutive annotations, reduces redundant information, improves coding efficiency, and obtains the differential coding processing result;

[0087] Differential encoding algorithm, according to the formula:

[0088] ;

[0089] Calculate the difference between the continuous annotation contents to obtain the differential encoding processing result, where: It means that after the improved algorithm processing, The difference between the content of the annotation and the previous annotation, It is the first The value of the annotation content, Indicates The value of the annotation content, is the adjustment coefficient, used to adjust The degree of influence of (the additional attribute value of the current annotation content), Indicates Additional attribute values ​​of the annotation content, which can be the length, complexity or measurement of specific features of the annotation content. is the adjustment coefficient, used to adjust The degree of influence of (the additional attribute value of the previous annotation content), Indicates Additional attribute values ​​of the annotation content, is the adjustment coefficient, used to adjust The degree of influence of (the conditional coefficient of the current annotation content), Representative The conditional coefficient of the annotation content can be a measure of the environmental or contextual factors based on the annotation variation. is a global adjustment factor used to ensure the calculation stability of the final difference value. is a constant that provides a baseline adjustment to ensure that the difference value does not deviate from a reasonable range due to over-adjustment.

[0090] The execution process is as follows:

[0091] Attribute and condition analysis: For each annotation content, analyze and determine its additional attributes ( , ) and the conditional coefficient ( ), a deep understanding of the annotation content, including its structure, length, complexity and other characteristics, as well as the context in which the annotation was generated;

[0092] Parameter adjustment: set adjustment coefficients based on the analysis results of attributes and conditions , , and global regulatory factors , usually based on empirical data and preliminary experiments, to find the optimal encoding effect;

[0093] Calculate the difference value, apply the improved differential encoding formula, comprehensively consider the basic difference between the continuous annotation content, the influence of additional attributes, and the conditional coefficient of the annotation content, and calculate the refined difference value ;

[0094] Difference encoding, using the calculated difference value Encoding is performed to form a differential encoding processing result.

[0095] By introducing , , , The improved formula can not only capture the changes between annotation contents in a more detailed manner, but also dynamically adjust the differential calculation according to the characteristics of the annotation content and contextual conditions, thereby improving the flexibility and accuracy of differential encoding.

[0096] The annotation information compression submodule uses symbol compression technology to optimize the storage format and transmission size based on the differential encoding processing results, reduce the bandwidth occupied by the annotation information, and maintain the readability of the content to generate the encoded annotation information;

[0097] In the annotation information compression submodule, based on the differential coding processing results, symbol compression technology is applied to further optimize the annotation information. This technology is implemented in Python and uses the Huffman coding algorithm to perform symbol-level compression on the differentially encoded data. The shortest code is assigned to the most common differences, and longer codes are assigned to rare differences. The Huffman tree is generated by building a priority queue. During the tree construction process, the node selection operation uses a priority queue to ensure that the two nodes with the lowest frequency are merged each time. This aims to reduce the storage size and bandwidth occupancy of the final annotation information while maintaining the integrity and readability of the original content to generate the encoded annotation information.

[0098] See also Figure 2 and Figure 6,The HDR image processing module includes a key area recognition submodule, a local contrast enhancement submodule, and an HDR rendering submodule;

[0099] The key area identification submodule analyzes the ultrasound image based on the feature extraction results, identifies key areas related to diagnosis, including abnormal structures of tumors and cysts, and generates key area identification records;

[0100] In the key area recognition submodule, an image segmentation algorithm is used based on the feature extraction results, specifically the U-Net model, which is implemented using the PyTorch framework. The model automatically identifies key diagnosis-related areas, such as abnormal tumor or cyst structures, by learning the texture complexity and edge information in ultrasound images. The U-Net model structure includes symmetrical contraction paths and expansion paths to capture contextual information of the background and target. The cross-entropy loss function is used to optimize the model. During training, the batch size is set to 32, the learning rate is adjusted to 0.0001, and the Adam optimizer is used to generate key area recognition records.

[0101] The local contrast enhancement submodule uses a local contrast enhancement algorithm to process the key areas based on the key area recognition records, optimizes their visibility and diagnostic accuracy in the image, and obtains the local enhanced image results;

[0102] Local contrast enhancement algorithm, according to the formula:

[0103] ;

[0104] Calculate the enhanced pixel value and generate a local enhanced image result with high diagnostic value, where: For the Row and The new pixel value of the pixel in the column after local contrast enhancement processing, Represents the global enhancement coefficient, which controls the level of overall contrast enhancement. Represents the cumulative distribution function value of the current pixel in its local window, represents the minimum value of the cumulative distribution function within the local window, represents the maximum value of the cumulative distribution function within the local window, Represents the local detail enhancement factor, which is used to adjust the degree of detail enhancement. Representation and Pixels , Related local detail enhancement, Represents the target area weight coefficient, which is used to enhance the visibility of key areas. represents the weight distribution of the target area in the local window, Represents the global saturation adjustment factor, which is used to adjust the color balance of the overall image. Represents the saturation of the overall image. Represents the local brightness adjustment factor, which is used to adjust the enhancement degree according to the local brightness. Represents the average brightness of the local window.

[0105] The execution process is as follows:

[0106] Key area and attribute analysis: using advanced image analysis techniques, such as deep learning models, to identify key areas in images and analyze regional features (such as texture and brightness);

[0107] Parameter determination: According to the characteristics of the key area and the attributes of the overall image, the parameter values ​​introduced in the improved formula are determined ( , , , ) to ensure precise enhancement of key areas while maintaining the natural feel of the overall image;

[0108] Local window processing: for each pixel in the image, calculate the cumulative distribution function of the window according to the local window in which it is located, and use the improved formula to calculate the new value of each pixel according to the parameters determined in step 2. ;

[0109] Enhancement application,Through the above steps, the pixels in each local window are processed to obtain the enhanced image.,This process not only improves the contrast and visibility of key areas, but also ensures the naturalness and detail retention of the enhancement processing through careful parameter adjustment.

[0110] By improving the algorithm, key areas with complex textures and different brightness conditions can be processed more effectively, especially in fields such as medical image analysis, where refined and personalized processing is very critical, greatly improving the accuracy and efficiency of diagnosis.

[0111] The HDR rendering submodule processes the local enhanced image results into a high dynamic range, optimizes the overall visual effect of the image, reveals diagnosis-related details, and generates HDR-processed images;

[0112] In the HDR rendering submodule, the local enhanced image results are processed with a high dynamic range in the OpenEXR format through HDR rendering technology, specifically implemented using Python scripts combined with OpenCV and Pillow libraries. The processing includes dynamic range compression and color management to retain details and colors that cannot be displayed on standard dynamic range displays. The parameters adjusted during the HDR rendering process include color space conversion to CIE Lab and nonlinear stretching of the L channel to simulate the human eye's perception of brightness. This step optimizes the overall visual effect, reveals diagnosis-related details, and generates HDR-processed images.

[0113] See also Figure 2 and Figure 7 ,The network status monitoring module includes a bandwidth monitoring submodule, a delay measurement submodule, and a network status comprehensive evaluation submodule;

[0114] The bandwidth monitoring submodule monitors the current bandwidth usage of the transmission network in real time based on the bandwidth adaptation strategy, including the total bandwidth, used bandwidth and remaining available bandwidth, and generates bandwidth usage records;

[0115] In the bandwidth monitoring submodule, based on the bandwidth adaptation strategy, a network monitoring tool, specifically a NetFlow analyzer, is used to monitor the current bandwidth usage of the transmission network and collect traffic data on network devices through the SNMP protocol. The NetFlow analyzer configuration includes setting the traffic sampling rate to 1 / 1000 and the export time interval to 1 minute. Such settings are designed to accurately monitor the total bandwidth, used bandwidth, and remaining available bandwidth in real time. The script includes traffic aggregation and filtering logic to analyze specific ultrasound image transmission traffic and generate bandwidth usage records.

[0116] The delay measurement submodule measures the delay of network transmission, including round-trip time and data packet transmission delay, and obtains the network delay measurement result;

[0117] In the delay measurement submodule, the delay of network transmission is measured using the Ping tool and the Traceroute command. The Ping tool sets the target address to the IP address of the ultrasound image data receiving end, sends an ICMP echo request message, and records the round-trip time (RTT). The Traceroute command is used to track the path nodes that the data packet passes through and the delay of each node. The parameter setting includes a maximum TTL (time to live) value of 30 hops to ensure the integrity of the tracking path, thereby accurately measuring the network transmission delay, including the overall round-trip time and the transmission delay of each node, and obtaining the network delay measurement result.

[0118] The network status evaluation submodule combines the bandwidth usage records and network delay measurement results to evaluate the overall status of the current network, including network stability and transmission efficiency, and generates a network status record;

[0119] In the network status assessment submodule, a network status assessment algorithm is used in combination with bandwidth usage records and network delay measurement results. The algorithm is implemented as a Python script. The bandwidth utilization and delay time are comprehensively considered in the algorithm. The stability of the network is determined by setting thresholds. Bandwidth utilization exceeding 80% or average delay exceeding 100ms is considered to be poor network status. Conditional judgment and numerical analysis methods, such as standard deviation and mean calculation, are used in the script to evaluate the overall status of the network, including stability and transmission efficiency, and generate network status records.

[0120] See also Figure 2 and Figure 8 ,The image and annotation synchronous transmission module includes a data packaging submodule, a synchronization strategy customization submodule, and a transmission efficiency optimization submodule;

[0121] The data packaging submodule packages data based on the HDR-processed images and the encoded annotation information, ensuring that each data packet includes complete image data and corresponding annotation information, and generates preliminary packaged data;

[0122] In the data packaging submodule, based on the HDR-processed images and encoded annotation information, a custom data encapsulation protocol is used to perform data packaging operations using a script written in Python. The script uses the json format to organize data packets, assigns a unique identifier (UUID) to each data packet, and stores the image data and annotation information in the form of key-value pairs to ensure the consistency of the data packet structure. At the same time, the zlib library is used to compress the data packet, and the compression level is set to optimal compression to reduce the bandwidth requirement during transmission and generate preliminary packaged data.

[0123] The synchronization strategy customization submodule designs the synchronization transmission strategy based on the preliminary packaged data and network status records, including the sending timing of the data packet, the retransmission mechanism and the data integrity verification, to ensure the synchronization and integrity of the data during the network transmission process, and obtain the synchronization transmission strategy solution;

[0124] In the synchronization strategy customization submodule, a synchronization transmission strategy is designed based on the preliminary packaged data and network status records. The TCP protocol is used to ensure the reliability of data transmission, and a customized retransmission mechanism is implemented. This mechanism is triggered in the Python script by detecting the data packet confirmation response. A timeout of 2 seconds is set for each data packet. If no confirmation response is received within the timeout, the data packet is resent. In addition, data integrity verification is introduced to generate a hash value for each data packet through the SHA-256 hash function, which is used by the receiving end to verify the integrity of the data packet and generate a synchronization transmission strategy solution.

[0125] The transmission efficiency optimization submodule applies the synchronous transmission strategy scheme, adjusts the transmission parameters, including the window size and the packet segmentation method, optimizes the data transmission efficiency in the network, and generates synchronously transmitted packets;

[0126] In the transmission efficiency optimization submodule, a synchronous transmission strategy is applied to adjust transmission parameters, including dynamically adjusting the TCP window size and selecting the optimal packet segmentation size according to the network status. The adjustment logic of the window size and packet segmentation method is updated in real time through network quality feedback. If the network delay increases or the packet loss rate rises, the window size is reduced to reduce network congestion. The selection of the packet segmentation size is dynamically adjusted based on the current network bandwidth and delay, aiming to optimize the data transmission efficiency in the network, reduce transmission delay, improve synchronization, and generate synchronously transmitted data packets.

[0127] See also Figure 2 and Fig. 9 ,The client decoding and display module includes a data decoding submodule, an image rendering submodule, and an annotation display submodule;

[0128] The data decoding submodule receives the synchronously transmitted data packets and performs preliminary decoding processing, including decoding of HDR image data and annotation information, and generates decoded image data and decoded annotation data;

[0129] In the data decoding submodule, synchronously transmitted data packets are received, and a Python-based decoding library, specifically PyAV, is used to perform preliminary decoding processing on the data packets. The PyAV library encapsulates FFmpeg and supports a wide range of video and audio formats. During the decoding process, the HDR image data and annotation information are decoded separately. The HDR image data adopts the H.265 encoding format, and the annotation information adopts the UTF-8 encoding format to ensure the compatibility and integrity of the data, and generate decoded image data and decoded annotation data.

[0130] The image rendering submodule performs high-quality rendering processing on the decoded image data, adjusts the brightness, contrast and color balance, optimizes the display effect, and obtains the rendered image display data;

[0131] In the image rendering submodule, the decoded image data is rendered with high quality. The OpenGL graphics library is used to write a custom shader program using GLSL (OpenGL Shading Language). The program adjusts the brightness, contrast, and color balance of the image. The brightness adjustment parameter is set to 1.2 and the contrast adjustment parameter is set to 1.5 in the shader. The color balance is adjusted by color temperature and is set to 6500K to match the daylight color temperature, thereby optimizing the display effect, improving the visual quality of the image, and obtaining the rendered image display data.

[0132] The annotation display submodule combines the decoded annotation data with the rendered image display data to ensure that the annotation content is accurately aligned with the corresponding image part and generate a remote ultrasound transmission image;

[0133] In the annotation display submodule, the decoded annotation data is combined with the rendered image display data. HTML5 and CSS3 technologies are used to dynamically overlay annotation information on the Web page through JavaScript scripts. The Canvas API is used in the JavaScript script to draw images and annotations to ensure that the annotation content is accurately aligned with the corresponding image part on the page. The annotation information is displayed in a semi-transparent text box so as not to block the key details of the image, ensuring the readability of the annotation content and synchronization with the image, and generating remote ultrasound transmission images.

[0134] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A remote ultrasonic image transmission system, characterized in that: The remote ultrasonic image transmission system comprises: The ultrasound image feature extraction module uses deep learning algorithms to analyze ultrasound images, extract key features, including texture complexity and edge information, and generate feature extraction results; The bandwidth dynamic adaptation module evaluates the network bandwidth demand of the current image based on the feature extraction result, compares it with the real-time network status, and dynamically adjusts the transmission strategy, including adjusting the transmission route, to obtain the bandwidth adaptation strategy; The real-time annotation information encoding module uses a differential encoding algorithm to encode the annotation and labeling information of ultrasound images, optimize the storage and transmission of annotation data, reduce the bandwidth occupied by annotation information, and maintain the integrity and readability of the annotation content to generate encoded annotation information; The HDR image processing module reveals the key areas of the image based on the feature extraction results, adopts a local contrast enhancement algorithm, applies high dynamic range processing to the key areas, improves the visual effect and highlights the diagnosis-related details, and obtains the HDR-processed image; The network status monitoring module combines the bandwidth adaptation strategy to monitor the transmission network in real time, including network bandwidth and delay, and generates network status records; The image and annotation synchronous transmission module adopts a synchronous transmission strategy based on the HDR processed image, the encoded annotation information and the network status record to ensure the synchronization and integrity of the image and annotation data and generate a synchronous transmission data packet; The client decoding and display module receives the synchronously transmitted data packets, decodes and processes them, and accurately displays the HDR images and annotation information on the terminal device to obtain the remote ultrasound transmission images; The bandwidth dynamic adaptation module includes an image demand assessment submodule, a network status analysis submodule, and a transmission strategy adjustment submodule; The image demand assessment submodule assesses the network bandwidth demand of the current ultrasound image based on the feature extraction result, and generates an image bandwidth demand assessment result by combining the influence of texture complexity and edge information on bandwidth consumption; The network status analysis submodule analyzes the capacity and stability of the current network based on the image bandwidth demand assessment result to obtain a network status analysis result; The transmission strategy adjustment submodule dynamically adjusts the transmission strategy of ultrasound images in combination with the network status analysis results, including selecting the optimal transmission route and adjusting the data priority during transmission, matching the current network status, ensuring the efficiency and quality of image data transmission, and generating a bandwidth adaptation strategy; The real-time annotation information encoding module includes an annotation content extraction submodule, a differential encoding processing submodule, and an annotation information compression submodule; The annotation content extraction submodule extracts the annotation and annotation information to be encoded from the ultrasound image, including text descriptions and graphic marks, ensures the integrity and accuracy of the annotation content, and generates extracted annotation content; The differential coding processing submodule applies a differential coding algorithm to the extracted annotation content, records the changes between consecutive annotations, reduces redundant information, improves coding efficiency, and obtains a differential coding processing result; The differential encoding algorithm is based on the formula: ； Calculate the difference between the continuous annotation contents to obtain the differential encoding processing result, where: Indicates The difference between the content of the annotation and the previous annotation, It is the first The value of the annotation content, Indicates The value of the annotation content, is the adjustment coefficient, adjust The degree of influence, Indicates Additional attribute values ​​of the annotation content, is the adjustment coefficient, adjust The degree of influence, Indicates Additional attribute values ​​of the annotation content, is the adjustment coefficient, adjust The degree of influence, Representative The conditional coefficient of the annotation content, is the global adjustment factor, is a constant; The annotation information compression submodule generates encoded annotation information by applying symbol compression technology to optimize storage format and transmission size based on the differential encoding processing result, reducing the bandwidth occupied by the annotation information while maintaining the readability of the content.

2. The remote ultrasonic image transmission system according to claim 1, characterized in that: The feature extraction result includes a texture complexity score, an edge clarity index and a key shape identifier; the bandwidth adaptation strategy specifically includes a selected network path, an adjusted data transmission priority and the amount of bandwidth allocated to each transmission task; the encoded annotation information includes a differential change descriptor, a compressed annotation symbol set and a reduced representation of the annotation text; the HDR-processed image includes a dynamic range extension value, a local detail enhancement index and a color depth adjustment record; the network status record includes a real-time network delay index, the bandwidth consumption of the current active connection and a network quality score; the synchronously transmitted data packet specifically includes an image data block of time synchronization information, an annotation data sequence and a transmission sequence number; the remote ultrasound transmission image specifically includes a decoded and restored transmission image and a reconstructed annotation content layout.

3. The remote ultrasonic image transmission system according to claim 1, characterized in that: The ultrasound image feature extraction module includes an image analysis submodule, a feature scoring submodule, and a feature synthesis submodule; The image analysis submodule uses a deep learning algorithm to scan the ultrasound image layer by layer, identify the texture complexity and edge information in the image, and generate preliminary feature analysis results; The feature scoring submodule quantitatively scores texture complexity and edge information based on the preliminary feature analysis results to obtain feature scoring results; The feature integration submodule integrates all scores in the feature scoring results, evaluates the importance of texture complexity and edge clarity, and generates feature extraction results.

4. The remote ultrasonic image transmission system according to claim 1, characterized in that: The HDR image processing module includes a key area identification submodule, a local contrast enhancement submodule, and an HDR rendering submodule; The key area identification submodule analyzes the ultrasound image based on the feature extraction results, identifies key areas related to diagnosis, including abnormal structures of tumors and cysts, and generates key area identification records; The local contrast enhancement submodule processes the key areas based on the key area identification records using a local contrast enhancement algorithm to optimize their visibility and diagnostic accuracy in the image, thereby obtaining a local enhanced image result; The local contrast enhancement algorithm is based on the formula: ； Calculate the enhanced pixel value and generate a local enhanced image result with high diagnostic value, where: For the Row and The new pixel value of the pixel in the column after local contrast enhancement processing, represents the global enhancement coefficient, Represents the cumulative distribution function value of the current pixel in its local window, represents the minimum value of the cumulative distribution function within the local window, represents the maximum value of the cumulative distribution function within the local window, represents the local detail enhancement factor, Representation and Pixels , Related local detail enhancement, represents the target area weight coefficient, represents the weight distribution of the target area in the local window, represents the global saturation adjustment factor, Represents the saturation of the overall image. represents the local brightness adjustment factor, Represents the average brightness of the local window; The HDR rendering submodule performs high dynamic range processing on the local enhanced image results, optimizes the overall visual effect of the image, reveals diagnosis-related details, and generates an HDR-processed image.

5. The remote ultrasonic image transmission system according to claim 1, characterized in that: The network status monitoring module includes a bandwidth monitoring submodule, a delay measurement submodule, and a network status comprehensive evaluation submodule; The bandwidth monitoring submodule monitors the current bandwidth usage of the transmission network in real time based on the bandwidth adaptation strategy, including the total bandwidth, the used bandwidth and the remaining available bandwidth, and generates a bandwidth usage record; The delay measurement submodule measures the delay of network transmission, including round-trip time and data packet transmission delay, to obtain a network delay measurement result; The network status comprehensive evaluation submodule combines the bandwidth usage record and the network delay measurement result to evaluate the overall status of the current network, including network stability and transmission efficiency, and generates a network status record.

6. The remote ultrasonic image transmission system according to claim 1, characterized in that: The image and annotation synchronous transmission module includes a data packaging submodule, a synchronization strategy customization submodule, and a transmission efficiency optimization submodule; The data packaging submodule performs data packaging based on the HDR processed image and the encoded annotation information, ensuring that each data packet includes complete image data and corresponding annotation information, and generates preliminary packaged data; The synchronization strategy customization submodule designs a synchronization transmission strategy based on the initially packaged data and network status records, including the sending timing of the data packet, the retransmission mechanism and the data integrity verification, to ensure the synchronization and integrity of the data during the network transmission process, and obtain a synchronization transmission strategy solution; The transmission efficiency optimization submodule applies a synchronous transmission strategy scheme, adjusts transmission parameters, including window size and data packet segmentation method, optimizes data transmission efficiency in the network, and generates synchronously transmitted data packets.

7. The remote ultrasonic image transmission system according to claim 1, characterized in that: The client decoding and display module includes a data decoding submodule, an image rendering submodule, and an annotation display submodule; The data decoding submodule receives the synchronously transmitted data packets, performs preliminary decoding processing, including decoding of HDR image data and annotation information, and generates decoded image data and decoded annotation data; The image rendering submodule performs high-quality rendering processing on the decoded image data, adjusts the brightness, contrast and color balance, optimizes the display effect, and obtains rendered image display data; The annotation display submodule combines the decoded annotation data with the rendered image display data to ensure that the annotation content is accurately aligned with the corresponding image part, and generates a remote ultrasound transmission image.

Images