Medical image transmission method and system based on big data

Abstract

The invention relates to the technical field of medical image transmission methods, in particular to a medical image transmission method and system based on big data, and the method comprises the following steps: obtaining a fragment sequence; obtaining an initial correlation degree matrix based on the distribution density gradient of every two adjacent fragments; obtaining a normalized correlation degree matrix; generating an incremental weight matrix through multi-stage projection aggregation, and obtaining an incremental weight of each pair of fragments; based on the incremental weight matrix, a difference adaptive filtering method is applied to obtain a difference optimization matrix; based on the difference optimization matrix, applying a power law distribution model to generate a final distribution matrix, and obtaining the transmission priority of each fragment; and the transmission sequence of the segments is determined based on the final distribution matrix so as to realize ordered transmission of the medical image data, and regions containing key diagnosis information can be quickly distinguished based on the intensity change of data information in the segments and are endowed with higher priorities, so that important data are preferentially processed during data transmission.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image transmission methods, and more specifically, to a medical image transmission method and system based on big data. Background Art

[0002] With the advancement of medical technology and the improvement of data storage capacity, the application of medical imaging data is becoming more and more extensive, covering many aspects from daily health checks to complex disease diagnosis. Especially in the field of radiography, ultrasound, magnetic resonance imaging and other imaging fields, the large amount of data generated provides detailed information support for doctors' diagnosis. However, the characteristics of medical image data are large data volume, high density and high transmission quality requirements. How to efficiently and reliably transmit these large-scale imaging data is a major challenge faced by existing technologies.

[0003] Most existing medical image transmission methods directly rely on data compression and sequential transmission, lacking a refined distinction between the information density of different regions in the image. The usual method is to directly transmit the image after compressing it as a whole, or to divide the image data into equal segments and transmit them sequentially. However, different areas of medical images often have significant density differences. Areas in the image that contain rich tissue information are usually more valuable for diagnosis, while other background areas have lower information density and are relatively less important. The existing technology does not have an effective means to distinguish and process these areas, resulting in the inability to guarantee the efficiency and order of key data transmission, affecting the overall transmission efficiency and real-time diagnosis.

[0004] In addition, existing methods generally adopt a static transmission order when processing image transmission, lacking a dynamic priority management and scheduling mechanism. When network bandwidth or storage space is limited, there is often no clear priority division between key image data and background data, and all data is processed according to the same rules. The biggest drawback of this static transmission mechanism is that it cannot ensure the priority transmission of key data, which may result in doctors not being able to receive important imaging information in a timely manner, affecting diagnostic effectiveness and efficiency.

[0005] Traditional methods also have shortcomings in controlling the quality of data transmission. Conventional image compression transmission technology lacks dynamic adaptability to indicators such as link stability and packet loss rate during data transmission. Traditional methods lack effective remediation and optimization mechanisms for packet loss or data delay during transmission, which leads to unstable quality of medical images during transmission, which may cause loss of image quality during data reconstruction, affecting subsequent clinical applications and analysis.

[0006] In summary, the existing medical image transmission technology still has a lot of room for improvement in key data priority transmission, dynamic data scheduling, image reconstruction quality assurance, etc., and it is difficult to meet the growing demand for medical data transmission. There is an urgent need for an image transmission method that can identify the difference in information density between different areas in the image, dynamically adjust the transmission priority, and improve the image reconstruction quality. Summary of the invention

[0007] In view of the shortcomings of existing medical image transmission methods, the present invention proposes a medical image transmission method and system based on big data, which mainly solves the following technical problems: how to identify high-density information areas and give priority processing during medical image transmission to improve data transmission efficiency and image reconstruction quality; how to adapt to different network transmission conditions through dynamic scheduling and optimization algorithms to make the data transmission process more stable and reliable; how to balance data compression and the integrity of key data during transmission to ensure the accuracy and real-time performance of image transmission.

[0008] The present invention provides a medical image transmission method based on big data, comprising the following steps:

[0009] Acquire a data information value sequence of a medical image from a medical device, divide the data information value sequence into a number of segments of equal length, and obtain the segment sequence;

[0010] Based on each segment in the segment sequence, a distribution density gradient of each segment is determined, where the distribution density gradient is a rate of change of data information density within the segment, and an initial correlation matrix is ​​obtained based on the distribution density gradients of every two adjacent segments;

[0011] Performing dynamic weight normalization and offset adjustment on the initial correlation matrix to obtain a normalized correlation matrix;

[0012] Based on the normalized correlation matrix, an increasing weight matrix is ​​generated through multi-level projection aggregation to obtain an increasing weight of each pair of fragments;

[0013] Based on the incremental weight matrix, applying a difference adaptive filtering method to obtain a difference optimization matrix;

[0014] Based on the difference optimization matrix, a power law distribution model is applied to generate a final allocation matrix to obtain the transmission priority of each fragment;

[0015] The transmission order of the fragments is determined based on the final allocation matrix to achieve orderly transmission of the medical image data.

[0016] Preferably, the distribution density gradient is used to characterize the density change of the information value in each segment, wherein the distribution density gradient G of each segment is i For fragment Si The density change rate is expressed as:

[0017]

[0018] Among them G i For fragment S i The distribution density gradient, s k represents the kth data information value, and L represents the length of the fragment.

[0019] Preferably, the correlation in the initial correlation matrix is ​​obtained based on the distribution density gradient difference of adjacent segments, and the correlation is determined by the following formula:

[0020]

[0021] where r ij For fragment S i and S j The initial correlation, G i and G j Segment S i and S j The distribution density gradient of , ∈ is a small constant to prevent division by zero.

[0022] Preferably, the normalization algorithm comprises the following steps:

[0023] For the initial correlation r ij Perform maximum and minimum normalization and offset adjustment, the normalized correlation matrix element N(r ij ) is obtained by the following formula:

[0024]

[0025] Where N(r ij ) is the normalized correlation, min(r ij ) and max(r ij ) are the minimum and maximum values ​​of the correlation degree, δ is the offset constant to prevent division by zero, β is the normalized amplitude adjustment weight, and γ is the offset adjustment ratio.

[0026] Preferably, the normalized correlation matrix is ​​aggregated by multi-level projection to generate an increasing weight matrix, and the increasing weight is obtained by the following formula:

[0027]

[0028] Where W ij Represents fragment S i and S j The increasing weight, r ij For fragment S i and Sj The correlation degree, P ik Represents fragment S i With fragment S k The geometric projection of , ζ is the adjustment factor for adjusting the uniformity of the projection weight, η is the incremental weight adjustment parameter, and n is the number of projection iterations.

[0029] Preferably, each element of the difference optimization matrix is ​​obtained by the following formula:

[0030]

[0031] where F ij represents the difference optimization value after filtering, W ij For fragment S i and S j The increasing weight of , λ is the adaptive adjustment coefficient, κ is the filter gradient adjustment coefficient, and θ is the difference threshold.

[0032] Preferably, the power law distribution model is obtained by applying the following formula:

[0033]

[0034] where q ij Represents fragment S i and fragment S j The priority of the transmission, F ij is the element of the difference optimization matrix, α is the scale factor of the power law distribution, and ω is the distribution weight exponent.

[0035] Preferably, the data processing flow of the distribution density gradient calculation step, the initial correlation matrix calculation step, the dynamic weight normalization and offset adjustment step, the multi-level projection aggregation step, the difference adaptive filtering step and the power-law distribution model generation final distribution matrix step sequentially forms a closed loop to ensure the validity and priority control of the medical image transmission data stream.

[0036] A medical image transmission system based on big data comprises a processing module and a communication module, wherein the processing module is used to implement the steps, and the communication module is used to transmit the data information value sequence between the medical device and the receiving module.

[0037] Preferably, the communication module includes a wireless communication module for transmitting the data information value sequence in a wireless manner to achieve data security and privacy protection.

[0038] The invention has the following beneficial effects:

[0039] The medical image transmission method and system based on big data proposed in the present invention have achieved significant improvements in transmission efficiency and data integrity through a series of innovative algorithms and scheduling mechanisms. First, the present invention effectively identifies high information density areas in medical images through distribution density gradient calculation. The distribution density gradient algorithm is based on the density changes of data information within the fragment, and can quickly identify areas containing key diagnostic information and give them a higher priority, thereby giving priority to important data during data transmission. This algorithm effectively solves the problem of the failure to give priority to the transmission of dense image information areas in traditional methods, and greatly improves the transmission speed of important data.

[0040] In terms of transmission order management, the present invention generates an increasing weight matrix through multi-level projection aggregation, so that the transmission priority of each image fragment is accumulated layer by layer based on the density difference. Under the multi-level aggregation of increasing weights, the weights of different fragments are optimally allocated, so that high-density information fragments always have higher transmission weights, ensuring the scientificity and rationality of data transmission. This layer-by-layer accumulated weight aggregation method not only ensures the priority of important data fragments in transmission, but also effectively avoids the waste of network bandwidth resources.

[0041] Furthermore, the present invention smoothes the incremental weight matrix through differential adaptive filtering at the MAC layer to form a stable differential optimization matrix. The adaptive filtering algorithm dynamically adjusts the weight allocation according to the real-time status of data transmission, and ensures the stability protection of high-density information areas during transmission through adaptive adjustment factors and gradient control, thereby reducing fluctuations during data transmission. The introduction of this filtering mechanism not only improves the stability of image transmission, but also effectively solves the contradiction between transmission efficiency and image quality in traditional algorithms.

[0042] In terms of transmission priority control, the present invention uses the power-law distribution model to further enhance the transmission priority of high-density information fragments and form the final transmission order. The power-law distribution model creates a significant priority gap between key data fragments and ordinary data fragments, thereby ensuring the priority transmission of key image data even when bandwidth is limited in a complex network environment. The introduction of the power-law distribution model enables high-density data fragments to maintain a high transmission stability even when the network load changes, further improving the reliability of transmission.

[0043] The present invention realizes the intelligent and dynamic management of medical image data in a large data transmission environment through a series of algorithmic steps such as distribution density gradient identification, increasing weight matrix, multi-level aggregation, adaptive filtering and power-law distribution. Through close collaboration and complementarity between the steps, a complete transmission management system is formed. The efficient identification of distribution density gradients and the multi-level accumulation of increasing weights form the basis for priority sorting, the adaptive filtering balances the transmission priority and data stability, and the power-law distribution ultimately opens the data priority to ensure the priority transmission of key data. This set of collaboratively optimized transmission mechanisms not only solves the contradiction between traditional methods in transmission efficiency and image quality, but also realizes the efficient transmission of large-scale medical image data, provides technical support for clinical decision-making and remote diagnosis, and has significant application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 The present invention is a flow chart of the method.

[0045] Figure 2 It is a system logic block diagram of the present invention.

[0046] Figure 3 It is a logic block diagram of the processing module of the present invention.

[0047] Figure 4 It is a logic block diagram of the communication module of the present invention. DETAILED DESCRIPTION

[0048] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects thereof are described in detail below in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0049] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0050] Please refer to Figure 1-4 The present invention provides a medical image transmission method and system based on big data, aiming to improve the efficiency and priority management of medical images in the data transmission process, and realize the orderly and rapid transmission of medical image data in a big data environment.

[0051] The present invention discloses a medical image transmission method based on big data, which specifically includes the following steps:

[0052] First, a data information value sequence of a medical image is obtained from a medical device. The data information value sequence is image data presented in numerical form, usually composed of the grayscale or color value of each pixel. In order to facilitate subsequent data processing and compression, the data information value sequence is divided into a number of equal-length segments to form a segment sequence. Preferably, the length of the segment can be determined according to the resolution and transmission requirements of the image. For example, for high-resolution images, the segment length can be 512 information values; and for images of general resolution, the segment length can be set to 256 information values, so that the stability and consistency of data transmission can still be guaranteed under the condition of less bandwidth. This step ensures that the image data is structured and segmented, which facilitates the subsequent processing module to further analyze the data density and transmission priority.

[0053] After the fragment sequence is generated, the distribution density gradient of each fragment is further determined based on the data content of each fragment. The distribution density gradient is a measure of the rate of change of data within a fragment, which is used to reflect the density of data information within the fragment. The density gradient is estimated by the rate of change of information values ​​in each fragment. For example, in a fragment, if the difference between adjacent data information values ​​is large, indicating that the information in the fragment changes dramatically, the distribution density gradient value of the fragment is high, thereby reflecting the high information area in the image. In this way, the distribution density gradient obtained can not only provide a basis for subsequent priority sorting, but also identify areas with high information density in the image.

[0054] Next, the initial correlation matrix is ​​calculated based on the distribution density gradient of adjacent segments. The initial correlation matrix is ​​used to measure the correlation between adjacent segments to determine whether to give priority to certain highly correlated segments during image transmission.

[0055] Preferably, in one embodiment of the present invention, a step of obtaining a sequence of data information values ​​and dividing it into equal-length segments is described. The selection of the length of the segment directly affects the processing efficiency of the image data and the subsequent compression effect, and is usually adjusted according to the actual scenario. When transmitting large-size, high-resolution medical images, a longer segment length is selected, such as 512 or 1024 information values, to reduce the number of segmentations during transmission and reduce system overhead; for smaller images or when lower resolution is required, the segment length can be shortened to 128 or 256 information values. This flexible segment division method ensures efficient image data segmentation and transmission under different transmission bandwidth conditions.

[0056] Preferably, in one embodiment of the present invention, the calculation of the distribution density gradient is explained in detail. The distribution density gradient is used to quantify the data information density in each segment, and the density in the segment is measured based on the rate of change of the information value, preferably calculated by the following formula:

[0057]

[0058] Among them, G i represents the distribution density gradient of the ith segment, s k is the kth data information value, and L is the length of the segment. In practical applications, distribution density gradient calculation can help identify information-dense areas in images and assign different transmission priorities according to their density. For example, in medical imaging, areas with rich tissue details often have higher distribution density gradients, while background areas have lower gradients. Through such gradient differentiation calculations, priority transmission of key information areas in medical images can be achieved, thereby ensuring that high-density information is processed first and significantly improving the efficiency of data transmission.

[0059] Preferably, in one embodiment of the present invention, the construction of the initial correlation matrix includes the following method: the matrix is ​​used to measure the correlation between adjacent segments so that the subsequent steps can adjust the transmission order according to the correlation. The calculation method of the initial correlation is as follows:

[0060]

[0061] Among them, r ij Represents fragment S i and fragment S j The initial correlation between i and G j Segment S i and fragment S j The distribution density gradient, ∈ is a small constant to avoid zero denominator, usually taken as 0.0001. This formula reflects the correlation between fragments by calculating the relative difference of density gradients of adjacent fragments. The beneficial effect of such a correlation calculation method is that a reasonable transmission order is arranged between information-dense areas and sparse areas, so that high-information-density fragments are given priority in the transmission process. For the lesion area and background area in medical images, the lesion area is transmitted first because of its higher density gradient, thereby improving the transmission efficiency of important data.

[0062] Preferably, in one embodiment of the present invention, the dynamic weight normalization and offset adjustment of the initial correlation matrix are performed to obtain a normalized correlation matrix. The purpose of normalization is to ensure that the correlation between the fragments is evenly distributed and to avoid uneven weights due to large differences in density gradients. The normalization formula is as follows:

[0063]

[0064] Among them, N(r ij ) is the normalized correlation, min(r ij ) and max(r ij) represent the minimum and maximum values ​​of the initial correlation, δ is an offset constant to avoid a zero denominator, usually 0.01, β is the weight of the adjustment amplitude, preferably 0.5, and γ is the offset adjustment ratio, usually between 0.1 and 1.0. Through normalization and offset adjustment, it is ensured that the values ​​in the correlation matrix are distributed within a relatively balanced range, preventing extreme values ​​from affecting subsequent transmission sorting. This process not only ensures the reasonable distribution of correlation, but also effectively avoids the instability caused by excessive deviation of weights between fragments.

[0065] Through this dynamic normalization, the correlation degree is reasonably distributed among the fragments to form a balanced correlation matrix, which provides an accurate weight basis for subsequent priority sorting, thereby ensuring that the fragments with high density gradients are processed first during transmission.

[0066] Preferably, in one embodiment of the present invention, based on the normalized correlation matrix, an incremental weight matrix is ​​generated through multi-level projection aggregation to obtain the incremental weight of each pair of fragments. This step accumulates the correlation weight of each fragment layer by layer through the multi-level projection aggregation method to ensure the incrementality and reasonable distribution of the weights between fragments.

[0067] Specifically, multi-level projection aggregation is accomplished by geometric projection and iterative increment, where each element of the weight matrix W is incremented ij Obtained by the following formula:

[0068]

[0069] Among them, W ij For fragment S i and fragment S j The increasing weight of ij For fragment S i and S j The normalized correlation degree of P ik For fragment S i With fragment S k The geometric projection value is used to measure the spatial similarity between fragments; ζ is the projection balance factor, which is usually taken as 0.05 to ensure the balance between different projection weights; η is the adjustment parameter of the incremental weight, preferably taken as 0.9, so that the weight increases layer by layer; n is the number of projection iterations, which is usually taken as 3 to ensure the incremental effect of multi-layer aggregation.

[0070] The beneficial effect of multi-level projection aggregation is that by optimizing and aggregating weights layer by layer, the weights of data-intensive areas are ensured to increase continuously, thereby giving high-density fragments higher transmission priority in the transmission sorting. Through this increasing weight matrix, high-density areas can be effectively identified and sorted, and important data can be transmitted preferentially, thus improving transmission efficiency.

[0071] Preferably, in one embodiment of the present invention, based on the incremental weight matrix, a differential adaptive filtering method is applied to obtain a differential optimization matrix. The purpose of adaptive filtering is to eliminate extreme weight fluctuations in the incremental weight matrix, ensure a more balanced weight distribution between different segments, and thus enhance the smoothness and stability of the transmitted data.

[0072] Specifically, each element of the differential optimization matrix F ij It is obtained by the following adaptive filtering formula:

[0073]

[0074] Among them, F ij For fragment S i and S j The difference optimization value of ij For fragment S i and S j The increasing weight, λ is the adaptive adjustment coefficient, preferably 1.0, to adapt to most density changes; κ is the filter gradient adjustment coefficient, usually adjusted between 0.5 and 2.0, so as to smooth the change of weight; θ is the difference threshold, preferably 0.3, which is used to determine under what circumstances the weight is filtered.

[0075] Through adaptive filtering, the weight distribution in the matrix is ​​smoother and more continuous, making the weight priority of high-density data areas more significant, while the weight fluctuation of low-density areas is suppressed. The beneficial effect of this method is that the extreme differences in weights are eliminated through filtering, ensuring a more stable order of image data transmission, thereby improving the consistency of transmission effects.

[0076] Preferably, in one embodiment of the present invention, based on the difference optimization matrix, a power law distribution model is applied to generate a final allocation matrix to obtain the transmission priority of each fragment. The power law distribution model is introduced to ensure that the fragments in the data-intensive area obtain higher priority, so that these fragments are preferentially transmitted in data transmission.

[0077] Specifically, the transmission priority is determined by the following power-law distribution formula:

[0078]

[0079] Among them, q ij Represents fragment S i and fragment S j The final transmission priority of F ij is the element of the difference optimization matrix; α is the power law distribution scaling factor, preferably 1.5, which is used to increase the weight gap of high priority fragments; ω is the distribution weight exponent, which is usually set to 2.0 to ensure that high-density data fragments are transmitted first.

[0080] The advantage of this power-law distribution model is that it effectively amplifies the priority of dense information areas according to the weight differences of each fragment, ensuring that key data is given priority during data transmission and avoiding information loss due to improper fragment sorting.

[0081] In one embodiment of the present invention, a medical image transmission system based on big data is proposed, including a processing module 1 and a communication module 2, so as to realize the systematic execution of the above-mentioned method steps. The processing module 1 is mainly used to analyze, divide and sort the image data, and the communication module 2 is used to transmit the processed data information value sequence between the medical device and the receiving module.

[0082] In practical applications, the processing module 1 includes a data processing unit (not shown), which obtains the image data information sequence and performs segment division, correlation matrix calculation, dynamic normalization and multi-level projection aggregation, and finally outputs a sequence of image segments with priority marks. The communication module 2 includes a wireless communication unit for transmitting the data sequence to the receiving end.

[0083] The beneficial effect of the system design is that the processing module 1 processes the data to ensure that the transmission order of the image data is optimized, while the communication module 2 realizes fast and seamless data transmission through wireless transmission, ensuring the stability and efficiency of the data transmission process.

[0084] In one embodiment of the present invention, the wireless communication module of the communication module 2 is used for the purpose of transmitting a sequence of data information values ​​wirelessly to ensure data security and privacy protection. In this module, an encryption protocol (such as AES-256) is preferably used to encrypt the sequence of data information values ​​to ensure the security of data transmission. In addition, the wireless communication unit of the communication module 2 supports multiple transmission modes, including Wi-Fi, 5G and Bluetooth protocols, to meet the transmission requirements in different environments.

[0085] In practical applications, the wireless communication module can maintain a stable connection with the receiving end of the medical image, ensuring that the image data is not disturbed during the transmission process, thereby achieving real-time and high efficiency. The beneficial effect of the present invention is that the use of wireless communication not only improves the flexibility of data transmission, but also ensures the security of the transmission process through encryption means, providing reliable support for the remote processing and analysis of medical data.

[0086] The medical image transmission method and system based on big data of the present invention realizes the high efficiency and priority management of the data transmission process through the steps of slicing processing, correlation optimization, multi-level aggregation, adaptive filtering and power-law sorting of image data. Finally, with the help of processing module 1 and communication module 2, the system can stably complete the data transmission process, providing an efficient and reliable solution for telemedicine and medical image processing.

[0087] In order to verify the superiority of the medical image transmission method based on big data of the present invention, a specific experimental design was implemented, and a professional medical image data set was selected and a detailed test analysis was carried out through comparative experiments. The core indicators of the test revolve around the innovative points of the present invention, including the accuracy of distribution density gradient calculation, transmission priority sorting effect, and the final transmission efficiency and image reconstruction quality.

[0088] The present invention uses the public medical imaging dataset MIMIC-CXR as experimental data. The dataset contains a large amount of chest X-ray image data with high image resolution and different information densities, which is suitable for testing the transmission effect of the present invention under different data densities.

[0089] In the embodiment of the present invention, the image is first divided into equal-length segments, each segment contains 512 information values, the distribution density gradient is calculated for each segment, the initial correlation matrix is ​​established and dynamically normalized. Then, an increasing weight matrix is ​​generated through multi-level projection aggregation, and finally, adaptive filtering and power-law distribution model are used to generate priority sorting to ensure that high-density information areas are transmitted first. The final data is transmitted to the receiving end in an encrypted manner through the wireless communication module.

[0090] The comparative test uses the traditional medical image transmission method, which directly divides the image data into segments of the same length and transmits them in sequence without considering the density differences and priority sorting of the segments. This method lacks gradient analysis of data segments, correlation optimization, and transmission sequence management, resulting in the failure to effectively highlight high information density areas during transmission.

[0091] The experiment tested the following key indicators:

[0092] Transmission efficiency: Evaluates the amount of image data per unit time, in MB / s, and detects the impact of differences in data density of each segment on the transmission time during the transmission process.

[0093] Image reconstruction quality: The quality of image reconstruction at the receiving end after image transmission is completed is evaluated using the peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) indicators, expressed in dB and a value between 0 and 1, respectively.

[0094] Data packet loss rate: The amount of image information lost due to packet loss during the transmission process, expressed as a percentage.

[0095] The detection method uses the standard image transmission analysis tool Wireshark to monitor data packets, and uses MATLAB to perform PSNR and SSIM evaluation on image reconstruction at the receiving end to ensure the scientificity and accuracy of the results.

[0096] The test results and analysis are as follows:

[0097] index Embodiments of the present invention Comparative Example Transmission efficiency (MB / s) 5.6 3.2 Image reconstruction quality (PSNR) 38.7dB 32.4dB Image reconstruction quality (SSIM) 0.92 0.81 Data packet loss rate (%) 1.5 4.8

[0098] In the embodiment of the present invention, the transmission efficiency is significantly better than the comparative example, reaching 5.6MB / s, while the comparative example is only 3.2MB / s, which shows that the priority management based on distribution density gradient and ascending weight sorting of the present invention significantly improves the transmission efficiency during data transmission. By giving priority to the transmission of high-density information fragments, the timely delivery of key data is ensured, shortening the overall transmission time.

[0099] In terms of image reconstruction quality, the present invention is superior to the comparative example in both PSNR and SSIM, with PSNR reaching 38.7dB and SSIM value being 0.92, while the comparative example is 32.4dB and 0.81 respectively. This shows that the image transmission method of the present invention can more effectively preserve the details and structural information of the image and improve the accuracy of image reconstruction.

[0100] In addition, the packet loss rate of the embodiment of the present invention is 1.5%, which is much lower than 4.8% of the comparative example. The low packet loss rate reflects that during the transmission process, due to the role of priority management and adaptive filtering, the transmission process is more stable, the packet loss phenomenon is effectively controlled, and the integrity of the data is guaranteed.

[0101] The test results show that the present invention effectively optimizes the transmission process of medical images through distribution density gradient analysis, multi-level projection aggregation and power law distribution model sorting. Under different density data distributions, the transmission efficiency is greatly improved and information loss is reduced by giving priority to the transmission of important fragments. At the same time, the low packet loss rate and high PSNR / SSIM value show that the image data can maintain a high reconstruction quality after transmission, ensuring the accuracy and integrity of medical images, which is crucial for clinical decision support and telemedicine analysis.

[0102] In the best embodiment of the present invention, efficient and reliable medical image transmission is achieved through intelligent analysis and sorting of image data, providing an effective solution for data transmission in the medical field.

[0103] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A medical image transmission method based on big data, characterized in that it includes the following steps: Acquire a data information value sequence of a medical image from a medical device, divide the data information value sequence into a number of segments of equal length, and obtain the segment sequence; Based on each segment in the segment sequence, a distribution density gradient of each segment is determined, where the distribution density gradient is a rate of change of data information density within the segment, and an initial correlation matrix is ​​obtained based on the distribution density gradients of every two adjacent segments; Performing dynamic weight normalization and offset adjustment on the initial correlation matrix to obtain a normalized correlation matrix; Based on the normalized correlation matrix, an increasing weight matrix is ​​generated through multi-level projection aggregation to obtain an increasing weight of each pair of fragments; Based on the incremental weight matrix, applying a difference adaptive filtering method to obtain a difference optimization matrix; Based on the difference optimization matrix, a power law distribution model is applied to generate a final allocation matrix to obtain the transmission priority of each fragment; The transmission order of the fragments is determined based on the final allocation matrix to achieve orderly transmission of the medical image data.

2. A medical image transmission method based on big data according to claim 1, characterized in that: The distribution density gradient is used to characterize the density change of the information value in each segment, where the distribution density gradient G of each segment is i For fragment S i The density change rate is expressed as: Among them G i For fragment S i The distribution density gradient, s k represents the kth data information value, and L represents the length of the fragment.

3. A medical image transmission method based on big data according to claim 2, characterized in that: The correlation in the initial correlation matrix is ​​obtained based on the distribution density gradient difference of adjacent segments, and the correlation is determined by the following formula: where r ij For fragment S i and S j The initial correlation, G i and G j Segment S i and S j The distribution density gradient of , ∈ is a small constant to prevent division by zero.

4. The method for transmitting medical images based on big data according to claim 3, characterized in that: The normalization algorithm comprises the following steps: For the initial correlation r ij Perform maximum and minimum normalization and offset adjustment, the normalized correlation matrix element N(r ij ) is obtained by the following formula: Where N(r ij ) is the normalized correlation, min(r ij ) and max(r ij ) are the minimum and maximum values ​​of the correlation degree, δ is the offset constant to prevent division by zero, β is the normalized amplitude adjustment weight, and γ is the offset adjustment ratio.

5. A medical image transmission method based on big data according to claim 4, characterized in that: The normalized correlation matrix is ​​aggregated through multi-level projection to generate an increasing weight matrix, and the increasing weight is obtained by the following formula: Where W ij Represents fragment S i and S j The increasing weight, r ij For fragment S i and S j The correlation degree, P ik Represents fragment S i With fragment S k The geometric projection of , ζ is the adjustment factor for adjusting the uniformity of the projection weight, η is the incremental weight adjustment parameter, and n is the number of projection iterations.

6. A medical image transmission method based on big data according to claim 5, characterized in that: Each element of the difference optimization matrix is ​​obtained by the following formula: where F ij represents the difference optimization value after filtering, W ij For fragment S i and S j The increasing weight of , λ is the adaptive adjustment coefficient, K is the filter gradient adjustment coefficient, and θ is the difference threshold.

7. A medical image transmission method based on big data according to claim 6, characterized in that: The power law distribution model is obtained by applying the following formula: where q ij Represents fragment S i and fragment S j The priority of the transmission, F ij is the element of the difference optimization matrix, α is the scale factor of the power law distribution, and ω is the distribution weight exponent.

8. The method for transmitting medical images based on big data according to claim 7, characterized in that: The data processing flow of the distribution density gradient calculation step, the initial correlation matrix calculation step, the dynamic weight normalization and offset adjustment step, the multi-level projection aggregation step, the difference adaptive filtering step and the power-law distribution model generation final distribution matrix step sequentially forms a closed loop to ensure the effectiveness and priority control of the medical image transmission data stream.

9. A medical image transmission system based on big data, characterized in that: It comprises a processing module and a communication module, wherein the processing module is used to implement the steps described in any one of claims 1-8, and the communication module is used to transmit the data information value sequence between the medical device and the receiving module.

10. A medical image transmission system based on big data according to claim 9, characterized in that: The communication module includes a wireless communication module, which is used to transmit the data information value sequence in a wireless manner to achieve data security and privacy protection.

Images