Terminal medical equipment data accurate acquisition and low-memory efficient transmission method and system

Abstract

The invention relates to a terminal medical equipment data accurate acquisition and low-memory high-efficiency transmission method and system. Data attribution chaos is completely eradicated through physical medium forced identity binding and chaos disturbance verification; unified access of equipment of multiple manufacturers is realized by adopting entropy weighted projection protocol adaptation, and a collector does not need to be customized; internal memory pressure cubic self-adaptive compression, predictive dynamic caching and chaotic encryption breakpoint resume are introduced, and efficient block-level transmission and instant resource recovery are achieved under a limited internal memory; the integrity and sequence of the data are ensured through convolution verification of the receiving end, and the accuracy, compatibility, transmission efficiency and safety of data acquisition in a basic medical scene are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of medical device data processing technology, and in particular relates to a method and system for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices. Background Technology

[0002] With the accelerated development of information technology in primary healthcare, portable medical devices (such as Bluetooth blood pressure monitors, heart rate monitors, and ultrasound machines) are widely used in community health centers and telemedicine. The large volumes of medical imaging data generated by these devices (such as raw ultrasound images) need to be collected in real-time by embedded terminals and transmitted to a cloud management platform to support doctor diagnoses and electronic medical record archiving. In existing technologies, Bluetooth medical devices often connect directly to terminals via the BLE protocol, but this generally suffers from loose binding between user identity and device, leading to data confusion among multiple patients and potentially causing medical liability disputes. Furthermore, ultrasound devices from different manufacturers use proprietary protocols and heterogeneous interfaces (such as GE's DICOM extension and Mindray's MDP protocol), requiring customized acquisition units for each device, resulting in poor compatibility and high deployment costs. In addition, terminal memory resources are limited (usually less than 256MB), while ultrasound data can be hundreds of MB in size with a single acquisition. Traditional full-data buffering and transmission methods are prone to memory overflow or require full retransmission after transmission interruption, leading to low efficiency and poor reliability.

[0003] While existing block compression and breakpoint resume solutions have been attempted, their fixed compression ratio, static cache threshold, and plaintext storage of progress records make them difficult to adapt to data type fluctuations and security requirements, and they cannot achieve accurate, efficient, and secure closed-loop transmission in resource-constrained environments. Summary of the Invention

[0004] The purpose of this invention is to propose a method and system for accurate data acquisition and efficient low-memory transmission of terminal medical devices, so as to solve the technical problems of user identity mismatch, poor medical device compatibility, and difficulty in efficient transmission of large-volume data in low-memory environments in the existing data acquisition and transmission of terminal medical devices.

[0005] To achieve one of the above-mentioned objectives, one embodiment of the present invention provides a method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices, the method comprising:

[0006] The raw medical imaging data stream is divided into multiple raw data blocks. User identification and Bluetooth device media access control address are entered through physical media, binding timestamps are recorded, and chaotic perturbation binding tokens and global seeds are generated.

[0007] After the binding token is verified to be valid, the fingerprint vector of the medical imaging device is automatically extracted and the protocol parsing library is selected by entropy weighted projection, and the original data block is assigned a type label.

[0008] Based on type tags and real-time memory pressure factor cube adjustment of compression ratio, compressed data blocks are generated, and cache release threshold and prediction error compensation term are dynamically calculated through prediction model. Cache space is reclaimed when preset release conditions are met.

[0009] After the cache space is reclaimed, the progress record is encrypted using a global seed-driven chaotic stream to support block-level breakpoint resumption. The receiving end uses convolutional weighted checksum to verify the integrity and order, and then splices the compressed data blocks sequentially to restore the original medical imaging data.

[0010] As a further improvement to one embodiment of the present invention, the method further includes, in which the generation of the chaotic perturbation binding token and the global seed includes,

[0011] Calculate time drift The formula is:

[0012]

[0013] in, This is the current timestamp. To bind a timestamp;

[0014] Based on the time drift Normalization yields the initial chaotic state The formula is:

[0015]

[0016] in, Preset modulus;

[0017] Based on the initial chaotic state Perform Logistic Mapping Iteration The terminal chaotic state is obtained after a preset number of iterations. Where r is the first chaotic growth rate and m is the iteration index;

[0018] The terminal's chaotic state The binding token is calculated by concatenating and XORing the identity identifier and the Bluetooth device media access control address into binary. ;in, For binary conversion;

[0019] Extract the binding token The former bit as global seed ;in This is the seed bit width.

[0020] As a further improvement to one embodiment of the present invention, the method further includes, after the binding token is verified to be valid, automatically extracting the fingerprint vector of the medical imaging device and selecting a protocol parsing library by entropy-weighted projection, and assigning type labels to the original data block, including,

[0021] Based on the same time drift Recalculate the chaotic state at the verification end with iterative parameters ;

[0022] Based on the terminal's chaotic state Chaotic state at the verification end ,judge To confirm whether the binding is valid; among which, This is the floating-point precision threshold. Maximum allowable drift;

[0023] Extract the fingerprint vector of the medical imaging device only when the binding is valid. and with each feature vector in the protocol library Calculate cosine similarity and the corresponding protocol entropy Where i is the protocol library index.

[0024] Based on the cosine similarity With protocol entropy Select protocol identifier ;

[0025] The protocol identifier Load the corresponding protocol parsing library for the original data block Perform parsing and assign type tags Where k is the data block index.

[0026] As a further improvement to one embodiment of the present invention, the method further includes, in the step of adjusting the compression ratio based on type tags and the cube of real-time memory pressure factor to generate compressed data blocks, the following steps are taken:

[0027] Initialize used memory statistics ;

[0028] Obtain the device's total memory capacity in real time via the operating system interface. With current free memory capacity Calculate the pressure factor The formula is:

[0029]

[0030] According to the type label Applying cubic amplification compression ratio, the formula is:

[0031]

[0032] in, For type-label-based adaptive compression ratio, The preset base compression ratio corresponding to the type tag;

[0033] The adaptive compression ratio Input compression algorithm to generate compressed data blocks ;

[0034] Based on the compressed data block Size, update used memory statistics The formula is:

[0035]

[0036] in, This is the attenuation coefficient.

[0037] As a further improvement to one embodiment of the present invention, the method further includes, in which the step of dynamically calculating the cache release threshold and the prediction error compensation term through a prediction model, and reclaiming the cache space when the preset release conditions are met, includes,

[0038] Based on the previous The size of each compressed data block is used to calculate the moving average prediction. The formula is:

[0039]

[0040] in, Set the preset sliding window size;

[0041] Based on the predicted and actual moving average values, a prediction error compensation term is calculated. ;

[0042] Based on the prediction error compensation term Weighted update of the predicted cache usage for the next compressed data block The formula is:

[0043]

[0044] in, This is the gain coefficient;

[0045] Based on the predicted cache usage With current free memory Set cache release threshold ;in, The safety increment coefficient;

[0046] Based on the cache release threshold Confirmation of response from the receiving end ,judge To trigger cache release;

[0047] The release operation is performed after confirming receipt of the response, and the generation of the encryption progress record is triggered after the release.

[0048] As a further improvement to one embodiment of the present invention, the method further includes, after the cache space is reclaimed, using a global seed to drive a chaotic stream to encrypt the progress record to support block-level breakpoint resumption, including...

[0049] Based on the global seed Initialize chaotic state ;

[0050] Based on the chaotic state Iterative generation of chaotic sequences ;in, This represents the second chaotic growth rate;

[0051] Based on the chaotic sequence Generate block-level keys ;

[0052] The block-level key XOR encryption with the progress triple yields ;in, The current transmission offset for the k-th block. Cyclic redundancy check;

[0053] Record the encryption progress Store to flash address ;in, For the memory module cycle, This refers to the size of a single record storage unit.

[0054] As a further improvement to one embodiment of the present invention, the method further includes, after the receiving end verifies the integrity and order using a convolutional weighted checksum, sequentially splicing compressed data blocks to recover the original medical imaging data, including...

[0055] Based on the aforementioned compressed data blocks Calculate the check value ;

[0056] Based on the verification value With the global seed Apply convolution kernel The weighted summation formula is:

[0057]

[0058] in, Total number of data blocks This is a pre-defined ternary weighted array;

[0059] The convolution weighted checksum The value is compared and verified with the expected value pre-calculated and transmitted by the sending end;

[0060] Based on the successful verification, a deconvolution order check is performed: the block order consistency is verified using the Kernel inverse operation.

[0061] After the integrity and sequence verifications are passed, the components are assembled sequentially. to Restore the original medical imaging data.

[0062] To achieve one of the above-mentioned objectives, an embodiment of the present invention also provides a system for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices, the system comprising an identity binding module, a protocol adaptation module, a compression module, and a breakpoint resume encryption module;

[0063] The identity binding module is used to divide the raw medical imaging data stream into multiple raw data blocks, input the user identity identifier and Bluetooth device media access control address through the physical medium, record the binding timestamp, and generate a chaotic perturbation binding token and a global seed.

[0064] The protocol adaptation module is used to automatically extract the fingerprint vector of the medical imaging device and select the protocol parsing library by entropy weighted projection after the binding token is verified to be valid, and assign type labels to the original data block.

[0065] The compression module is used to adjust the compression ratio based on type label and real-time memory pressure factor cube, generate compressed data blocks, and dynamically calculate cache release threshold and prediction error compensation term through prediction model, and reclaim cache space when preset release conditions are met.

[0066] The breakpoint resume encryption module is used to encrypt the progress record with a global seed-driven chaotic stream after the cache space is reclaimed to support block-level breakpoint resume. The receiving end uses convolutional weighted checksum to verify the integrity and order, and then splices the compressed data blocks in sequence to restore the original medical imaging data.

[0067] To achieve one of the above-mentioned objectives, an embodiment of the present invention also provides an electronic device, including a memory and a processor, characterized in that the memory stores a computer program that can run on the processor, and when the program is executed on the processor, it implements the steps in the method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices as described above.

[0068] To achieve one of the above-mentioned objectives, an embodiment of the present invention also provides a storage medium storing a computer program, characterized in that, when the computer program is executed by a processor, it implements the steps in the method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices as described above.

[0069] Compared with existing technologies, this invention provides a method and system for accurate data acquisition and efficient low-memory transmission of terminal medical devices. It eliminates data ownership confusion through mandatory identity binding via physical media and chaotic perturbation verification; it adopts an entropy-weighted projection protocol to achieve unified access for devices from multiple manufacturers, eliminating the need for customized collectors; it introduces memory pressure cube adaptive compression, predictive dynamic caching, and chaotic encryption breakpoint resumption to achieve efficient block-level transmission and real-time resource reclamation under limited memory; and it uses convolutional verification at the receiving end to ensure data integrity and order, significantly improving the accuracy, compatibility, transmission efficiency, and security of data acquisition in primary healthcare scenarios. Attached Figure Description

[0070] Figure 1 This is an overall flowchart of the method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices as described in this invention.

[0071] Figure 2 This is a schematic diagram of the architecture of the terminal medical device data accurate acquisition and low-memory high-efficiency transmission system described in this invention. Detailed Implementation

[0072] The present invention will now be described in detail with reference to the specific embodiments shown in the accompanying drawings. However, these embodiments do not limit the present invention, and any structural, methodological, or functional modifications made by those skilled in the art based on these embodiments are included within the scope of protection of the present invention.

[0073] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0074] In Embodiment 1 of the present invention, the present invention provides a method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices, such as... Figure 1 As shown, the method includes,

[0075] S1: Divide the raw medical imaging data stream into multiple raw data blocks, input the user identity and Bluetooth device media access control address through the physical medium, record the binding timestamp, and generate a chaotic perturbation binding token and a global seed.

[0076] S2: After the binding token is verified to be valid, the fingerprint vector of the medical imaging device is automatically extracted and the protocol parsing library is selected by entropy weighted projection, and the original data block is assigned a type label;

[0077] S3: Based on type labels and real-time memory pressure factor cube adjustment of compression ratio, generate compressed data blocks, and dynamically calculate cache release threshold and prediction error compensation term through prediction model, and reclaim cache space when preset release conditions are met;

[0078] S4: After the cache space is reclaimed, the progress record is encrypted using a global seed-driven chaotic stream to support block-level breakpoint resumption. The receiving end uses convolutional weighted checksum to verify the integrity and order, and then splices the compressed data blocks in sequence to restore the original medical imaging data.

[0079] In one specific embodiment of the present invention, a chaotic perturbation binding token and a global seed are generated, specifically as follows:

[0080] Calculate time drift The formula is:

[0081]

[0082] in, This is the current timestamp. To bind a timestamp;

[0083] Based on the time drift Normalization yields the initial chaotic state The formula is:

[0084]

[0085] in, Preset modulus;

[0086] Based on the initial chaotic state Perform Logistic Mapping Iteration The terminal chaotic state is obtained after a preset number of iterations. Where r is the first chaotic growth rate and m is the iteration index;

[0087] The terminal's chaotic state The binding token is calculated by concatenating and XORing the identity identifier and the Bluetooth device media access control address into binary. ;in, For binary conversion;

[0088] Extract the binding token The former bit as global seed ;in This is the seed bit width.

[0089] It should be noted that generating chaotic perturbation binding tokens and global seeds aims to prevent data ownership mismatches and provide a unified cryptographic root for subsequent end-to-end security. This step involves the user registering their identity via physical media (such as an IC card or QR code). and with Bluetooth device media access control address The pairing is executed immediately after pairing to ensure that each device is associated with only a single patient's data, thus avoiding cross-contamination in multi-user scenarios.

[0090] Furthermore, calculate the time drift. The key is to capture the real-time nature of the binding operation, introduce dynamic initial conditions for subsequent chaotic disturbances, and prevent attackers from forging associations by statically replaying old binding information.

[0091] Furthermore, the initial chaotic state is obtained based on time drift normalization. The key is to transform continuous-time variables into a controllable input range of chaotic mappings, ensuring that even small time differences can produce significantly different iterative trajectories.

[0092] Furthermore, the Logistic mapping is iterated until a preset number of iterations are obtained. The key lies in amplifying minute differences in the initial state through nonlinear dynamics, generating highly unpredictable intermediate perturbation values. This iterative process is driven by the first chaotic growth rate. The driver is specifically designed for identity binding to domains.

[0093] Furthermore, calculate the binding token. The key lies in fusing user identity, device address, and chaotic perturbation into a single encrypted credential, achieving irreversible cracking through SHA-256 hashing. This token serves as a digital signature for the binding relationship and is directly output to the verification step for real-time comparison and recalculation results, preventing tampering.

[0094] Furthermore, before extracting the binding token bit as global seed The key lies in deriving a system-wide shared security root from the binding credentials. This seed runs through subsequent breakpoint resume encryption (driving chaotic stream key generation) and receiver verification (participating in convolutional weighted summation), realizing an end-to-end security closed loop from identity verification to data recovery.

[0095] Preferably, if ΔT < 0 or the data entry fails, a re-entry process is triggered until success is achieved or the maximum number of retries is reached, at which point the device is locked, thereby achieving dynamic identity binding to prevent replay attacks.

[0096] In one specific embodiment of the present invention, after the binding token is verified to be valid, the fingerprint vector of the medical imaging device is automatically extracted and a protocol parsing library is selected by entropy-weighted projection. A type label is then assigned to the original data block. Specifically,

[0097] Based on the same time drift Recalculate the chaotic state at the verification end with iterative parameters ;

[0098] Based on the terminal's chaotic state Chaotic state at the verification end ,judge To confirm whether the binding is valid; among which, This is the floating-point precision threshold. Maximum allowable drift;

[0099] Extract the fingerprint vector of the medical imaging device only when the binding is valid. and with each feature vector in the protocol library Calculate cosine similarity and the corresponding protocol entropy Where i is the protocol library index.

[0100] Based on the cosine similarity With protocol entropy Select protocol identifier ;

[0101] The protocol identifier Load the corresponding protocol parsing library for the original data block Perform parsing and assign type tags Where k is the data block index.

[0102] It should be noted that assigning type labels to raw data blocks aims to eliminate device compatibility barriers and provide precise type guidance for subsequent compression strategies, ensuring that raw data streams of different proprietary protocols are seamlessly parsed on the terminal side without the need to develop dedicated collectors for each device.

[0103] Furthermore, the generation and verification of the binding token are triggered only after the identity binding security is confirmed. The binding validity serves as the entry gate for protocol adaptation, preventing data interference from unauthorized devices from disrupting system processes. Firstly, based on the same time drift... Recalculate the chaotic state with iterative parameters The key is to reproduce the disturbance trajectory of the binding end at the verification end, providing a dynamic comparison benchmark for preventing replay attacks. This recalculation status directly serves subsequent validity judgment, ensuring the timeliness and uniqueness of the binding credential throughout the device access stage.

[0104] Furthermore, confirming the validity of the binding involves double-locking the trusted boundaries of identity and time. This judgment result serves as a conditional trigger for protocol parsing, allowing fingerprint vector extraction only after confirmation. This avoids invalid bindings consuming computing resources.

[0105] Furthermore, select the protocol identifier. And load the corresponding library for the original data block parse and assign type tags The key is to enable automatic device identification and intelligent classification of data blocks. This type label serves as the core input parameter of the compression module, directly driving the type-specific adjustment of the cubic adaptive compression ratio (such as high compression of image blocks and medium compression of waveform blocks), ensuring that the compression strategy and data characteristics are accurately matched.

[0106] Preferably, if all If the similarity is below the preset threshold, it will fall back to the default general parsing protocol, thereby achieving adaptive parsing of the device fingerprint-driven protocol.

[0107] In one specific embodiment of the present invention, compressed data blocks are generated by adjusting the compression ratio based on type tags and the cube of the real-time memory pressure factor. Specifically,

[0108] Initialize used memory statistics ;

[0109] Obtain the device's total memory capacity in real time via the operating system interface. With current free memory capacity Calculate the pressure factor The formula is:

[0110]

[0111] According to the type label Applying cubic amplification compression ratio, the formula is:

[0112]

[0113] in, For type-label-based adaptive compression ratio, The preset base compression ratio corresponding to the type tag;

[0114] The adaptive compression ratio Input compression algorithm to generate compressed data blocks ;

[0115] Based on the compressed data block Size, update used memory statistics The formula is:

[0116]

[0117] in, This is the attenuation coefficient.

[0118] It should be noted that the compression ratio adjustment based on type label and real-time memory pressure factor cube aims to dynamically amplify the compression intensity according to the data type and the current resource tension, so as to ensure that the terminal can process hundreds of MB of raw ultrasound data blocks under limited memory constraints without causing memory overflow or transmission bottleneck.

[0119] Furthermore, initialize the memory usage statistics. The goal is to establish a baseline zero point for the compression cycle, providing a clean iterative starting point for subsequent block-by-block stress assessments.

[0120] Furthermore, the pressure factor is calculated. The key lies in quantifying the scarcity of memory resources in real time, by statistically analyzing the usage of the previous block, which is one block behind. Avoid circular dependencies. This factor, acting as a driving signal for the degree of compression aggression, is directly input into the cubic amplification formula, ensuring that the more scarce the memory, the more significant the improvement in compression ratio, thus forming a resource adaptive response mechanism.

[0121] Furthermore, applying cubic amplification compression ratio based on type labels achieves dual-dimensional control by combining the inherent characteristics of the data with external pressure, and presets the base compression ratio. It provides a type-specific baseline, while the cubic term introduces non-linear amplification.

[0122] Furthermore, a compression algorithm is executed to generate compressed data blocks. The key is to transform large blocks of raw data into low-occupancy transmission units. These compressed blocks serve as the direct objects for cache prediction and breakpoint resume, and their size determines the timing of memory release and network load.

[0123] Preferably, if compression fails or the output size exceeds the original data block size, a fallback to the previous state is performed. Perform recompression;

[0124] Furthermore, update the memory usage statistics. The key is to provide accurate historical feedback for the next block of pressure calculation. The exponential decay term simulates the natural release of memory, and the actual occupancy term corrects the prediction bias. This updated output closes the current block compression loop and is fed forward to the cache prediction module to drive dynamic adjustment of the threshold; at the same time, it is fed back to the pressure factor calculation, forming an iterative closed-loop adaptive system for compression and memory management.

[0125] In one specific embodiment of the present invention, a cache release threshold and a prediction error compensation term are dynamically calculated using a prediction model, and cache space is reclaimed when preset release conditions are met. Specifically,

[0126] Based on the previous The size of each compressed data block is used to calculate the moving average prediction. The formula is:

[0127]

[0128] in, Set the preset sliding window size;

[0129] Based on the predicted and actual moving average values, a prediction error compensation term is calculated. ;

[0130] Based on the prediction error compensation term Weighted update of the predicted cache usage for the next compressed data block The formula is:

[0131]

[0132] in, This is the gain coefficient;

[0133] Based on the predicted cache usage With current free memory Set cache release threshold ;in, The safety increment coefficient;

[0134] Based on the cache release threshold Confirmation of response from the receiving end ,judge To trigger cache release;

[0135] The release operation is performed after confirming receipt of the response, and the generation of the encryption progress record is triggered after the release.

[0136] It should be noted that by dynamically calculating the cache release threshold and prediction error compensation term through the prediction model, cache space is reclaimed when the preset release conditions are met. This aims to prevent memory overflow while maximizing cache turnover rate, ensuring that the system can continuously process large-volume ultrasound data streams generated continuously under 186MB of memory, without crashing or transmission stalling due to cache accumulation.

[0137] Furthermore, calculating the moving average prediction based on the size of several compressed data blocks aims to capture the local trend of recent data block sizes and provide a reliable prior estimate for the next block. This predicted value serves as the basis for subsequent error compensation and occupancy estimation, avoiding inaccurate threshold settings due to sudden image blocks.

[0138] Furthermore, the calculation of the prediction error compensation term quantifies the current prediction deviation and corrects it positively, preventing continuously underestimated predictions from causing overly optimistic thresholds, which could ultimately lead to accumulated errors and memory overflow. This compensation term is directly injected into the next memory usage update, enabling the prediction model to learn itself.

[0139] Furthermore, by weighting and updating the predicted cache occupancy of the next compressed data block, the aim is to integrate historical predictions with the latest observations to form a smooth and responsive forecast. This predicted occupancy serves as the core basis for setting the release threshold, ensuring that the threshold always has sufficient margin to accommodate the maximum possible occupancy of the next block.

[0140] Furthermore, setting a cache release threshold establishes a safe memory usage limit for the current block. The safety increment coefficient increases linearly with the block index, further preventing data fluctuations in the later stages and ensuring increasing security as the transmission process progresses. This threshold directly determines whether the current block is eligible for release.

[0141] Preferably, if If the data is less than the preset minimum safe memory, the processing of subsequent data blocks is paused and the oldest cache is released first, thereby achieving prediction-driven low-memory cache management.

[0142] Furthermore, determining whether the release conditions are met and executing the release after confirming the receipt of the response ensures "zero-risk" memory reclamation. The cache is only truly released when the network confirms that the block has safely arrived at the receiving end, completely eliminating the risk of data loss. Immediately after the release action is completed, subsequent encrypted progress records are generated, ensuring that only blocks that have truly completed transmission and released resources are persistently recorded. This strictly synchronizes the cache state with the resumption progress, forming a hard-linked mechanism for cache reclamation and breakpoint resumption.

[0143] In one specific embodiment of the present invention, after the cache space is reclaimed, a global seed is used to drive a chaotic stream to encrypt the progress record to support block-level breakpoint resumption. Specifically,

[0144] Based on the global seed Initialize chaotic state ;

[0145] Based on the chaotic state Iterative generation of chaotic sequences ;in, This represents the second chaotic growth rate;

[0146] Based on the chaotic sequence Generate block-level keys ;

[0147] The block-level key XOR encryption with the progress triple yields ;in, The current transmission offset for the k-th block. Cyclic redundancy check;

[0148] Record the encryption progress Store to flash address ;in, For the memory module cycle, This refers to the size of a single record storage unit.

[0149] It should be noted that after the cache space is reclaimed, a global seed-driven chaotic stream is used to encrypt the progress record to support block-level breakpoint resumption. This aims to ensure that only data blocks that have been safely transmitted and successfully released from the cache are encrypted and recorded, while preventing attackers from tampering with or forging the progress record. This allows for accurate recovery to the last successful block even after network fluctuations, unexpected power outages, or device restarts, achieving truly "zero retransmission" efficient resumption. This step is strictly controlled by the preceding cache release conditions and only occurs after a confirmation of receipt. Furthermore, the cache is only triggered after it has been actually reclaimed, forming a rigid causal link with the cache module: not released = not recorded, not recorded = must be retransmitted during recovery, completely eliminating the risk of data inconsistency caused by "false success".

[0150] Furthermore, based on the global seed initialization of the chaotic state, the global seed Ψ generated in the identity binding stage is directly inherited to the transmission and resume stage, realizing end-to-end encrypted lineage from user identity to transmission progress, ensuring that even if an attacker obtains the device, they cannot generate a legitimate progress record without knowing Ψ.

[0151] Furthermore, the use of iteratively generating chaotic sequences is to produce independent and unpredictable key streams for each block, resulting in a second chaotic growth rate. Identity binding stage Separation creates a dual-domain, dual-rate security isolation, preventing leakage in one stage from affecting another.

[0152] Furthermore, the purpose of generating a block-level key and XORing the progress triple (block number, offset, CRC32) is to convert the transmission progress into high-strength ciphertext, where... As a data fingerprint, the progress record is strongly bound to the actual data block content; any data tampering or offset forgery will cause verification failure during recovery. This encrypted progress record serves as the sole trusted credential for recovery after a power outage, directly determining from which block the system continues transmission.

[0153] Furthermore, storing the encrypted progress record to a designated address in the flash memory enables unlimited resume downloads from breakpoints within a very small storage space, employing a modular approach. The circular overwrite strategy ensures that the flash memory is never filled up.

[0154] Preferably, after the device is powered on or the network is restored, the system reinitializes the chaotic state in the same manner using the stored global seed Ψ. Reconstruct the keystream in block number order All valid encryption progress records in flash memory Decrypting each element one by one yields the plaintext triple (k, (CRC32). The system finds the largest and consecutive completed block number from it. (i.e., from 1 to) If all records are valid and their CRC32 matches the local cache or the received block, then proceed directly from the first... The block continues to be transmitted. If a block fails to decrypt or has a CRC32 mismatch, it is considered that the block and all subsequent blocks are incomplete, and transmission resumes from the block following the last consecutive successful block. If the flash memory is wiped or Ψ is lost, a conservative rollback to the first block is performed for retransmission. This recovery process ensures that even with multiple power outages, the last truly successful block can always be accurately located, achieving completely zero-redundancy retransmission.

[0155] In one specific embodiment of the present invention, after verifying the integrity and order using a convolutional weighted checksum, the receiving end sequentially splices and compresses data blocks to recover the original medical imaging data. Specifically, this involves...

[0156] Based on the aforementioned compressed data blocks Calculate the check value ;

[0157] Based on the verification value With the global seed Apply convolution kernel The weighted summation formula is:

[0158]

[0159] in, Total number of data blocks This is a pre-defined ternary weighted array;

[0160] The convolution weighted checksum The value is compared and verified with the expected value pre-calculated and transmitted by the sending end;

[0161] Based on the successful verification, a deconvolution order check is performed: the block order consistency is verified using the Kernel inverse operation.

[0162] After the integrity and sequence verifications are passed, the components are assembled sequentially. to Restore the original medical imaging data.

[0163] It should be noted that the purpose of splicing compressed data blocks to restore the original medical imaging data is to achieve triple verification of the integrity, anti-misorder, and anti-tampering of all compressed data blocks at one time, ensuring that the ultrasound image finally seen by the doctor is completely consistent with the original data collected by the terminal, and eliminating the risk of misdiagnosis due to transmission errors.

[0164] Furthermore, calculating the checksum of each compressed data block involves creating an independent content fingerprint for each block, which serves as the raw material for subsequent convolutional verification. This fingerprint is completely consistent with the CRC32 in the progress record at the sending end, providing content-level verifiability for the recovery phase.

[0165] Furthermore, weighted summation using convolutional kernels is applied and incorporated into the global seed. The block index k is used to simultaneously achieve dual detection of integrity and order: the ternary periodic weighting of the convolutional kernel can amplify the deviation of misordered blocks, the linear superposition of the block index k further amplifies the positional anomaly, and the global seed... The introduction of this feature deeply integrates the verification result with the patient's identity-binding token; any forgery, substitution, or missing piece will result in a mismatch in the total checksum. As the final digital signature in the entire transmission process, it directly inherits the signature generated during the identity binding stage. This enables end-to-end tamper-proof protection from patient card swiping to image presentation.

[0166] Furthermore, when transmitting the first compressed data block, or in a dedicated control message after all blocks have been transmitted, the sender will use the exact same global seed Ψ and the pre-calculated expected convolutional weighted checksum of the convolutional kernel Kernel. The data is sent to the receiving end along with the data block. The receiving end caches this expected value until all data blocks have arrived, at which point a final comparison is performed. If the verification fails, the receiving end requests a retransmission of the corresponding data block and its preceding and following adjacent blocks.

[0167] Furthermore, the stitching and restoration process only proceeds after verifying the total checksum is equal and performing a deconvolution order check. This minimizes the cost of verification failures: large memory is allocated for stitching only if the global checksum passes, avoiding wasted resources on invalid operations; the deconvolution order check serves as a secondary safeguard, further eliminating the very few out-of-order cases masked by convolution. After stitching, the system immediately restores the original medical imaging data that is completely consistent with the data acquired by the terminal, allowing doctors to make real-time diagnoses and archive the data. The restored original medical imaging data is then output to the medical imaging display module or the diagnostic AI engine, thereby achieving an end-to-end closed-loop diagnostic application.

[0168] Preferably, the expected checksum can be pre-calculated and transmitted via a secure channel (such as with the first data block or via a separate control channel) before the sending end completes the transmission of all data blocks or during the transmission of the first block. The value is sent to the receiving end; alternatively, the sending end may resend the expected value after all block transmissions are complete. The receiving end receives... This is then cached until all blocks are complete before a final comparison. The expected value is determined by the sender using the exact same global seed Ψ, the same convolutional kernel, and the same block checksum. The sequence is calculated to ensure that the sending and receiving ends will always get the same result under the same patient identity, thus realizing identity-based verification and binding.

[0169] Preferably, the specific method for deconvolution order checking is as follows: the receiving end maintains a sliding window of length 3, and checks the received... The same convolutional kernel is applied in reverse to the sequence for deconvolution. If the entire sequence is completely ordered and without any missing parts, the deconvolution result should be an approximately zero sequence (leaving only noise-level residuals). If there are out-of-order or missing parts, significant peaks will appear at the corresponding positions. Detecting these peaks by setting an empirical threshold can then confirm the absolute correctness of the sequence. This method does not require an additional sequence number field and can be implemented using only existing checksums.

[0170] In a second embodiment of the present invention, the present invention provides a system for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices, such as... Figure 2 As shown, the system includes an identity binding module 1, a protocol adaptation module 2, a compression module 3, and a breakpoint resume encryption module 4;

[0171] The identity binding module 1 is used to divide the medical imaging raw data stream into multiple raw data blocks, input the user identity identifier and Bluetooth device media access control address through the physical medium, record the binding timestamp, and generate a chaotic disturbance binding token and a global seed.

[0172] The protocol adaptation module 2 is used to automatically extract the fingerprint vector of the medical imaging device and select the protocol parsing library by entropy weighted projection after the binding token is verified to be valid, and assign type labels to the original data block.

[0173] The compression module 3 is used to adjust the compression ratio based on type label and real-time memory pressure factor cube, generate compressed data blocks, and dynamically calculate cache release threshold and prediction error compensation term through prediction model, and reclaim cache space when preset release conditions are met.

[0174] The breakpoint resume encryption module 4 is used to encrypt the progress record with a global seed-driven chaotic stream after the cache space is reclaimed to support block-level breakpoint resume. After the receiving end verifies the integrity and order by using convolutional weighted checksum, it sequentially splices the compressed data blocks to restore the original medical imaging data.

[0175] In a third embodiment of the present invention, the present invention provides an electronic device, including a memory and a processor, characterized in that the memory stores a computer program that can run on the processor, and when the program is executed on the processor, it implements the steps in the method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices as described above.

[0176] In a fourth embodiment of the present invention, the present invention provides a storage medium storing a computer program, characterized in that, when the computer program is executed by a processor, it implements the steps in the method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices as described above.

[0177] In summary, this invention provides a method and system for accurate data acquisition and efficient low-memory transmission from terminal medical devices. It eliminates data ownership confusion through mandatory identity binding via physical media and chaotic perturbation verification; it adopts an entropy-weighted projection protocol to enable unified access for devices from multiple manufacturers, eliminating the need for customized collectors; it introduces memory pressure cube adaptive compression, predictive dynamic caching, and chaotic encryption breakpoint resumption to achieve efficient block-level transmission and real-time resource reclamation within limited memory; and it employs convolutional verification at the receiving end to ensure data integrity and order, significantly improving the accuracy, compatibility, transmission efficiency, and security of data acquisition in primary healthcare scenarios.

[0178] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the modules described above can be referred to the corresponding process in the aforementioned method implementation, and will not be repeated here.

[0179] The modules described as separate components may or may not be physically separate. Similarly, the components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0180] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated module can be implemented in hardware or in a combination of hardware and software functional modules.

[0181] The integrated modules implemented as software functional modules described above can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer system (which may be a personal computer, server, or network system, etc.) or processor to execute some steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0182] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices, characterized in that: include, The raw medical imaging data stream is divided into multiple raw data blocks. User identification and Bluetooth device media access control address are entered through physical media, binding timestamps are recorded, and chaotic perturbation binding tokens and global seeds are generated. After the binding token is verified to be valid, the fingerprint vector of the medical imaging device is automatically extracted and the protocol parsing library is selected by entropy weighted projection, and the original data block is assigned a type label. Based on type tags and real-time memory pressure factor cube adjustment of compression ratio, compressed data blocks are generated, and cache release threshold and prediction error compensation term are dynamically calculated through prediction model. Cache space is reclaimed when preset release conditions are met. After the cache space is reclaimed, the progress record is encrypted using a global seed-driven chaotic stream to support block-level breakpoint resumption. The receiving end uses convolutional weighted checksum to verify the integrity and order, and then splices the compressed data blocks sequentially to restore the original medical imaging data.

2. The method of claim 1, wherein the method further comprises: The generation of chaotic perturbation binding tokens and global seeds includes... Calculate time drift The formula is: ； in, This is the current timestamp. To bind a timestamp; Based on the time drift Normalization yields the initial chaotic state The formula is: ； in, Preset modulus; Based on the initial chaotic state Perform Logistic Mapping Iteration The terminal chaos is obtained after a preset number of iterations. Where r is the first chaotic growth rate and m is the iteration index; The terminal's chaotic state The binding token is calculated by concatenating and XORing the identity identifier and the Bluetooth device media access control address into binary. ;in, For binary conversion; Extract the binding token The former bit as global seed ;in This is the seed bit width.

3. The method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices according to claim 2, characterized in that: After the binding token is verified to be valid, the process of automatically extracting the fingerprint vector of the medical imaging device and selecting a protocol parsing library by entropy-weighted projection, and assigning type labels to the original data blocks includes... Based on the same time drift Recalculate the chaotic state at the verification end with iterative parameters ; Based on the terminal's chaotic state Chaotic state at the verification end ,judge To confirm whether the binding is valid; among which, This is the floating-point precision threshold. Maximum allowable drift; Extract the fingerprint vector of the medical imaging device only when the binding is valid. and with each feature vector in the protocol library Calculate cosine similarity and the corresponding protocol entropy Where i is the protocol library index. Based on the cosine similarity With protocol entropy Select protocol identifier ; The protocol identifier Load the corresponding protocol parsing library for the original data block Perform parsing and assign type tags Where k is the data block index.

4. The method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices according to claim 3, characterized in that: The method of generating compressed data blocks by adjusting the compression ratio based on type labels and the cubic adjustment of real-time memory pressure factors includes: Initialize used memory statistics ; Obtain the device's total memory capacity in real time via the operating system interface. With current free memory capacity Calculate the pressure factor The formula is: ； According to the type label Applying cubic amplification compression ratio, the formula is: ； in, For type-label-based adaptive compression ratio, The preset base compression ratio corresponding to the type tag; The adaptive compression ratio Input compression algorithm to generate compressed data blocks ; Based on the compressed data block Size, update used memory statistics The formula is: ； in, This is the attenuation coefficient.

5. The method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices according to claim 4, characterized in that: The step of dynamically calculating the cache release threshold and prediction error compensation term through a prediction model, and reclaiming cache space when preset release conditions are met, includes: Based on the previous The size of each compressed data block is used to calculate the moving average prediction. The formula is: ； in, Set the preset sliding window size; Based on the predicted and actual moving average values, a prediction error compensation term is calculated. ; Based on the prediction error compensation term Weighted update of the predicted cache usage for the next compressed data block The formula is: ； in, This is the gain coefficient; Based on the predicted cache usage With current free memory Set cache release threshold ;in, The safety increment coefficient; Based on the cache release threshold Confirmation of response from the receiving end ,judge To trigger cache release; The release operation is performed after confirming receipt of the response, and the generation of the encryption progress record is triggered after the release.

6. The method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices according to claim 5, characterized in that: The step of using a global seed to drive chaotic stream encryption of progress records after cache space reclamation to support block-level breakpoint resumption includes... Based on the global seed Initialize chaotic state ; Based on the chaotic state Iterative generation of chaotic sequences ;in, This represents the second chaotic growth rate; Based on the chaotic sequence Generate block-level keys ; The block-level key XOR encryption with the progress triple yields ;in, The current transmission offset for the k-th block. Cyclic redundancy check; Record the encryption progress Store to flash address ;in, For the memory module cycle, This refers to the size of a single record storage unit.

7. The method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices according to claim 6, characterized in that: The receiving end uses convolutional weighted checksum to verify the integrity and sequence of the data, and then sequentially splices and compresses the data blocks to restore the original medical imaging data, including: Based on the aforementioned compressed data blocks Calculate the check value ; Based on the verification value With the global seed Apply convolution kernel The weighted summation formula is: ； in, Total number of data blocks This is a pre-defined ternary weighted array; The convolution weighted checksum The value is compared and verified with the expected value pre-calculated and transmitted by the sending end; Based on the successful verification, a deconvolution order check is performed: the block order consistency is verified using the Kernel inverse operation. After the integrity and sequence verifications are passed, the components are assembled sequentially. to Restore the original medical imaging data.

8. A system for precise data acquisition and low-memory, high-efficiency transmission of terminal medical devices, characterized in that: It includes an identity binding module, a protocol adaptation module, a compression module, and a breakpoint resume encryption module; The identity binding module is used to divide the raw medical imaging data stream into multiple raw data blocks, input the user identity identifier and Bluetooth device media access control address through the physical medium, record the binding timestamp, and generate a chaotic disturbance binding token and a global seed. The protocol adaptation module is used to automatically extract the fingerprint vector of the medical imaging device and select the protocol parsing library by entropy weighted projection after the binding token is verified to be valid, and assign type labels to the original data block. The compression module is used to adjust the compression ratio based on type label and real-time memory pressure factor cube, generate compressed data blocks, and dynamically calculate cache release threshold and prediction error compensation term through prediction model, and reclaim cache space when preset release conditions are met. The breakpoint resume encryption module is used to encrypt the progress record with a global seed-driven chaotic stream after the cache space is reclaimed to support block-level breakpoint resume. The receiving end uses convolutional weighted checksum to verify the integrity and order, and then splices the compressed data blocks in sequence to restore the original medical imaging data.

9. An electronic device, comprising a memory and a processor, characterized in that: The memory stores a computer program that can run on the processor, and when the program is executed on the processor, it implements the steps of the method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices as described in any one of claims 1-7.

10. A storage medium storing a computer program, characterized in that: When the computer program is executed by the processor, it implements the steps in the method for accurate data acquisition and low-memory, high-efficiency transmission of terminal medical devices as described in any one of claims 1-7.

Images