A method and system for detecting digital imaging film and imaging inspection report

Abstract

The present invention belongs to the field of data detection technology and discloses a method and system for detecting digital image films and image inspection reports. The method comprises the following steps: a front-end server uses a data compression model to compress the real-time digital image film and the real-time image inspection report; the front-end server encrypts and uploads the mutually related real-time digital image film compressed package and the real-time image inspection report compressed package to a cloud data center; the cloud data center uses a data quality detection model to perform data quality detection on the real-time digital image film compressed package and the real-time image inspection report compressed package; the cloud data center performs integrity detection; the cloud data center performs correlation detection; and the cloud data center uses a detection report generation model to generate a detection report to obtain a real-time detection report. The present invention solves the problems of low manual processing efficiency and poor detection accuracy in the prior art.

Description

Technical Field

[0001] The present invention belongs to the technical field of data detection, and in particular relates to a detection method and system for digital image films and image inspection reports. Background Art

[0002] Digital imaging films and imaging examination reports have become digitized and mobile in the current medical field. Digital imaging films (cloud films) are a new type of medical digital imaging service based on mobile internet and cloud storage technology. Unlike traditional plastic films, cloud films do not require physical printing. Patients can access the original imaging data and diagnostic reports at any time through smart devices such as smartphones and tablets. This technology not only provides convenience for patients but also improves the efficiency and accuracy of doctors' diagnoses. An imaging examination report is a professional report compiled by a radiologist or radiologist based on the patient's imaging results. It typically includes a description of the image, a diagnostic opinion, and possible recommendations. In the modern medical system, digital imaging films and imaging examination reports are often used in combination to provide patients with comprehensive and convenient medical services. Through electronic platforms, patients can easily access their imaging data and examination reports, which is of great significance to improving the quality of medical services and patient satisfaction.

[0003] The existing detection technology of digital imaging films and imaging inspection reports has the following defects:

[0004] 1) Inefficient manual processing: In existing technologies, the processing of digital imaging films and imaging examination reports often relies on manual operations, which is not only time-consuming but also prone to errors, especially when processing large amounts of data;

[0005] 2) Poor detection accuracy: In the existing technology, the detection method of digital imaging films and imaging examination reports usually adopts simple file name parsing. This detection method is very simple and cannot guarantee the integrity and legitimacy of the data during transmission, resulting in poor detection accuracy. Summary of the Invention

[0006] In order to solve the problems of low manual processing efficiency and poor detection accuracy in the prior art, the present invention aims to provide a method and system for detecting digital imaging films and imaging inspection reports.

[0007] The technical solution adopted in the present invention is:

[0008] A method for detecting digital imaging films and imaging inspection reports comprises the following steps:

[0009] The front-end server collects the user's real-time digital image film and real-time image inspection report, and uses a pre-built data compression model to compress the real-time digital image film and real-time image inspection report to obtain a mutually related real-time digital image film compressed package and real-time image inspection report compressed package;

[0010] The front-end server encrypts and uploads the related real-time digital image film compression package and the real-time image inspection report compression package to the cloud data center;

[0011] The cloud data center uses a pre-built data quality detection model to perform data quality detection on the real-time digital image film compression package and the real-time image inspection report compression package to obtain the real-time data quality detection result. If the real-time data quality detection result is qualified, it proceeds to the next step;

[0012] The cloud data center performs integrity testing on the real-time digital image film compression package and the real-time image inspection report compression package to obtain the real-time integrity test result. If the real-time integrity test result is qualified, the process proceeds to the next step.

[0013] The cloud data center performs a correlation test on the real-time digital image film compression package and the real-time image inspection report compression package to obtain a real-time correlation test result. If the real-time correlation test result is qualified, the process proceeds to the next step.

[0014] The cloud data center generates a test report based on the real-time data quality test results, the real-time integrity test results and / or the real-time correlation test results using a pre-built test report generation model to obtain a real-time test report.

[0015] Furthermore, the data compression model is constructed based on the CNN-LSTM-DBN algorithm, and the data compression model includes an image feature extraction module constructed based on the CNN algorithm, a sequence feature extraction module constructed based on the LSTM algorithm, and a data compression module constructed based on the DBN algorithm.

[0016] Furthermore, the front-end server collects the user's real-time digital image film and real-time image inspection report, and uses a pre-built data compression model to compress the real-time digital image film and real-time image inspection report to obtain a mutually related real-time digital image film compressed package and real-time image inspection report compressed package, including the following steps:

[0017] The front-end server collects the user's real-time digital image film and real-time image inspection report, and performs pre-processing to obtain the pre-processed real-time digital image film and the pre-processed real-time image inspection report;

[0018] Using the image feature extraction module of the pre-built data compression model, extracting the real-time image features of the pre-processed real-time digital image film;

[0019] According to the real-time image features, the data compression module of the data compression model is used to compress the data to obtain a real-time digital image film compression package;

[0020] The image feature extraction module of the data compression model is used to extract the real-time sequence features of the real-time image inspection report after preprocessing;

[0021] According to the real-time sequence characteristics, the data compression module of the data compression model is used to compress the data to obtain a compressed package of the real-time image inspection report after preprocessing;

[0022] Extracting first compressed package information of the real-time digital image film compressed package and second compressed package information of the real-time image inspection report compressed package;

[0023] The first compressed package information and the second compressed package information are associated to obtain real-time associated information, as well as mutually associated real-time digital image film compressed packages and real-time image inspection report compressed packages.

[0024] Furthermore, the front-end server encrypts and uploads the mutually related real-time digital image film compressed package and the real-time image inspection report compressed package to the cloud data center, including the following steps:

[0025] The front-end server generates a first file hash value of the real-time digital image film compression package and a second file hash value of the real-time image inspection report compression package;

[0026] Use a random encryption algorithm to generate a random encryption key, and generate an encryption seed based on the random encryption key;

[0027] Encrypting the real-time digital image film compressed package and the real-time image inspection report compressed package according to the encryption seed to obtain mutually related encrypted real-time digital image film compressed package and encrypted real-time image inspection report compressed package;

[0028] The real-time association information, the encrypted seed, the first file hash value, the second file hash value, the encrypted real-time digital image film compressed package, and the encrypted real-time image inspection report compressed package are uploaded to the cloud data center.

[0029] Furthermore, the data quality detection model is constructed based on the cGAN-MLP algorithm, and the data quality detection model includes a conditional information embedder and a conditional information processor both constructed based on the MLP algorithm, and a generator and a discriminator both constructed based on the RNN algorithm. The generator is connected to the conditional information embedder and the discriminator respectively, and the discriminator is connected to the conditional information processor.

[0030] Furthermore, the cloud data center uses a pre-built data quality detection model to perform data quality detection on the real-time digital image film compression package and the real-time image inspection report compression package to obtain a real-time data quality detection result. If the real-time data quality detection result is qualified, the next step is entered, which includes the following steps:

[0031] The cloud data center decrypts the encrypted real-time digital image film compressed package and the encrypted real-time image inspection report compressed package based on the received encryption seed using a random decryption algorithm. If the decryption is successful, the process proceeds to the next step; otherwise, a data replacement alarm is issued and the detection ends.

[0032] After the decryption is successful, the decrypted real-time digital image film compressed package and the decrypted real-time image inspection report compressed package are obtained;

[0033] Extracting a first real-time compression package feature of the decrypted real-time digital image film compression package and a second real-time compression package feature of the decrypted real-time image inspection report compression package, and inputting the features into a pre-built data quality detection model;

[0034] Using the conditional information embedder of the data quality detection model, conditionally embed the first real-time compressed package feature and the second real-time compressed package feature to obtain a real-time conditional information embedding feature;

[0035] Using the generator of the data quality detection model, data quality detection is performed based on the embedded features of real-time condition information to obtain real-time data quality detection labels, i.e., real-time data quality detection results;

[0036] If the real-time data quality test result is qualified, the integrity test step is entered; otherwise, a quality failure alarm is issued and the test is ended.

[0037] Furthermore, the cloud data center performs integrity testing on the real-time digital image film compressed package and the real-time image inspection report compressed package to obtain a real-time integrity test result. If the real-time integrity test result is qualified, the next step is entered, which includes the following steps:

[0038] The cloud data center extracts the third file hash value of the decrypted real-time digital image film compressed package and the fourth file hash value of the decrypted real-time image inspection report compressed package;

[0039] Comparing the third file hash value with the first file hash value, and comparing the fourth file hash value with the second file hash value, to obtain a real-time integrity detection result;

[0040] If all the comparisons are consistent, the real-time integrity test result is qualified and the correlation test step is entered. Otherwise, a data corruption alarm is issued and the test ends.

[0041] Furthermore, the detection report generation model is built based on the Transformer algorithm.

[0042] Furthermore, the cloud data center generates a test report based on the real-time data quality test result, the real-time integrity test result, and / or the real-time correlation test result using a pre-built test report generation model to obtain a real-time test report, including the following steps:

[0043] Preprocessing and sequence conversion are performed on the real-time data quality detection results, the real-time integrity detection results, and / or the real-time correlation detection results to obtain a real-time detection result sequence;

[0044] Inputting the real-time detection result sequence into a pre-built detection report generation model, and using the detection report generation model to extract real-time detection result sequence features of the real-time detection result sequence;

[0045] According to the sequence characteristics of the real-time detection results, the detection report is generated to obtain the user's real-time detection report.

[0046] A digital imaging film and imaging inspection report detection system is used to implement a digital imaging film and imaging inspection report detection method. The system includes a cloud data center and several front-end servers. Each front-end server is communicatively connected to the cloud data center, and the front-end server includes a data compression unit and an encryption upload unit connected in sequence. The cloud data center includes a quality detection unit, an integrity detection unit, a correlation detection unit and a detection report generation unit connected in sequence.

[0047] The beneficial effects of the present invention are:

[0048] The present invention provides a method and system for detecting digital imaging films and imaging inspection reports, which significantly reduce manual intervention and improve the efficiency of processing a large number of digital imaging films and imaging inspection reports through automated data compression, data detection and report generation processes; introduces quality detection, integrity detection and correlation detection to conduct comprehensive detection of digital imaging films and imaging inspection reports, thereby improving detection accuracy; adopts encryption technology to upload data, thereby ensuring the security of digital imaging films and imaging inspection reports during transmission and storage, and preventing the leakage of sensitive information; utilizes a detection report generation model to automatically generate detection reports, thereby simplifying the report writing process, shortening the report generation cycle, and improving user experience.

[0049] Other beneficial effects of the present invention will be further described in the specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a flow chart of the detection method of digital image film and image inspection report in the present invention.

[0051] Figure 2 It is a structural block diagram of the detection system of digital image film and image inspection report in the present invention. DETAILED DESCRIPTION

[0052] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments.

[0053] Example 1:

[0054] like Figure 1 As shown, this embodiment provides a method for detecting digital imaging films and imaging inspection reports, comprising the following steps:

[0055] S1: The front-end server collects the user's real-time digital image film and real-time image inspection report, and uses a pre-built data compression model to compress the real-time digital image film and real-time image inspection report to obtain a mutually related real-time digital image film compressed package and real-time image inspection report compressed package, including the following steps:

[0056] S1-1: front-end server, collects the user's real-time digital image film and real-time image inspection report, and pre-processes them to obtain the pre-processed real-time digital image film and pre-processed real-time image inspection report;

[0057] Preprocessing is an important step in data compression and feature extraction, including image denoising, standardization, and cropping of digital imaging films, as well as serialization, conversion, and normalization of imaging inspection reports to ensure data quality and improve the accuracy of subsequent processing.

[0058] S1-2: Using the image feature extraction module of the pre-built data compression model, extract the real-time image features of the pre-processed real-time digital image film;

[0059] The data compression model is based on the Convolutional Neural Networks (CNN)-Long Short-Term Memory (LSTM)-Deep Belief Networks (DBN) algorithm. The data compression model includes an image feature extraction module based on the CNN algorithm, a sequence feature extraction module based on the LSTM algorithm, and a data compression module based on the DBN algorithm.

[0060] The image feature extraction module extracts deep features of digital imaging films and can capture complex features in digital imaging films. The sequence feature extraction module extracts sequence features of imaging inspection reports and can capture time dependencies in sequence data. Feature extraction is crucial for subsequent compression steps. The data compression module consists of multiple restricted Boltzmann machine (RBM) layers and can learn the probability distribution of data. The encoding process involves converting sub-blocks into a set of compressed codewords, which are compact representations of the original data and can learn effective representations of the data, thereby removing redundant information during the encoding process and achieving efficient compression. Since it is lossless compression, the encoding process of the DBN model can ensure that all information is retained so that the original data can be fully restored during decompression.

[0061] The method for constructing a data compression model includes the following steps:

[0062] A-1: Collect historical digital image films and historical image inspection reports of several users, and perform pre-processing to obtain several pre-processed historical digital image films and several pre-processed historical image inspection reports;

[0063] A-2: Use the CNN-LSTM-DBN algorithm to build an initial data compression model; the initial data compression model includes an initial image feature extraction module, an initial sequence feature extraction module, and an initial data compression module;

[0064] A-3: Using several pre-processed historical digital image films, the initial image feature extraction module is optimized and trained to obtain the final image feature extraction module, and several historical image data features are generated;

[0065] A-4: Using several pre-processed historical image inspection reports, the initial sequence feature extraction module is optimized and trained to obtain the final sequence feature extraction module, and several historical sequence data features are generated;

[0066] A-5: Based on several historical image data features and several historical sequence data features, the initial data compression module is optimized and trained to obtain the final data compression module;

[0067] A-6: Integrate the final image feature extraction module, the final sequence feature extraction module, and the final data compression module to obtain the final data compression model;

[0068] S1-3: Based on the real-time image features, the data compression module of the data compression model is used to perform data compression to obtain a real-time digital image film compression package;

[0069] S1-4: Using the image feature extraction module of the data compression model, the real-time sequence features of the real-time image inspection report after preprocessing are extracted;

[0070] S1-5: Based on the real-time sequence characteristics, the data compression module of the data compression model is used to compress the data to obtain a compressed package of the real-time imaging inspection report after preprocessing;

[0071] S1-6: extracting first compressed package information of the real-time digital image film compressed package and second compressed package information of the real-time image inspection report compressed package;

[0072] S1-7: Associating the first compressed package information with the second compressed package information to obtain real-time associated information, as well as the mutually associated real-time digital image film compressed package and the real-time image inspection report compressed package;

[0073] Ensure the correlation between the image film and the inspection report, which facilitates subsequent correlation detection. Through the correlation information, it can quickly detect whether the real-time digital image film compression package and the real-time image inspection report compression package are from the same user, avoiding data replacement or transmission errors.

[0074] S2: The front-end server encrypts and uploads the related real-time digital image film compressed package and the real-time image inspection report compressed package to the cloud data center, including the following steps:

[0075] S2-1: The front-end server uses a hash function to generate a hash value for the first file of the real-time digital image film compressed package and a hash value for the second file of the real-time image inspection report compressed package. By comparing the hash values ​​before and after uploading, it can verify whether the data has been tampered with during transmission. The hash value serves as a unique identifier for the file, facilitating subsequent data management and retrieval.

[0076] A hash function is a function that maps data of any length to data of a fixed length and is used to verify the integrity of the data.

[0077] S2-2: Generate a random encryption key using a random encryption algorithm, and generate an encryption seed based on the random encryption key;

[0078] Random encryption keys and encryption seeds increase the difficulty of cracking encryption and improve data security. Each encryption uses a different key, which increases the unpredictability of encryption.

[0079] S2-3: Encrypting the real-time digital image film compressed package and the real-time image inspection report compressed package according to the encryption seed to obtain an encrypted real-time digital image film compressed package and an encrypted real-time image inspection report compressed package that are associated with each other;

[0080] Even if the encrypted data is illegally accessed, it cannot be easily deciphered, thus protecting the user's privacy;

[0081] S2-4: Uploading the real-time association information, the encrypted seed, the first file hash value, the second file hash value, the encrypted real-time digital image film compressed package, and the encrypted real-time image inspection report compressed package to the cloud data center;

[0082] S3: The cloud data center uses a pre-built data quality detection model to perform data quality detection on the real-time digital image film compression package and the real-time image inspection report compression package to obtain the real-time data quality detection result. If the real-time data quality detection result is qualified, the next step is entered, which includes the following steps:

[0083] S3-1: The cloud data center uses a random decryption algorithm based on the received encryption seed to decrypt the received encrypted real-time digital image film compressed package and the encrypted real-time image inspection report compressed package. If the decryption is successful, the process proceeds to the next step; otherwise, a data replacement alarm is issued and the detection ends.

[0084] Ensure that only authorized users with the correct encryption seed can decrypt the data and can determine whether the data has been replaced based on the decryption results;

[0085] S3-2: After decryption is successful, the decrypted real-time digital image film compressed package and the decrypted real-time image inspection report compressed package are obtained;

[0086] S3-3: extracting a first real-time compression package feature of the decrypted real-time digital image film compression package and a second real-time compression package feature of the decrypted real-time image inspection report compression package, and inputting the features into a pre-built data quality detection model;

[0087] The data quality detection model is built based on the Conditional Generative Adversarial Network (cGAN)-Multilayer Perceptron (MLP) algorithm, and includes a conditional information embedder and a conditional information processor, both built based on the MLP algorithm, and a generator and a discriminator, both built based on the Recurrent Neural Network (RNN) algorithm. The generator is connected to the conditional information embedder and discriminator, respectively, and the discriminator is connected to the conditional information processor.

[0088] The conditional information embedder is used to process the conditional information of the compressed package features, obtain the conditional information embedding features in sequence format, and integrate the conditional information of the compressed package into the generation process; the generator is used to perform label prediction based on the conditional information embedding features and random noise output by the conditional information embedder; the conditional information processor is used to process the additional conditional information of the compressed package to help the discriminator more accurately judge the authenticity of the label; the discriminator analyzes whether the predicted label is true and meets the given conditional information; the generator and the discriminator compete with each other through an adversarial training process, the generator attempts to generate more realistic labels, and the discriminator attempts to better distinguish between real labels and generated labels, thereby improving the generator's label prediction accuracy;

[0089] The method for constructing a data quality detection model includes the following steps:

[0090] B-1: Collecting a number of compressed packages of historical digital image films and a number of compressed packages of historical image inspection reports, and preprocessing them to obtain a number of compressed packages of preprocessed historical digital image films and a number of compressed packages of preprocessed historical image inspection reports;

[0091] B-2: Set a corresponding real data quality detection label for each pre-processed historical digital image film compression package and the corresponding pre-processed historical image inspection report compression package;

[0092] B-3: Use the cGAN-MLP algorithm to build an initial data quality detection model; the initial data quality detection model includes an initial generator and an initial discriminator. The initial generator is equipped with a conditional information embedder, and the initial discriminator is equipped with a conditional information processor.

[0093] B-4: Combine the first loss function of the initial generator and the second loss function of the initial discriminator to obtain a comprehensive loss function;

[0094] B-5: extracting the first historical compression package feature of the pre-processed historical digital image film compression package and the second historical compression package feature of the pre-processed historical image inspection report compression package;

[0095] B-6: Using the conditional information embedder, conditionally embed the first historical compressed package feature and the second historical compressed package feature to obtain a historical conditional information embedding feature;

[0096] B-7: Use the initial generator to embed features based on historical condition information, perform data quality detection, and obtain historical data quality detection labels;

[0097] B-8: Using a condition information processor, performing condition information processing based on the first historical compressed package characteristics and the second historical compressed package characteristics to obtain historical condition information;

[0098] B-9: Use the initial discriminator to perform data discrimination based on the historical data quality detection labels, the corresponding real data quality detection labels, and the historical condition information to obtain the historical data discrimination results;

[0099] B-10: Traverse all pre-processed historical digital image film compression packages and corresponding pre-processed historical image inspection report compression packages, repeat the above steps, and optimize the training of the initial data quality detection model;

[0100] B-11: Use the comprehensive loss function to generate the historical loss value of the optimization training process. If the historical loss value is lower than the loss value threshold, the final data quality detection model is output. Otherwise, the optimization training continues.

[0101] S3-4: Using the conditional information embedder of the data quality detection model, conditionally embed the first real-time compressed package feature and the second real-time compressed package feature to obtain a real-time conditional information embedding feature;

[0102] S3-5: Use the generator of the data quality detection model to embed features based on real-time condition information, perform data quality detection, and obtain real-time data quality detection labels, i.e., real-time data quality detection results;

[0103] S3-6: If the real-time data quality test result is qualified, then enter the integrity test step; otherwise, issue a quality failure alarm and end the test;

[0104] S4: The cloud data center performs integrity testing on the real-time digital image film compression package and the real-time image inspection report compression package to obtain a real-time integrity test result. If the real-time integrity test result is qualified, the next step is entered, which includes the following steps:

[0105] S4-1: The cloud data center extracts the third file hash value of the decrypted real-time digital image film compressed package and the fourth file hash value of the decrypted real-time image inspection report compressed package;

[0106] S4-2: Compare the hash value of the third file with the hash value of the first file, and compare the hash value of the fourth file with the hash value of the second file to obtain a real-time integrity test result. By comparing the hash values, it can be determined whether the data has been tampered with or damaged during transmission;

[0107] S4-3: If all comparisons are consistent, the real-time integrity test result is qualified and the correlation test step is entered. Otherwise, a data corruption alarm is issued and the test ends. If the hash values ​​are inconsistent, it indicates that there may be a problem with the data during transmission. For incomplete data detected, an alarm is issued in a timely manner, and measures can be taken to repair or retransmit it, reducing potential errors and losses.

[0108] S5: The cloud data center performs a correlation test on the real-time digital image film compression package and the real-time image inspection report compression package to obtain a real-time correlation test result. If the real-time correlation test result is qualified, the next step is entered, which includes the following steps:

[0109] S5-1: Parse the real-time associated information to obtain associated first compressed package information and second compressed package information;

[0110] S5-2: extracting the third compressed package information of the decrypted real-time digital image film compressed package and the fourth compressed package information of the decrypted real-time image inspection report compressed package;

[0111] S5-3: Compare the third compressed package information with the first compressed package information, and compare the fourth compressed package information with the second compressed package information to obtain a real-time correlation detection result;

[0112] S5-4: If all comparisons are consistent, the real-time correlation test result is qualified and the test report generation step is entered. Otherwise, a data misalignment alarm is issued and the test is terminated. This ensures that the correlation between the imaging film and the examination report remains unchanged during the transmission process. If the information does not match, it indicates that an error may have occurred during the processing or transmission process. In the event of a mismatch, an alarm is issued in a timely manner, and corrective measures can be taken to avoid medical errors caused by data misalignment.

[0113] S6: The cloud data center generates a test report based on the real-time data quality test results, the real-time integrity test results, and / or the real-time correlation test results using a pre-built test report generation model to obtain a real-time test report, including the following steps:

[0114] S6-1: Preprocess and sequence the real-time data quality test results, real-time integrity test results, and / or real-time relevance test results to obtain a real-time test result sequence; format and serialize the quality, integrity, and relevance test results so that the model can understand and process the data;

[0115] S6-2: inputting the real-time test result sequence into a pre-built test report generation model, and using the test report generation model to extract real-time test result sequence features of the real-time test result sequence;

[0116] The test report generation model is built based on the Transformer algorithm; the Transformer model can capture long-distance dependencies in sequence data and provide rich features for report generation;

[0117] The method for constructing a test report generation model includes the following steps:

[0118] C-1: Use the Transformer algorithm to build an initial test report generation model. Use a text database to pre-train the initial test report generation model to obtain a pre-trained test report generation model. Pre-training enables the model to understand and generate natural language text.

[0119] C-2: Collect several historical test reports and corresponding historical data quality test results, historical integrity test results, and historical relevance test results, and perform preprocessing to obtain several test report generation samples;

[0120] C-3: Generate samples based on several test reports, optimize and train the pre-trained test report generation model, and obtain the final test report generation model; use historical data to enable the model to be optimized for specific fields;

[0121] S6-3: Generate a test report based on the sequence characteristics of the real-time test results to obtain a real-time test report for the user.

[0122] Example 2:

[0123] like Figure 2 As shown, this embodiment provides a digital imaging film and imaging inspection report detection system for implementing a digital imaging film and imaging inspection report detection method. The system includes a cloud data center and a plurality of front-end servers. Each front-end server is communicatively connected to the cloud data center. The front-end servers include a data compression unit and an encryption upload unit connected in sequence. The cloud data center includes a quality detection unit, an integrity detection unit, a correlation detection unit, and a detection report generation unit connected in sequence.

[0124] A data compression unit is used to collect the user's real-time digital image film and real-time image inspection report, and use a pre-built data compression model to compress the real-time digital image film and real-time image inspection report to obtain a mutually related real-time digital image film compression package and real-time image inspection report compression package;

[0125] An encryption uploading unit is used to encrypt and upload the mutually related real-time digital image film compressed package and real-time image inspection report compressed package to the cloud data center;

[0126] A quality detection unit is used to perform data quality detection on the real-time digital image film compression package and the real-time image inspection report compression package using a pre-built data quality detection model to obtain real-time data quality detection results;

[0127] An integrity detection unit is used to perform integrity detection on the real-time digital image film compression package and the real-time image inspection report compression package to obtain a real-time integrity detection result;

[0128] A correlation detection unit is used to perform correlation detection on the real-time digital image film compression package and the real-time image inspection report compression package to obtain a real-time correlation detection result;

[0129] The test report generation unit is used to generate a test report based on the real-time data quality test results, the real-time integrity test results and / or the real-time correlation test results using a pre-built test report generation model to obtain a real-time test report.

[0130] The present invention provides a method and system for detecting digital imaging films and imaging inspection reports, which significantly reduce manual intervention and improve the efficiency of processing a large number of digital imaging films and imaging inspection reports through automated data compression, data detection and report generation processes; introduces quality detection, integrity detection and correlation detection to conduct comprehensive detection of digital imaging films and imaging inspection reports, thereby improving detection accuracy; adopts encryption technology to upload data, thereby ensuring the security of digital imaging films and imaging inspection reports during transmission and storage, and preventing the leakage of sensitive information; utilizes a detection report generation model to automatically generate detection reports, thereby simplifying the report writing process, shortening the report generation cycle, and improving user experience.

[0131] The present invention is not limited to the above optional embodiments. Anyone can derive various other forms of products based on the teachings of the present invention. The above specific embodiments should not be construed as limiting the scope of protection of the present invention. The scope of protection of the present invention shall be based on the scope defined in the claims, and the description can be used to interpret the claims.

Claims

1. A method for detecting digital imaging films and imaging inspection reports, characterized in that: The steps include: The front-end server collects the user's real-time digital image film and real-time image inspection report, and uses a pre-built data compression model to compress the real-time digital image film and real-time image inspection report to obtain a mutually related real-time digital image film compressed package and real-time image inspection report compressed package; The front-end server encrypts and uploads the related real-time digital image film compression package and the real-time image inspection report compression package to the cloud data center; The cloud data center uses a pre-built data quality detection model to perform data quality detection on the real-time digital image film compression package and the real-time image inspection report compression package to obtain the real-time data quality detection result. If the real-time data quality detection result is qualified, it proceeds to the next step; The cloud data center performs integrity testing on the real-time digital image film compression package and the real-time image inspection report compression package to obtain the real-time integrity test result. If the real-time integrity test result is qualified, the process proceeds to the next step. The cloud data center performs a correlation test on the real-time digital image film compression package and the real-time image inspection report compression package to obtain a real-time correlation test result. If the real-time correlation test result is qualified, the process proceeds to the next step. The cloud data center generates a test report based on the real-time data quality test results, the real-time integrity test results and / or the real-time correlation test results using a pre-built test report generation model to obtain a real-time test report.

2. The method for detecting digital imaging films and imaging inspection reports according to claim 1, characterized in that: The data compression model is constructed based on the CNN-LSTM-DBN algorithm, and the data compression model includes an image feature extraction module constructed based on the CNN algorithm, a sequence feature extraction module constructed based on the LSTM algorithm, and a data compression module constructed based on the DBN algorithm.

3. The method for detecting digital imaging films and imaging inspection reports according to claim 2, wherein: The front-end server collects the user's real-time digital image film and real-time image inspection report, and uses a pre-built data compression model to compress the real-time digital image film and real-time image inspection report to obtain a mutually related real-time digital image film compressed package and real-time image inspection report compressed package, including the following steps: The front-end server collects the user's real-time digital image film and real-time image inspection report, and performs pre-processing to obtain the pre-processed real-time digital image film and the pre-processed real-time image inspection report; Using the image feature extraction module of the pre-built data compression model, extracting the real-time image features of the pre-processed real-time digital image film; According to the real-time image features, the data compression module of the data compression model is used to compress the data to obtain a real-time digital image film compression package; The image feature extraction module of the data compression model is used to extract the real-time sequence features of the real-time image inspection report after preprocessing; According to the real-time sequence characteristics, the data compression module of the data compression model is used to compress the data to obtain a compressed package of the real-time image inspection report after preprocessing; Extracting first compressed package information of the real-time digital image film compressed package and second compressed package information of the real-time image inspection report compressed package; The first compressed package information and the second compressed package information are associated to obtain real-time associated information, as well as mutually associated real-time digital image film compressed packages and real-time image inspection report compressed packages.

4. The method for detecting digital imaging films and imaging inspection reports according to claim 3, wherein: The front-end server encrypts and uploads the related real-time digital image film compression package and the real-time image inspection report compression package to the cloud data center, including the following steps: The front-end server generates a first file hash value of the real-time digital image film compression package and a second file hash value of the real-time image inspection report compression package; Use a random encryption algorithm to generate a random encryption key, and generate an encryption seed based on the random encryption key; Encrypting the real-time digital image film compressed package and the real-time image inspection report compressed package according to the encryption seed to obtain mutually related encrypted real-time digital image film compressed package and encrypted real-time image inspection report compressed package; The real-time association information, the encrypted seed, the first file hash value, the second file hash value, the encrypted real-time digital image film compressed package, and the encrypted real-time image inspection report compressed package are uploaded to the cloud data center.

5. The method for detecting digital imaging films and imaging inspection reports according to claim 4, characterized in that: The data quality detection model is constructed based on the cGAN-MLP algorithm, and the data quality detection model includes a conditional information embedder and a conditional information processor both constructed based on the MLP algorithm, and a generator and a discriminator both constructed based on the RNN algorithm. The generator is connected to the conditional information embedder and the discriminator respectively, and the discriminator is connected to the conditional information processor.

6. The method for detecting digital imaging films and imaging inspection reports according to claim 5, characterized in that: The cloud data center uses a pre-built data quality detection model to perform data quality detection on the real-time digital image film compression package and the real-time image inspection report compression package to obtain the real-time data quality detection result. If the real-time data quality detection result is qualified, the next step is entered, which includes the following steps: The cloud data center decrypts the encrypted real-time digital image film compressed package and the encrypted real-time image inspection report compressed package based on the received encryption seed using a random decryption algorithm. If the decryption is successful, the process proceeds to the next step; otherwise, a data replacement alarm is issued and the detection ends. After the decryption is successful, the decrypted real-time digital image film compressed package and the decrypted real-time image inspection report compressed package are obtained; Extracting a first real-time compression package feature of the decrypted real-time digital image film compression package and a second real-time compression package feature of the decrypted real-time image inspection report compression package, and inputting the features into a pre-built data quality detection model; Using the conditional information embedder of the data quality detection model, conditionally embed the first real-time compressed package feature and the second real-time compressed package feature to obtain a real-time conditional information embedding feature; Using the generator of the data quality detection model, data quality detection is performed based on the embedded features of real-time condition information to obtain real-time data quality detection labels, i.e., real-time data quality detection results; If the real-time data quality test result is qualified, the integrity test step is entered; otherwise, a quality failure alarm is issued and the test is ended.

7. The method for detecting digital imaging films and imaging inspection reports according to claim 6, characterized in that: The cloud data center performs integrity testing on the real-time digital image film compression package and the real-time image inspection report compression package to obtain a real-time integrity test result. If the real-time integrity test result is qualified, the next step is entered, which includes the following steps: The cloud data center extracts the third file hash value of the decrypted real-time digital image film compressed package and the fourth file hash value of the decrypted real-time image inspection report compressed package; Comparing the third file hash value with the first file hash value, and comparing the fourth file hash value with the second file hash value, to obtain a real-time integrity detection result; If all the comparisons are consistent, the real-time integrity test result is qualified and the correlation test step is entered. Otherwise, a data corruption alarm is issued and the test ends.

8. The method for detecting digital imaging films and imaging inspection reports according to claim 7, characterized in that: The test report generation model is built based on the Transformer algorithm.

9. The method for detecting digital imaging films and imaging inspection reports according to claim 8, characterized in that: The cloud data center generates a test report based on the real-time data quality test results, the real-time integrity test results, and / or the real-time correlation test results using a pre-built test report generation model to obtain a real-time test report, including the following steps: Preprocessing and sequence conversion are performed on the real-time data quality detection results, the real-time integrity detection results, and / or the real-time correlation detection results to obtain a real-time detection result sequence; Inputting the real-time detection result sequence into a pre-built detection report generation model, and using the detection report generation model to extract real-time detection result sequence features of the real-time detection result sequence; According to the sequence characteristics of the real-time detection results, a detection report is generated to obtain a real-time detection report for the user.

10. A digital imaging film and imaging inspection report detection system, used to implement the digital imaging film and imaging inspection report detection method according to any one of claims 1 to 9, characterized in that: The system includes a cloud data center and several front-end servers, each of which is communicatively connected to the cloud data center, and the front-end servers include a data compression unit and an encryption upload unit connected in sequence, and the cloud data center includes a quality detection unit, an integrity detection unit, a correlation detection unit and a detection report generation unit connected in sequence.

Images