Medical image quality control management method and system

Abstract

The invention relates to the technical field of medical images, in particular to a medical image quality control management method and system. Aiming at the technical problems of insufficient differentiation evaluation, missing quality problem tracking analysis, brief quality control report content, low historical data utilization rate and the like in the existing ultrasonic image quality control, the method comprises the following steps: acquiring ultrasonic image data in a target medical institution; performing feature extraction on the acquired ultrasonic image data; inputting the extracted features into a preset ultrasonic image quality evaluation model; judging the quality grade of the ultrasonic image data according to an output result of the ultrasonic image quality evaluation model; and generating a quality control report and marking the ultrasonic image data of which the quality does not reach the standard as a to-be-reviewed state. The method also comprises the functions of generating a targeted training plan, carrying out historical quality control data trend analysis, establishing physician and equipment quality correlation analysis and the like, realizes the intelligence and systematization of ultrasonic image quality control, and effectively improves the diagnosis accuracy of ultrasonic examination of medical institutions.

Description

Technical Field

[0001] The present invention relates to the field of medical imaging technology, and in particular to a medical imaging quality control management method and system. Background Art

[0002] With the rapid development of medical imaging technology, ultrasound imaging, as a non-invasive, convenient, and economical examination method, is playing an increasingly important role in clinical diagnosis. Ultrasound imaging, with its advantages of real-time operation, lack of radiation, and ease of operation, has become the preferred method for diagnosing and monitoring treatment of many diseases. However, ultrasound examination results are highly dependent on the operator's technical level, the standardization of the examination environment, and the performance status of the equipment. As a key link in ensuring diagnostic accuracy, medical ultrasound imaging quality control has received considerable attention from medical institutions in recent years. Traditional medical ultrasound imaging quality control relies primarily on regular spot checks and manual scoring by quality control experts within medical institutions. This approach not only consumes a large amount of human resources but also makes it difficult to achieve comprehensive coverage of all examination data, resulting in a certain degree of subjectivity and lag in quality control results. With the application of artificial intelligence technology in the field of medical imaging, image quality assessment methods based on deep learning have provided a new technical path for ultrasound imaging quality control.

[0003] However, existing medical ultrasound imaging quality control technologies still have many shortcomings. First, existing quality control systems often lack the ability to conduct differentiated evaluations of imaging data obtained from different examination sites, different examination equipment, and different examination physicians, and are unable to specifically discover and resolve quality problems in specific links. Secondly, most quality control systems only stay at the problem discovery stage, lack the function of tracking and analyzing quality problems, cannot establish quality correlation analysis between physicians and equipment, and are difficult to provide medical institutions with guidance and support for continuous improvement. In addition, existing quality control reports are usually single in format and brief in content, lack in-depth analysis of quality problems and improvement suggestions, and cannot meet the needs of medical institutions for refined management. Finally, traditional quality control methods have low utilization rates of historical data, making it difficult to discover the regularity and periodic changes of quality problems. The lack of an early warning mechanism leads to a large time lag between problem discovery and resolution, affecting the overall quality and diagnostic efficiency of ultrasound examinations. Summary of the Invention

[0004] In view of the problems existing in the prior art, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is how to solve the technical problems existing in the quality control process of medical ultrasound imaging, such as insufficient differentiated evaluation, lack of quality problem tracking and analysis, brief content of quality control reports, and low utilization of historical data. The present invention obtains ultrasound imaging data from the target medical institution, extracts features and inputs them into a preset quality assessment model to achieve intelligent evaluation of ultrasound image quality and generate a quality control report containing content such as quality grade distribution and problem type statistics. At the same time, it marks substandard images and establishes a quality correlation analysis between physicians and equipment, achieving accurate discovery and continuous improvement of quality problems, effectively improving the overall quality level of ultrasound examinations in medical institutions.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, an embodiment of the present invention provides a medical image quality control management method, which includes acquiring ultrasound image data within a target medical institution; Performing feature extraction on the acquired ultrasound image data; Input the extracted features into a preset ultrasound image quality assessment model; Judging the quality level of ultrasound image data according to the output results of the ultrasound image quality assessment model; Generate quality control reports and mark ultrasound image data that does not meet quality standards as pending review.

[0007] As a preferred solution of the medical image quality control management method of the present invention, the ultrasound image data includes: ultrasound image data of different examination parts, ultrasound image data obtained by different examination equipment, ultrasound image data obtained by different examination physicians, and ultrasound image data of different pathological conditions.

[0008] As a preferred solution of the medical image quality control management method of the present invention, the quality parameters evaluated by the ultrasound image quality assessment model include at least image clarity, contrast, signal-to-noise ratio and integrity of key anatomical structures.

[0009] As a preferred solution of the medical imaging quality control management method of the present invention, it also includes automatically generating a targeted training plan based on the quality control report, and the training plan includes operation instructions and quality improvement suggestions for specific ultrasound examination types.

[0010] As a preferred solution of the medical image quality control management method of the present invention, it also includes trend analysis of historical quality control data, identifying the regularity and periodic changes of quality problems, and generating early warning information.

[0011] As a preferred solution of the medical image quality control management method of the present invention, it also includes establishing a quality correlation analysis between physicians and examination equipment to identify the operational advantages and disadvantages of specific physicians on specific equipment.

[0012] As a preferred solution of the medical image quality control management method of the present invention, the quality control report includes quality grade distribution statistics, problem type statistics, physician performance analysis and equipment status evaluation.

[0013] In a second aspect, an embodiment of the present invention provides a medical image quality control management system, which includes an image data acquisition module for acquiring ultrasound image data within a target medical institution; A feature extraction module, used for extracting features from the acquired ultrasound image data; A quality assessment module is used to input the extracted features into a preset ultrasound image quality assessment model and determine the quality level of the ultrasound image data based on the output results of the ultrasound image quality assessment model; The report generation module is used to generate quality control reports and mark ultrasound image data that does not meet the quality standards as pending review.

[0014] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the medical image quality control management method of the first aspect of the present invention are implemented.

[0015] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the medical image quality control management method of the first aspect of the present invention are implemented.

[0016] The beneficial effects of the present invention are as follows: The medical image quality control management method and system provided by the present invention have the following significant beneficial effects: First, by acquiring ultrasound image data including different examination sites, different examination equipment, different examination physicians and different pathological conditions, comprehensive data collection is achieved, providing a rich and diverse basic data sample for quality assessment. Secondly, by extracting features from ultrasound image data, including quality parameters such as image clarity, contrast, signal-to-noise ratio and integrity of key anatomical structures, an objective and comprehensive quality evaluation index system is established. Third, intelligent rating is performed through a preset ultrasound image quality assessment model, ensuring the objectivity and consistency of the assessment process and avoiding the subjectivity and uncertainty of traditional manual assessment methods. Fourth, the quality control report generated based on the assessment results covers multiple dimensions such as quality grade distribution, problem type statistics, physician performance analysis and equipment status assessment, providing medical institutions with a comprehensive and intuitive quality management basis. Fifth, the system can automatically mark images that do not meet quality standards and push them to the relevant responsible persons, forming a closed-loop management to ensure that quality problems are corrected in a timely manner. The system also includes advanced features such as generating targeted training plans, analyzing historical data trends, and establishing correlations between physician and equipment quality, enabling comprehensive management from problem discovery to continuous improvement. Overall, this invention utilizes technical means to achieve intelligent, standardized, and systematized ultrasound imaging quality control, significantly improving the overall quality and diagnostic accuracy of ultrasound examinations in medical institutions, optimizing the allocation of medical resources, and enhancing the efficiency and relevance of quality control efforts, providing more reliable imaging support for clinical decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0018] Figure 1 This is a flow chart of the medical imaging quality control management method; Figure 2 Diagram of computer equipment for medical image quality control management method. DETAILED DESCRIPTION

[0019] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0020] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0021] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it designate a separate or selective embodiment that is mutually exclusive with other embodiments. Example 1

[0022] Reference Figures 1 and 2 , which is the first embodiment of the present invention, provides a medical image quality control management method, including: S100: Acquiring ultrasound image data in a target medical institution; S200: extracting features from the acquired ultrasound image data; S300: Inputting the extracted features into a preset ultrasound image quality assessment model; S400: judging the quality level of the ultrasound image data according to the output result of the ultrasound image quality assessment model; S500: Generate a quality control report and mark the ultrasound image data that does not meet the quality standards as pending review.

[0023] It should be noted that medical institutions face multiple challenges in ultrasound imaging quality control. First, ultrasound examination quality is highly dependent on the skill level and experience of the physician performing the examination, resulting in significant variations in image quality between physicians within the same institution. Second, the imaging characteristics, maintenance status, and parameter settings of different examination equipment can significantly impact image quality, but existing quality control systems often lack a systematic evaluation mechanism for equipment performance. Furthermore, ultrasound imaging quality assessment standards lack uniformity and objectivity. Traditional manual quality control methods are not only time-consuming and labor-intensive, but also often subjective, making it difficult to establish a standardized quality control process. Furthermore, medical institutions often lack the ability to conduct in-depth analysis of historical quality control data, unable to identify regularities and trends in quality issues from accumulated data over long periods of time, nor can they provide targeted improvement recommendations for specific physicians or equipment. Finally, existing quality control systems often rely on post-hoc evaluations and lack real-time feedback and early warning mechanisms, resulting in a time lag between problem discovery and resolution, impacting the timeliness and accuracy of medical decision-making. These combined challenges severely limit the diagnostic value and reliability of ultrasound imaging in clinical applications.

[0024] Therefore, this embodiment constructs a complete medical ultrasound imaging quality control management method through steps S100 to S500. The method first obtains ultrasound imaging data from the target medical institution, covering a comprehensive sample of different examination sites, different examination equipment, different physician operations, and different pathological conditions; then, feature extraction is performed on this data to capture core quality parameters such as image clarity, contrast, signal-to-noise ratio, and the integrity of key anatomical structures; the extracted features are then input into a preset quality assessment model to achieve an objective assessment of image quality; based on the assessment results, the system determines the quality level of the image data and generates a quality control report containing quality level distribution statistics, problem type statistics, physician performance analysis, and equipment status assessment; at the same time, the system marks images that do not meet quality standards, reminds relevant personnel to review them, and automatically generates targeted training plans based on the quality control results; in addition, the system can perform trend analysis on historical quality control data, identify the regularity and periodic changes of quality issues, generate early warning information in a timely manner, and establish a quality correlation analysis between physicians and examination equipment to identify the advantages and disadvantages of specific physicians operating on specific equipment. Through this series of steps, this embodiment realizes the automation, standardization, intelligence and early warning of ultrasound image quality control, fundamentally improving the overall quality level and diagnostic accuracy of ultrasound examinations in medical institutions, and providing more reliable imaging basis for clinical decision-making. Example 2

[0025] Reference Figure 1 - Figure 2 , which is the second embodiment of the present invention.

[0026] In the embodiment of the present application, obtaining ultrasound image data in the target medical institution in step S100 includes the following steps A1-A2: A1: Ultrasound image data includes: ultrasound image data of different examination parts, ultrasound image data obtained by different examination equipment, ultrasound image data obtained by different examination physicians, and ultrasound image data of different pathological conditions.

[0027] Specifically, ultrasound image data from different examination sites may include examination data from the abdomen (e.g., liver, gallbladder, pancreas, spleen), heart, blood vessels, breast, thyroid, gynecology (e.g., uterus, ovaries), obstetrics (e.g., fetal development), musculoskeletal (e.g., muscles, ligaments, joints), and the urinary system (e.g., kidneys, bladder, prostate). Ultrasound image data acquired by different examination devices may come from different brands or models, such as 2D, 3D, color Doppler, and elastography devices from manufacturers like Philips, GE, Siemens, and Toshiba. Ultrasound image data acquired by different examining physicians reflects differences in their experience, technical skills, and personal operating habits. Ultrasound image data from different pathological conditions include imaging manifestations of normal tissue, benign lesions, malignant lesions, inflammation, cysts, and stones.

[0028] In an optional embodiment, the ultrasound image data acquired from the target medical institution in step S100 can also be automatically acquired through a PACS (Picture Archiving and Communication System) data interface. The system can establish a secure connection with the hospital's PACS system to periodically or in real time capture newly added ultrasound image data, achieving automated and seamless data integration. PACS systems typically store complete DICOM-formatted image files, including device parameters, examination information, and basic patient information (after desensitization). This provides rich metadata support for subsequent quality control analysis.

[0029] In another optional embodiment, the ultrasound image data acquired in step S100 can also be collected and uploaded via a mobile terminal app. Physicians can use a dedicated app to upload key screenshots or video clips from the ultrasound examination directly to the quality control system, adding necessary annotations. This approach is particularly suitable for mobile ultrasound equipment or scenarios that are not fully integrated into a PACS system, ensuring that all ultrasound examinations are included in the quality control scope.

[0030] In the embodiment of the present application, feature extraction of the ultrasound image data acquired in step S200 includes the following steps B1-B2: B1: Extract objective features of ultrasound images, including but not limited to computer vision parameters such as edge clarity, contrast, brightness uniformity, noise distribution, texture characteristics, and structural integrity.

[0031] B2: Extract professional features of ultrasound images, including but not limited to compliance with standard sections, visibility of key anatomical landmarks, standardization of measurement markings, completeness of image annotations, and other medical professional requirements.

[0032] Specifically, in B1, edge clarity can be calculated and statistically distributed by using the Sobel operator or the Canny edge detection algorithm to determine the gradient strength of the edge in the image; contrast can be quantified by calculating the difference in grayscale values between the target area and the background area in the image; brightness uniformity can be evaluated by analyzing the standard deviation of brightness in different areas of the image; noise distribution can be quantified by high-frequency component analysis or wavelet transform coefficients; texture features can be extracted through the gray-level co-occurrence matrix (GLCM) to extract statistical parameters such as energy, contrast, correlation, and entropy; and structural integrity can be evaluated by detecting the continuity and integrity of key anatomical structures.

[0033] Specifically, in B2, the conformity of standard sections can be evaluated by aligning and calculating similarity with preset standard section templates; the visibility of key anatomical landmarks is quantified by detecting and identifying the presence and clarity of specific anatomical structures; the standardization of measurement marks is evaluated by analyzing whether the position, angle, and length of the measurement lines meet the standard requirements; and the completeness of image annotations is assessed by checking whether the necessary text labels, direction marks, and scale bars are present and correctly placed.

[0034] It should be noted that the feature extraction process utilizes a multi-level analysis approach, first assessing overall quality at the macro level, such as signal-to-noise ratio and contrast, then performing a meso-level structural assessment, such as organ boundary clarity and echo texture uniformity, and finally performing a micro-level detailed assessment, such as the discernibility of small structures and the accuracy of measurement markers. This multi-level feature extraction strategy comprehensively captures the quality characteristics of ultrasound images, providing rich data support for subsequent quality assessments.

[0035] In an optional embodiment, feature extraction in step S200 can also be performed by automatically learning feature representations through a deep learning network. Using a pre-trained convolutional neural network (such as ResNet or VGG) as a feature extractor, this method automatically learns multi-level feature representations for ultrasound images, eliminating the need for manually designed feature operators. This approach can capture complex features that are difficult to describe using traditional methods, and its feature extraction capabilities are continuously optimized as more training data is collected.

[0036] In another optional embodiment, the feature extraction in step S200 can also be combined with real-time feedback from clinicians to dynamically adjust feature weights. The system records and analyzes clinician feedback on quality control results and automatically adjusts the weight parameters of different features to make the feature extraction process more consistent with actual clinical needs and improve the clinical relevance of quality control assessments.

[0037] In the embodiment of the present application, the features extracted in step S300 are input into a preset ultrasound image quality assessment model, including the following steps C1-C3: C1: The quality parameters assessed by the ultrasound image quality assessment model include at least: image clarity, contrast, signal-to-noise ratio, and integrity of key anatomical structures.

[0038] C2: Combine the extracted objective features and professional features into feature vectors, and perform dimensionality reduction through feature normalization and principal component analysis.

[0039] C3: The processed feature vector is input into a preset multi-level quality assessment model, which combines a rule-based scoring system and a learning-based classification algorithm.

[0040] Specifically, in C1, the feature vector contains all extracted objective and specialized features, potentially containing dozens or even hundreds of dimensions. These features are normalized using Min-Max normalization, mapping all eigenvalues to the range [0, 1] to eliminate dimensionality differences. Dimensionality reduction techniques such as principal component analysis (PCA) or autoencoders are then used to remove redundant information while retaining the primary components of variation, ultimately resulting in a compact and information-rich feature representation.

[0041] Specifically, in C3, the multi-level quality assessment model first evaluates basic quality parameters (such as signal-to-noise ratio and contrast) through a rule-based scoring system to ensure that basic quality meets standards. It then uses learning-based classification algorithms (such as support vector machines, random forests, or deep neural networks) to comprehensively assess overall quality, outputting a quality grade and detailed score. The model is trained using standard samples and expert-annotated data from multiple medical institutions to ensure consistency and professionalism in the evaluation criteria.

[0042] It should be noted that the quality assessment model utilizes a tiered evaluation strategy, setting differentiated evaluation criteria for different examination sites and clinical purposes. For example, for cardiac ultrasound examinations, the model focuses on assessing the clarity of cardiac chamber boundaries and the quality of valve motion visualization; for obstetric ultrasound examinations, it focuses on assessing the completeness of the visualization of key fetal anatomical structures. This targeted evaluation approach ensures the clinical relevance and practicality of quality control results. Furthermore, the model possesses self-learning capabilities, enabling it to continuously optimize evaluation criteria based on physician feedback and accumulated data to adapt to the specific needs of medical institutions.

[0043] In an optional embodiment, the quality assessment model in step S300 may also use an ensemble learning method to combine the prediction results of multiple basic models. By integrating the assessment results of multiple models such as decision trees, support vector machines, and neural networks, the bias of a single model is reduced and the stability and accuracy of the assessment are improved.

[0044] In another optional embodiment, the quality assessment model in step S300 can also employ a fuzzy logic system, addressing uncertainty in quality assessment through a fuzzy rule set and membership functions. This approach can better address fuzzy boundaries in ultrasound image quality assessment, such as the gradual transition between "relatively clear" and "very clear," making the assessment results more consistent with human expert judgment.

[0045] In the embodiment of the present application, judging the quality level of the ultrasound image data according to the output result of the ultrasound image quality assessment model in step S400 includes the following steps D1-D3: D1: It also includes automatic generation of targeted training plans based on quality control reports. The training plans include operation instructions and quality improvement suggestions for specific ultrasound examination types.

[0046] D2: Based on the output scores of the evaluation model, the ultrasound image data are classified into four quality levels: excellent (Grade A), good (Grade B), qualified (Grade C), and unqualified (Grade D).

[0047] D3: Differentiated processing strategies are set for different quality levels. For example, D-level images are directly marked as requiring re-examination, C-level images are recommended for clinician review, B-level images are archived as routine cases, and A-level images can be used as teaching demonstration cases.

[0048] Specifically, in D2, the four quality levels are as follows: Excellent (A, score ≥ 90 points) indicates that image quality reaches textbook examples, with all indicators reaching optimal levels; Good (B, score 75 ≤ score < 90 points) indicates that image quality fully meets clinical diagnostic requirements, with key indicators performing well; Acceptable (C, score 60 ≤ score < 75 points) indicates that image quality generally meets diagnostic requirements, but there is room for improvement; Unacceptable (D, score < 60 points) indicates that image quality does not meet basic diagnostic requirements and requires re-evaluation. The scoring criteria cover both technical quality (60%) and clinical applicability (40%) to ensure a comprehensive assessment.

[0049] Specifically, in D3, the more detailed implementation of the differentiated processing strategy includes: for Class D images, the system automatically notifies the relevant physicians and quality control personnel and requires re-examination within 24 hours; for Class C images, the system pushes them to senior physicians for review and provides specific improvement suggestions; for Class B images, the system records routine quality control data and incorporates it into the performance statistics of physicians and equipment; for Class A images, the system marks them as teaching resources and analyzes their excellent characteristics for other physicians to learn and refer to.

[0050] It should be noted that quality grading is determined not only by the absolute score but also by the examination difficulty factor. The system automatically adjusts the scoring criteria based on factors such as patient size (e.g., obesity), examination compliance, and lesion complexity to avoid unfair evaluations of difficult cases. Furthermore, the quality grading process is transparent, with the system detailing the components and deficiencies of each score, facilitating physician understanding and improvement. Furthermore, the system automatically identifies quality issues caused by equipment factors (e.g., probe aging) and distinguishes them from operational factors, ensuring accurate attribution of responsibility.

[0051] In an optional embodiment, the quality level determination in step S400 can also employ a dynamic threshold method, automatically adjusting the grading criteria based on the overall level and development goals of the medical institution. The system regularly analyzes the quality distribution of ultrasound images throughout the hospital and gradually raises the grading threshold as the overall level improves, forming a virtuous cycle of continuous improvement.

[0052] In another optional embodiment, the quality level judgment in step S400 can also be combined with a peer review mechanism, inviting internal or external experts from the institution to review some key or controversial images through an online collaborative platform to supplement the limitations of the algorithm evaluation and feedback the expert opinions to the system to optimize the evaluation model.

[0053] In the embodiment of the present application, step S500 generates a quality control report and marks the ultrasound image data that does not meet the quality standards as pending review, including the following steps E1-E6: E1: Based on the quality assessment results and historical data, a multi-dimensional quality control report is generated, including quality grade distribution statistics, problem type statistics, physician performance analysis, and equipment status assessment.

[0054] E2: Mark images with a quality grade of C and key features below a certain threshold, as well as all D-level images, as pending review and push them to the relevant responsible persons through the workflow system.

[0055] E3: Perform data visualization on the quality control report, and intuitively display the quality control status and improvement directions through charts, heat maps, trend lines, etc.

[0056] Specifically, in E1, the multi-dimensional statistical analysis of quality control reports includes: displaying the distribution of quality levels by department, physician, equipment, examination type, and other dimensions to identify weak links; counting the frequency and severity of problem types by technical problems (such as insufficient contrast, excessive noise) and operational problems (such as non-standard sections, non-standard measurements); analyzing the performance and improvement of each physician through horizontal comparison and vertical tracking; and evaluating the equipment status and maintenance needs by monitoring the quality change trends of images produced by the equipment.

[0057] Specifically, in E2, the process of marking for review includes: the system automatically identifies images that need to be reviewed and adds eye-catching marks in the PACS system and ultrasound workstation; sends notifications to responsible physicians and quality control experts through the hospital's messaging system, email, or mobile app; sets the priority and deadline for review tasks, and automatically escalates and notifies supervisors if they are not completed within the deadline; after the review is completed, the system records the comparison and improvement of the two inspection results to form a closed-loop management.

[0058] Specifically, in E3, data visualization processing adopts a responsive design, which can be presented in the best way on different terminal devices (such as computers, tablets, and mobile phones); it provides multi-level data drill-down functions, from a global overview to detailed analysis of specific cases; it supports customized reports and dashboards to meet the differentiated needs of users with different roles (such as physicians, department heads, and quality control experts); it realizes threshold warnings for key indicators, and abnormal data will automatically prompt issues that need attention in different colors or flashing.

[0059] It should be noted that the quality control report generation process utilizes an incremental update mechanism. New examination data updates relevant statistical indicators in real time, ensuring that management remains informed of quality control trends. Furthermore, the system automatically identifies unusual patterns in statistical data, such as a sudden decline in quality for a specific examination type performed by a specific physician, or persistent degradation in image quality produced by a particular device, highlighting these areas of concern in the report. Furthermore, the quality control report incorporates a benchmarking function, allowing the institution's quality control performance to be compared with the industry average or historical best performance, helping the institution clarify its direction and goals for improvement.

[0060] In an optional embodiment, the quality control report generated in step S500 can also use natural language generation technology to automatically generate text descriptions of quality control findings and improvement suggestions. The system analyzes the patterns and causes behind the quality control data and generates a professional and clear text report, making it easy for managers with non-technical backgrounds to understand the significance and value of the quality control results.

[0061] In an optional embodiment, the quality control report generated in step S500 can also use natural language generation technology to automatically generate text descriptions of quality control findings and improvement suggestions. The system analyzes the patterns and causes behind the quality control data and generates a professional and clear text report, making it easy for managers with non-technical backgrounds to understand the significance and value of the quality control results.

[0062] In another optional implementation, the pending review flag in step S500 can be integrated with a scheduling system to automatically schedule the most appropriate physician for the review. The system intelligently assigns review tasks based on physician expertise, workload, and the type of original examination question, ensuring review quality while optimizing resource utilization.

[0063] E4: Also includes trend analysis of historical quality control data, identifying the regularity and periodic changes of quality problems, and generating early warning information.

[0064] Conduct time series analysis on historical quality control data and apply statistical methods to identify seasonal fluctuations, cyclical trends and abnormal change points.

[0065] Based on the trend analysis results, a prediction model is constructed to provide early warning of possible quality risks and generate intervention recommendations.

[0066] Specifically, time series analysis methods include: using moving average method, exponential smoothing method and other techniques to deal with short-term fluctuations and extract long-term trends; applying seasonal decomposition technology to identify periodic patterns such as weekdays / weekends, day shift / night shift, and seasonal changes; using change point detection algorithm to identify mutation points of quality indicators and associate them with corresponding environmental factors (such as equipment updates, personnel changes, process reforms, etc.); identifying the lag effects of quality problems through autocorrelation analysis, such as the improvement cycle after training and the stabilization period after equipment maintenance.

[0067] Specifically, the construction and application of the prediction model include: combining time series prediction models such as ARIMA and SARIMA to predict quality trends in the future; setting dynamic thresholds based on historical fluctuation ranges to trigger early warnings when the predicted value exceeds the normal fluctuation range; generating different levels of early warning information according to the type and severity of the warning, such as general prompts, precautions, key concerns and emergency interventions; for early warning issues, the system automatically analyzes possible causes and provides targeted intervention suggestions, such as equipment preventive maintenance, physician skill enhancement training, process optimization, etc.

[0068] It should be noted that the trend analysis and early warning system employs a multi-verification mechanism to avoid false alarms due to accidental data fluctuations. The system comprehensively considers the changing patterns of multiple relevant indicators, such as quality scores, review rates, and issue types, triggering an alert only when these indicators converge on a specific trend. Furthermore, early warning information is generated using a tiered push strategy, selecting different notification channels and processing procedures based on the urgency and scope of the issue, ensuring that important issues receive timely attention and resolution. Furthermore, the system tracks the effectiveness of interventions following each alert, continuously optimizing the accuracy and practicality of the early warning model.

[0069] In an optional embodiment, the trend analysis in step S500 can also apply sequence models from deep learning, such as long short-term memory (LSTM) or gated recurrent units (GRU), to capture complex nonlinear patterns in quality data. These advanced models can handle long-term dependencies and multivariate interactions, improving the accuracy and predictability of predictions.

[0070] In another optional embodiment, the warning information in step S500 can also be integrated with a clinical decision support system. When systemic issues with the quality of a particular ultrasound examination are detected, the diagnostic reliability assessment in the relevant clinical pathway can be automatically adjusted. For example, if there is a problem with the measurement accuracy of a thyroid nodule, the system will prompt the clinician to add other auxiliary examinations or review measures to the relevant decision.

[0071] E5: It also includes establishing a quality correlation analysis between physicians and examination equipment to identify the operational advantages and disadvantages of specific physicians on specific equipment.

[0072] Construct a multidimensional data cube of physicians, equipment and quality, and identify the best and worst combinations of physicians and equipment through cross-analysis.

[0073] Based on the results of association analysis, physician grouping and equipment allocation are optimized, and personalized operation recommendations are provided for specific combinations.

[0074] The construction and analysis of the multidimensional data cube include: collecting and structuredly storing the quality scores and problem records of all examinations completed by each physician on each device; applying multivariate analysis of variance (MANOVA) to identify the impact of physician factors, device factors, and their interactions on quality; using data mining techniques such as association rule mining and cluster analysis to discover potential association patterns; calculating the "match score" of the physician-device combination for each examination type to quantify its synergistic effect; and establishing an association mapping between physician and device characteristics, such as some physicians perform better in high-frequency probe operation, while others have a greater advantage in the application of elastic imaging technology.

[0075] Specifically, the optimization allocation and personalized recommendations include: recommending the optimal physician-equipment combination for routine examinations and important cases based on the matching score; introducing physician-equipment matching as a consideration in the scheduling system to optimize resource allocation; generating customized training plans and operation guides based on physicians' operational weaknesses on specific equipment; providing physician adaptability assessments for new equipment procurement to predict training needs and potential risks; and designing a "progressive familiarization plan" to help physicians smoothly transition from equipment they are proficient in to unfamiliar equipment, thereby lowering the adaptation threshold.

[0076] It should be noted that physician-device correlation analysis not only focuses on static matching relationships but also tracks dynamic changes in these relationships. The system regularly updates analysis results to reflect the growth of physician skills and changes in device status. Furthermore, correlation analysis employs the principle of "complementary strengths," not only seeking optimal matches but also identifying complementary advantages between physicians and devices. For example, pairing physicians with precise but slow skills with devices with fast image processing speeds can optimize overall efficiency. Furthermore, correlation analysis results are presented in visual formats such as heat maps and correlation network diagrams, allowing managers to intuitively grasp the overall picture and key nodes of the physician-device relationship.

[0077] In an optional embodiment, the physician-equipment association analysis in step S500 can also employ a bio-inspired algorithm, such as a genetic algorithm or an ant colony algorithm, to solve the multi-objective optimization problem of physician-equipment allocation. By simulating evolutionary processes or swarm intelligence behavior, a globally optimal allocation solution is found while considering multiple objectives, such as maximizing quality, balancing workloads, and minimizing training costs.

[0078] In another optional implementation, the physician-device association analysis in step S500 can also incorporate expert system technology, encoding the experience and knowledge of experienced ultrasound experts into a rule base to assist in analyzing the reasons for and improvement directions for specific physician-device combinations. The system then infers these rules and generates clinically informed professional recommendations, improving the interpretability and practicality of the association analysis results.

[0079] E6: Quality control reports include quality grade distribution statistics, problem type statistics, physician performance analysis, and equipment status assessment.

[0080] Quality grade distribution statistics show the distribution of quality grades in different time periods, different examination types, and different departments, and identify quality fluctuations and trends.

[0081] Problem type statistics classify and sort common quality problems, such as unclear imaging, inaccurate measurements, missing markings, etc., to identify areas that require key improvements.

[0082] Physician performance analysis evaluates each physician's performance based on quality scores, review rates, problem types, and other dimensions, identifying best practices and training needs.

[0083] Equipment status assessment is based on the changing trends of image quality produced by each device, evaluating the performance status of the device and predicting maintenance needs and replacement timing.

[0084] Specifically, quality grade distribution statistics utilize multi-dimensional cross-analysis, allowing flexible switching and drilling down by time (daily, weekly, monthly, seasonal, and annual), location (department, hospital district), and business (examination type, disease type). The system provides various visualization methods, such as trend charts, pie charts, and stacked bar charts, to intuitively display the percentages of each grade and their changes. For time periods or business areas with quality anomalies, the system automatically labels and associates possible influencing factors, such as equipment upgrades, personnel changes, or process adjustments, to help managers understand the causes of quality fluctuations.

[0085] Specifically, a hierarchical classification system is used to analyze problem types, categorizing quality issues into broad categories: technical (e.g., inappropriate imaging parameters, improper probe selection), operational (e.g., incomplete scanning, inappropriate pressure), knowledge (e.g., missing anatomical landmarks, missed lesions), and process (e.g., non-standard labeling, incomplete documentation). Each broad category includes multiple sub-categories. The system calculates the frequency, severity, and impact of each type of problem, generating a Pareto chart or tree diagram to help identify the "critical few" issues. The system also tracks the evolving trends of problem types and evaluates the effectiveness of corrective measures.

[0086] Specifically, physician performance analysis utilizes a multi-metric evaluation system, comprehensively considering factors such as quality compliance rate, excellence rate, review rate, problem type distribution, and patient satisfaction to generate a physician's "quality profile." The system offers both horizontal comparison (comparison with physicians of the same level) and vertical comparison (comparison with a physician's own historical performance) to fairly assess a physician's relative level and progress. Based on each physician's performance characteristics, the system automatically identifies their areas of expertise (which can serve as demonstration and teaching resources) and areas of weakness (requiring focused training), supporting targeted talent development and skills improvement.

[0087] Specifically, the device status assessment establishes a health curve for device performance based on the image quality data produced by each device. The system monitors the changing trends of key indicators such as contrast resolution, spatial resolution, and signal-to-noise ratio. When it detects a sustained decline in performance beyond the normal fluctuation range, it automatically generates a maintenance alert. By analyzing the performance differences of devices across different inspection types, the system can identify specific issues such as probe aging and component degradation, providing data support for repair and replacement decisions. Furthermore, the system evaluates the utilization efficiency and scope of application of equipment, providing a reference for optimizing equipment resource allocation.

[0088] It should be noted that quality control reports utilize a layered design to meet the needs of different users: management can view a global overview and key indicators, department heads can conduct in-depth analysis of department performance and resource allocation, physicians can focus on individual performance and improvement suggestions, and equipment managers can monitor equipment status and maintenance plans. Reports also support customization, allowing users to set frequently used reports and key indicators based on their focus, improving information acquisition efficiency. Furthermore, quality control reports offer a historical snapshot function, allowing users to review quality control status at any point in time, facilitating the analysis of long-term trends and the evaluation of the effectiveness of improvement measures.

[0089] In an optional embodiment, the quality control report in step S500 can also utilize context-aware technology to automatically adjust the presentation and level of detail of the report content based on the user's identity, role, access terminal, and current work scenario. For example, when a physician briefly reviews the report on a mobile device, the system will highlight key reminders and urgent matters; while when a physician conducts in-depth research on a computer, comprehensive and detailed data analysis and interactive exploration capabilities are provided.

[0090] In another optional implementation, the quality control report in step S500 can also incorporate competitive gamification elements, stimulating physicians' enthusiasm for quality improvement through mechanisms such as quality challenges, progress leaderboards, and achievement badges. Based on the quality control results, the system automatically updates each physician's "quality achievement points" and "skill level," transforming the quality improvement process into an engaging professional growth experience and increasing physician engagement.

[0091] In summary, the medical ultrasound imaging quality control management method proposed in this embodiment obtains multi-source ultrasound imaging data, extracts objective and professional features of the images, applies advanced quality assessment models for intelligent rating, generates comprehensive quality control reports, and implements the marking, early warning, and improvement tracking of quality issues. The system not only provides static quality assessments, but also implements dynamic quality management and early warning through trend analysis, correlation analysis, and predictive models. Through personalized training plans and optimized matching of physicians and equipment, the system forms a closed-loop management process of quality control, improvement, and repeated quality control, effectively improving the overall quality level and diagnostic accuracy of ultrasound examinations in medical institutions, providing more reliable imaging support for clinical decision-making, while optimizing the allocation of medical resources and improving the efficiency and pertinence of quality control work. Example

[0092] The above is a schematic diagram of a medical image quality control management method. It should be noted that the technical solution of this medical image quality control management system and the technical solution of the medical image quality control management method described above are based on the same concept. For details not described in detail in the technical solution of the medical image quality control management system in this embodiment, please refer to the description of the technical solution of the medical image quality control management method described above.

[0093] This embodiment also provides a medical image quality control management system, including: An image data acquisition module, used to acquire ultrasound image data within a target medical institution; A feature extraction module, used for extracting features from the acquired ultrasound image data; A quality assessment module is used to input the extracted features into a preset ultrasound image quality assessment model and determine the quality level of the ultrasound image data based on the output results of the ultrasound image quality assessment model; The report generation module is used to generate quality control reports and mark ultrasound image data that does not meet the quality standards as pending review.

[0094] This embodiment also provides an electronic device suitable for medical image quality control management, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the medical image quality control management method proposed in the above embodiment.

[0095] This embodiment further provides a storage medium having a computer program stored thereon. When the program is executed by a processor, the method for implementing the medical image quality control management method proposed in the above embodiment is implemented.

[0096] The storage medium proposed in this embodiment and the method for implementing medical image quality control management proposed in the above embodiment belong to the same inventive concept. For technical details not described in detail in this embodiment, please refer to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0097] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented with the help of software and necessary general-purpose hardware, and of course can also be implemented with hardware. Based on this understanding, the technical solution of the present invention, or the part that contributes to the existing technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various embodiments of the present invention.

[0098] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A medical image quality control management method, characterized by: Including, obtaining ultrasound imaging data in the target medical institution; Performing feature extraction on the acquired ultrasound image data; Inputting the extracted features into a preset ultrasound image quality assessment model; Determining the quality level of the ultrasound image data according to an output result of the ultrasound image quality assessment model; Generate quality control reports and mark ultrasound image data that does not meet quality standards as pending review.

2. The medical image quality control management method according to claim 1, wherein: The ultrasound image data includes: ultrasound image data of different examination parts, ultrasound image data acquired by different examination equipment, ultrasound image data acquired by different examination physicians, and ultrasound image data of different pathological conditions.

3. The medical image quality control management method according to claim 2, wherein: The quality parameters evaluated by the ultrasound image quality assessment model include at least: image clarity, contrast, signal-to-noise ratio, and integrity of key anatomical structures.

4. The medical image quality control management method according to claim 3, wherein: It also includes automatic generation of targeted training plans based on quality control reports, which include operating instructions and quality improvement suggestions for specific ultrasound examination types.

5. The medical image quality control management method according to claim 4, wherein: It also includes trend analysis of historical quality control data, identifying the regularity and periodic changes of quality problems, and generating early warning information.

6. The medical image quality control management method according to claim 5, wherein: It also includes establishing a quality correlation analysis between physicians and examination equipment to identify the operational advantages and disadvantages of specific physicians on specific equipment.

7. The medical image quality control management method according to claim 6, wherein: The quality control report includes quality grade distribution statistics, problem type statistics, physician performance analysis and equipment status assessment.

8. A medical image quality control management system, based on the medical image quality control management method according to any one of claims 1 to 7, characterized in that: It also includes an image data acquisition module for acquiring ultrasound image data in a target medical institution; A feature extraction module, used for extracting features from the acquired ultrasound image data; a quality assessment module, configured to input the extracted features into a preset ultrasound image quality assessment model, and determine the quality level of the ultrasound image data based on an output result of the ultrasound image quality assessment model; The report generation module is used to generate quality control reports and mark ultrasound image data that does not meet the quality standards as pending review.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the medical image quality control management method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the medical image quality control management method according to any one of claims 1 to 7 are implemented.

Images