Universal parent platform adaptive image processing method and device for positioning type surgical robot

Abstract

The embodiment of the application discloses a positioning type surgical robot general mother platform adaptive image processing method and device. The positioning type surgical robot general mother platform adaptive image processing method comprises the following steps: acquiring a medical image and a task type of a medical image processing task to be executed; performing feature extraction on the medical image, and obtaining an optimal execution algorithm according to the extracted medical image features, the task type and a pre-constructed algorithm knowledge base, wherein the algorithm knowledge base is used for storing at least one algorithm and corresponding algorithm metadata; and processing the medical image based on the optimal execution algorithm to obtain an image processing result. The embodiment of the application can improve the efficiency of the image processing process and the reliability of the image processing result.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to an adaptive image processing method and apparatus for a general-purpose mother platform for positioning surgical robots. Background Technology

[0002] With the rapid development of medical artificial intelligence technology, a large number of specialized algorithm models have emerged for various medical image analysis tasks, including but not limited to organ segmentation models, lesion segmentation models, key point detection models, rigid registration models, non-rigid registration models, and structure recognition models. In actual clinical and research applications, the same medical institution or medical imaging artificial intelligence platform usually needs to deploy dozens or even hundreds of image processing algorithms to adapt to the medical image processing needs of different imaging modalities (e.g., computed tomography (CT), magnetic resonance imaging (MRI), and ultrasound), different anatomical locations, and different analysis tasks.

[0003] To meet the above requirements, in related technologies, the various algorithms within the system are deployed in a scattered manner as independent modules. Typically, the algorithm is manually selected and invoked to complete the processing task. In this case, the image processing efficiency is low, and the resulting image processing results are unreliable. Summary of the Invention

[0004] This application provides an adaptive image processing method and apparatus for a universal mother platform of a positioning surgical robot, which solves the technical problems of low efficiency and poor reliability of existing medical image processing processes.

[0005] In a first aspect, embodiments of this application provide an adaptive image processing method for a general-purpose mother platform of a positioning surgical robot. This adaptive image processing method for a general-purpose mother platform of a positioning surgical robot includes: acquiring medical images and the task type of the medical image processing task to be performed; Feature extraction is performed on medical images, and the optimal execution algorithm is obtained based on the extracted medical image features, task type, and pre-built algorithm knowledge base. The algorithm knowledge base is used to store at least one algorithm and its corresponding algorithm metadata. The optimal execution algorithm is used to process medical images to obtain image processing results.

[0006] As described above, feature extraction is performed on the acquired medical images. Based on the medical image features, the pre-defined task type of the medical image processing task to be performed, and a pre-built algorithm knowledge base, the optimal execution algorithm is obtained. This algorithm is then used to process the medical images to obtain the image processing results. The pre-built algorithm knowledge base allows for direct selection of the optimal execution algorithm, improving the efficiency of the medical image processing process and the reliability of the obtained image processing results.

[0007] The optimal execution algorithm is obtained based on the extracted medical image features, task type, and pre-built algorithm knowledge base, including: Based on medical image features and task type, a set of candidate algorithms is obtained by screening from a pre-built algorithm knowledge base. The set of candidate algorithms includes at least one candidate algorithm that satisfies the medical image processing task to be performed. Candidate algorithms are evaluated according to preset scoring dimensions, and the scores are sorted in descending order. The candidate algorithm with the highest score is selected as the optimal execution algorithm.

[0008] The task types include at least one of the following: Medical image segmentation task, medical image registration task, medical image key point detection task, medical image structure recognition task.

[0009] The scoring dimensions include at least one of the following: The matching degree between the candidate algorithm and the medical image features, the algorithm accuracy of the candidate algorithm, the computation time of the candidate algorithm, and the resource consumption of the candidate algorithm during operation.

[0010] The medical images are processed based on the optimal execution algorithm to obtain the image processing results, including: The medical images are processed based on the optimal execution algorithm to obtain the initial image processing results; The corresponding evaluation strategy is determined according to the task type, and the initial image processing results are evaluated based on the evaluation strategy to obtain the evaluation results; The evaluation results are compared with the preset scoring threshold, and the target image processing results are obtained based on the comparison results.

[0011] The evaluation results are compared with a preset scoring threshold, and the target image processing results are obtained based on the comparison results, including: If the evaluation result is greater than or equal to the preset scoring threshold, the initial image processing result will be used as the target image processing result. If the evaluation result is less than the preset scoring threshold, the defect type of the initial image processing result is determined based on the evaluation result, and the corresponding optimization strategy is matched according to the defect type. The image processing workflow is optimized based on the optimization strategy, and the medical images are reprocessed based on the optimized image processing workflow to obtain the optimized image processing results. The optimized image processing results are evaluated until the preset termination conditions are met, and the target image processing result is obtained.

[0012] The methods also include: Retrieve the newly added algorithm and its corresponding algorithm metadata; Deploy the new algorithm into the current algorithm execution environment and standardize the interface configuration of the new algorithm to make the new algorithm callable; According to the preset field rules of the algorithm knowledge base, the algorithm metadata of the new algorithm is stored in the algorithm knowledge base to obtain the updated algorithm knowledge base; Accordingly, feature extraction is performed on medical images, and based on the extracted medical image features, task type, and pre-built algorithm knowledge base, the optimal execution algorithm is obtained, including: Feature extraction is performed on medical images, and the optimal execution algorithm is obtained based on the extracted medical image features, task type, and updated algorithm knowledge base.

[0013] Secondly, embodiments of this application provide an adaptive image processing device for a universal mother platform of a positioning surgical robot. This adaptive image processing device for a universal mother platform of a positioning surgical robot includes: a task acquisition module, used to acquire medical images and the task type of the medical image processing task to be performed; The algorithm determination module is used to extract features from medical images and obtain the optimal execution algorithm based on the extracted medical image features, task type and pre-built algorithm knowledge base. The algorithm knowledge base is used to store at least one algorithm and its corresponding algorithm metadata. The image processing module is used to process medical images based on the optimal execution algorithm to obtain the image processing results.

[0014] Thirdly, embodiments of this application provide an electronic device, which includes: One or more processors; Memory, used to store one or more computer programs; When one or more computer programs are executed by one or more processors, electronic devices implement an adaptive image processing method for a general-purpose mother platform of a positioning surgical robot, as described in the first aspect.

[0015] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the adaptive image processing method for a general-purpose mother platform of a positioning surgical robot as described in any of the first aspects. Attached Figure Description

[0016] Figure 1 A flowchart illustrating an adaptive image processing method for a universal mother platform of a positioning surgical robot, provided in this application embodiment; Figure 2A flowchart illustrating another adaptive image processing method for a universal mother platform of a positioning surgical robot provided in this application embodiment; Figure 3 A schematic diagram of the architecture of a universal master platform for a positioning surgical robot provided in an embodiment of this application; Figure 4 A schematic diagram of an adaptive image processing device for a universal mother platform of a positioning surgical robot provided in this application embodiment; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0017] The following description and accompanying drawings fully illustrate specific embodiments of this application to enable those skilled in the art to practice them. The embodiments represent only possible variations. Individual components and functions are optional unless explicitly required, and the order of operation may vary. Parts and features of some embodiments may be included in or replace parts and features of other embodiments. The scope of embodiments of this application includes the entire scope of the claims and all available equivalents of the claims. In this document, each embodiment may be referred to individually or collectively by the term "invention," which is merely for convenience and is not intended to automatically limit the scope of the application to any single invention or inventive concept if more than one invention is disclosed. Relational terms such as "first" and "second" are used herein only to distinguish one entity or operation from another, without requiring or implying any actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed. The various embodiments in this document are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the structures, products, etc., disclosed in the embodiments, since they correspond to the disclosed parts, the descriptions are relatively simple; relevant details can be found in the method section.

[0018] Current medical image processing platforms primarily employ fixed workflows for image analysis. When performing specific tasks, users need to manually specify the particular algorithm model and adjust relevant parameters based on experience. For example, in brain MRI segmentation tasks, users need to manually select a brain tissue segmentation model; in image registration tasks, they need to manually select either a rigid or non-rigid registration algorithm based on the image characteristics; and in keypoint detection tasks, the corresponding model also needs to be determined manually.

[0019] In practical applications, the same medical institution or artificial intelligence platform usually deploys dozens or even hundreds of image processing algorithms to process medical image data of different modalities, different parts and different tasks. However, the algorithm calling process usually relies on pre-configured rules or human experience, lacks the ability to automatically analyze the features of the input medical images, lacks an automatic quality evaluation and feedback optimization mechanism for the algorithm output results, and lacks collaboration because each algorithm module runs independently.

[0020] In implementing the aforementioned technologies, the inventors discovered that the sources of actual medical images are complex. Image data generated by different devices, scanning protocols, and hospitals exhibit significant differences in resolution, grayscale distribution, noise levels, and field of view. Even for the same processing task, different data may require different algorithms to obtain optimal results. Therefore, relying solely on human experience to select algorithms can easily lead to a mismatch between the algorithm and the data, thus affecting the final processing outcome.

[0021] In addition, the inventors also found that the system deployed the algorithm independently and lacked unified management of its applicable conditions, input requirements and performance characteristics when implementing the above-mentioned related technologies. This caused the system to be unable to automatically match the optimal algorithm according to the medical image features, which increased the maintenance and expansion costs.

[0022] Furthermore, the inventors discovered that most processing steps directly output results after execution, lacking a quality assessment step. This prevents the system from automatically identifying issues such as missing segmentation targets, boundary errors, registration failures, or keypoint misalignment, severely impacting the reliability of clinical applications.

[0023] Furthermore, the inventors discovered during the implementation of the aforementioned related technologies that the systems typically execute only once. When the results are unsatisfactory, users must manually adjust and retry, which is not only cumbersome and inefficient but also leads to inconsistent results due to differences in operator experience. The system lacks a closed-loop feedback mechanism that automatically switches algorithms, adjusts parameters, or optimizes strategies based on the quality of the results.

[0024] Furthermore, the inventors discovered during the implementation of the aforementioned technologies that each algorithm module typically operates independently, lacking coordination between modules. When a single algorithm fails, the system cannot automatically call upon alternative solutions; when multiple algorithms can perform the task, the results cannot be comprehensively evaluated and optimized, resulting in underutilization of algorithm resources and limiting the overall processing accuracy and stability.

[0025] In view of the above technical problems, and based on the analysis of the underlying causes of the technical problems, the inventor proposes an adaptive image processing method for a general mother platform of a positioning surgical robot, which performs feature extraction on the acquired medical image, and obtains the optimal execution algorithm according to the medical image features, the set task type of the medical image processing task to be executed and the pre-constructed algorithm knowledge base, then processes the medical image through the algorithm to obtain an image processing result. Through the pre-constructed algorithm knowledge base, the optimal execution algorithm can be directly screened and obtained therefrom, which improves the efficiency of the medical image processing process and the reliability of the obtained image processing result.

[0026] Please refer to Figure 1 , which is a flow chart of an adaptive image processing method for a general mother platform of a positioning surgical robot provided in an embodiment of the present application, as shown in Figure 1 , the adaptive image processing method for a general mother platform of a positioning surgical robot includes, but is not limited to, S101-S103: S101, acquiring a medical image and a task type of a medical image processing task to be executed.

[0027] Wherein, the positioning surgical robot refers to a minimally invasive surgery auxiliary robot that takes three-dimensional reconstruction of preoperative or intraoperative medical images, precise positioning of lesions and puncture paths as core functions, and is equipped with a multi-degree-of-freedom mechanical execution arm, an optical or electromagnetic navigation positioning unit, and image planning software; the general mother platform refers to an integrated software and hardware basic support platform built for multi-model and multi-procedure positioning surgical robots. Medical images refer to human tissue or organ image data acquired by medical imaging equipment, which can include, but are not limited to, computed tomography (CT), magnetic resonance imaging (MRI), cone beam computed tomography (CBCT), computed tomography angiography (CTA), positron emission tomography (PET), ultrasound images, etc.; a medical image processing task refers to an operation of analyzing and processing a medical image to achieve a specific goal, and a task type refers to a specific classification or name of a medical image processing task.

[0028] In the embodiment of the present application, medical images can be collected by intraoperative scanning equipment such as CT, CBCT, and MRI matched with the positioning surgical robot, and uniformly transmitted to the designed general mother platform, and the user can select a predefined task type through a user interface displayed on the general mother platform, without specifying a specific algorithm model.

[0029] In the embodiment of the present application, the task type may include at least one of the following: a medical image segmentation task, a medical image registration task, a medical image key point detection task, and a medical image structure recognition task.

[0030] Specifically, the purpose of medical image segmentation is to classify each pixel in an image into a specific category and accurately delineate the boundaries of the target region, thereby achieving pixel-level precise region division; the purpose of medical image registration is to align images from different sources to a unified spatial coordinate system and eliminate geometric differences, thereby solving the spatial consistency problem of multimodal and multi-temporal images; the purpose of medical image keypoint detection is to locate key points of specific anatomical structures in an image, thereby identifying landmark coordinates with clear anatomical significance; the purpose of medical image structure recognition is to identify and classify specific anatomical structures or pathological features in an image, thereby distinguishing the structures present in the image. In addition, other medical image analysis tasks may also be included, but this application embodiment does not specifically limit these.

[0031] S102, extract features from the medical image, and obtain the optimal execution algorithm based on the extracted medical image features, task type, and pre-built algorithm knowledge base. The algorithm knowledge base is used to store at least one algorithm and its corresponding algorithm metadata.

[0032] Among them, medical image features refer to the set of indicators extracted from the input image that can quantify its technical attributes and content characteristics; the algorithm knowledge base is a structured dynamic database used to store the metadata of all available algorithms and their applicable conditions mapping relationship. Algorithm metadata refers to a standardized information set describing the applicable boundaries and performance characteristics of the algorithm; the optimal execution algorithm refers to the algorithm that satisfies the constraints of adaptability priority, optimal performance and optimal reliability under the current medical image features and task type constraints.

[0033] Specifically, the algorithm metadata stored in the algorithm knowledge base may include, but is not limited to: algorithm name, used to uniquely identify the corresponding image processing algorithm; task type, used to indicate the processing work that the algorithm is adapted to, such as medical image segmentation task, medical image registration task, medical image key point detection task, and medical image structure recognition task; supported image modalities, referring to the types of imaging that the algorithm can process, such as CT, MRI, CBCT, CTA, PET, and ultrasound; supported anatomical locations, referring to the human anatomical regions that the algorithm is adapted to, such as head, brain, chest, abdomen, bones, and face; input requirements, used to specify the parameter restrictions on image size, voxel spacing, slice thickness, signal-to-noise ratio, and enhancement state during normal algorithm inference; output type, used to indicate the output result after the algorithm operation, such as the segmentation task outputting a segmentation mask, the key point detection task outputting key point coordinates, and the registration task outputting a deformation field; and recommended priority, referring to the initial priority given by the algorithm's offline test performance, which can be dynamically updated based on historical processing results and used for subsequent comprehensive algorithm evaluation.

[0034] For example, the brain MRI nucleus segmentation model supports MRI as the image modality, the brain as the anatomical location, and medical image segmentation as the task type, with an initial recommendation priority of high.

[0035] In this embodiment, after a user uploads any medical image (e.g., CT, MRI, and ultrasound), a dedicated image feature analysis module will run automatically, eliminating the need for manual input of image parameters. The attribute information directly extracted from the medical image may include, but is not limited to: imaging modality information, referring to the imaging equipment and principles corresponding to the medical image, such as CT, MRI, CBCT, CTA, PET, and ultrasound; anatomical location information, referring to the human tissue region corresponding to the image, such as the head, brain, chest, abdomen, bones, and face; spatial information, referring to the three-dimensional geometric parameters of the image, such as image size, voxel spacing, image orientation, and field of view; image quality information, referring to image imaging quality indicators, such as signal-to-noise ratio, contrast, blurriness, and the degree of metal artifacts; and scanning parameter information, referring to the specific parameters set by the device during scanning, such as MRI sequence type, CT slice thickness, and enhanced scanning status. After extracting the above attribute information, it can be integrated into a structured feature data package, i.e., medical image features.

[0036] In this embodiment, all inherent features of the input medical image are first parsed, and combined with the processing task selected by the user, a pre-built algorithm knowledge base storing all algorithms and standardized attributes is called to output the optimal execution algorithm adapted to the current image.

[0037] In one optional implementation, the process of obtaining the optimal execution algorithm based on the extracted medical image features, task type, and pre-built algorithm knowledge base can be as follows: Based on the medical image features and task type, a set of candidate algorithms is selected from the pre-built algorithm knowledge base, wherein the set of candidate algorithms includes at least one candidate algorithm that satisfies the medical image processing task to be executed; the candidate algorithms are evaluated according to a preset scoring dimension, the obtained scores are sorted in descending order, and the candidate algorithm with the highest score is selected as the optimal execution algorithm.

[0038] Among them, the candidate algorithm set refers to the set of algorithms selected from the algorithm knowledge base based on medical image features and task type, which are suitable for the current image and can complete the corresponding processing task; the scoring dimension refers to the preset evaluation index used to quantitatively evaluate the comprehensive performance of the candidate algorithms.

[0039] In this embodiment, the scoring dimensions include at least one of the following: the degree of matching between the candidate algorithm and the medical image features, the algorithm accuracy of the candidate algorithm, the computation time of the candidate algorithm, and the resource consumption of the candidate algorithm during operation. Specifically, the degree of matching between the candidate algorithm and the medical image features measures the fit between the algorithm's own adaptation conditions and the various features of the medical image to be processed; the algorithm accuracy of the candidate algorithm represents the accuracy with which the algorithm completes the corresponding task; the computation time of the candidate algorithm refers to the estimated time required for the algorithm to process a medical image of the same size; and the resource consumption of the candidate algorithm during operation refers to the peak hardware resource consumption during the algorithm's operation.

[0040] In this embodiment, the extracted medical image features and the user-specified processing task are used as screening conditions. All algorithms that meet the processing requirements are selected from the pre-built algorithm knowledge base to form a candidate set. Then, each algorithm in the set is comprehensively scored using a preset multi-dimensional evaluation index. The algorithms are sorted from high to low according to their scores, and the candidate algorithm with the highest comprehensive score is selected as the optimal execution algorithm for this medical image processing.

[0041] S103, the medical image is processed based on the optimal execution algorithm to obtain the image processing result.

[0042] The image processing result refers to the structured data corresponding to the task output by the optimal execution algorithm after processing medical images.

[0043] In this embodiment, the optimal matching algorithm is obtained after automatic screening and comprehensive scoring and sorting. The algorithm performs a complete image processing inference process on the original medical image to be processed and outputs the corresponding initial image processing results for subsequent automated quality evaluation.

[0044] In one optional implementation, the above-mentioned processing of medical images based on the optimal execution algorithm to obtain image processing results can be as follows: processing medical images based on the optimal execution algorithm to obtain initial image processing results; determining the corresponding evaluation strategy according to the task type, and evaluating the initial image processing results based on the evaluation strategy to obtain evaluation results; comparing the evaluation results with a preset scoring threshold, and obtaining the target image processing result based on the comparison results.

[0045] The initial image processing result refers to the raw, unoptimized output data of the medical image to be processed after inputting it into the optimal execution algorithm and performing inference operations. The evaluation strategy refers to the pre-configured evaluation index system and calculation rules that correspond one-to-one with the task type. The evaluation result refers to the comprehensive quantitative score obtained after calculating the initial image processing result according to the selected evaluation strategy. The preset scoring threshold refers to the qualified judgment threshold pre-set in combination with clinical usage requirements. Different thresholds can be set according to different task scenarios. For example, the threshold is set higher for high-precision diagnosis scenarios and appropriately lowered for routine screening scenarios.

[0046] In the embodiments of this application, since different task types such as segmentation, key point detection and image registration have different focuses, it is impossible to use a single evaluation index to measure all tasks. Therefore, different evaluation strategies are adopted for different tasks.

[0047] Optionally, for image segmentation tasks, the following evaluation metrics can be used: connected component integrity, used to determine whether the segmented region is a continuous whole; anatomical structure rationality, used to determine whether the segmentation result conforms to common sense of human anatomy; boundary continuity, used to determine whether the segmentation contour edges are smooth; segmentation volume range, used to determine whether the size of the segmented organ or lesion is within the normal range; and model confidence, used to determine the credibility of the model in determining that each pixel belongs to the target region.

[0048] Optionally, for the key point detection task, the following evaluation indicators can be used: key point location rationality, which is used to determine whether the detected anatomical key points fall on the corresponding anatomical locations; anatomical constraint consistency, which is used to determine whether the positional constraint relationship between each key point conforms to the corresponding human anatomy; and key point distance relationship, which is used to determine whether the distance between adjacent key points is within the normal range.

[0049] Optionally, the following evaluation metrics can be used for the registration task: image similarity metric, used to determine the similarity of the grayscale distribution between the image to be registered and the reference image after registration; structural overlap metric, used to determine the proportion of overlap of the same anatomical tissue after the two images are aligned; deformation field smoothness metric, used to determine whether the displacement field changes smoothly during the image deformation process; and registration error metric, used to determine the actual offset distance of the anatomical structure.

[0050] In this embodiment, the optimal execution algorithm obtained through screening is used to process medical images to obtain initial image processing results. A corresponding evaluation strategy is then matched based on the currently executed image processing task type. The initial image processing results are then evaluated using the evaluation strategy, and a quantitative evaluation result is output. The evaluation result is compared with a preset scoring threshold. Based on the comparison result, it is determined whether the initial result is qualified, and the corresponding operation is selected to finally obtain the target image processing result.

[0051] In one optional implementation, the comparison of the evaluation result with a preset scoring threshold and the determination of the target image processing result based on the comparison result can be as follows: if the evaluation result is greater than or equal to the preset scoring threshold, the initial image processing result is taken as the target image processing result; if the evaluation result is less than the preset scoring threshold, the defect type of the initial image processing result is determined based on the evaluation result, and a corresponding optimization strategy is matched according to the defect type; the image processing flow is optimized based on the optimization strategy, and the medical image is reprocessed based on the optimized image processing flow to obtain the optimized image processing result; the optimized image processing result is evaluated until a preset termination condition is met to obtain the target image processing result.

[0052] Among them, the defect type is used to represent the intrinsic reasons for the unsatisfactory image processing results. It is based on the scoring items of multi-dimensional evaluation indicators and is obtained by classifying the problems existing in the initial image processing results. It can include, but is not limited to: parameter adaptation defects, input image defects, algorithm adaptation defects, processing paradigm defects, and single algorithm limitation defects. The optimization strategy refers to setting corresponding improvement schemes for different defect types. It can include, but is not limited to: automatic switching of candidate algorithms, automatic adjustment of algorithm parameters, automatic change of processing strategy, automatic addition of preprocessing steps, and fusion of multiple algorithm results. In addition, a mapping relationship between defect types and optimization strategies can be established in advance to adjust the image processing flow. The image processing flow is used to perform inference calculations on medical images and generate processing results. It can be composed of image preprocessing, selected execution algorithm, internal algorithm running parameters, and post-processing. The preset termination conditions include the evaluation result corresponding to the optimized image processing result being greater than or equal to the preset scoring threshold and the total number of iterations reaching the preset maximum number of iterations. In practice, the loop optimization can be terminated if any of the above conditions are met.

[0053] In some cases, if the target image is still not achieved even after reaching the maximum number of iterations, the image processing result with the highest score during the iteration process will be taken as the target image processing result.

[0054] Optionally, if the current algorithm has poor compatibility with the current image, and adjusting the parameters cannot improve the effect (i.e., algorithm compatibility defect), an optimization strategy of automatically switching candidate algorithms can be selected. Automatic switching of candidate algorithms means abandoning the currently selected optimal execution algorithm and selecting other candidate algorithms from the candidate algorithm set according to their ranking as the new execution algorithm.

[0055] Optionally, if the overall result is generally reasonable but lacks certain details (i.e., parameter fitting defects), an optimization strategy of automatically adjusting algorithm parameters can be chosen. Automatic parameter adjustment refers to fine-tuning the original algorithm by modifying only the hyperparameters within the algorithm itself.

[0056] Optionally, when the single-stage processing mode has limited effectiveness and requires step-by-step processing (i.e., processing paradigm defects), an optimization strategy that automatically changes the processing strategy can be selected. Automatically changing the processing strategy refers to adjusting the overall processing paradigm without changing individual algorithms, thus altering the execution logic of image processing.

[0057] Optionally, if imaging issues lead to poor results (i.e., input image defects), such as excessive image noise, insufficient contrast, severe metallic artifacts, or resolution mismatch, an optimization strategy that automatically adds preprocessing steps can be selected. Automatically adding preprocessing steps refers to an optimization strategy that adds an extra preprocessing step before the image is fed into the algorithm for inference, without modifying the algorithm itself, thus improving the quality of the input image.

[0058] Optionally, if each individual algorithm has its own advantages and disadvantages, and no single algorithm can meet the passing standard (i.e., the limitations and defects of a single algorithm), an optimization strategy of fusing the results of multiple algorithms can be chosen. Fusing the results of multiple algorithms refers to simultaneously calling multiple algorithms from the candidate algorithm set to generate processing results, and then using a specific strategy to combine the multiple results into the final result. This is a fallback optimization method.

[0059] In this embodiment, the evaluation result is compared with a preset scoring threshold. If the result meets the threshold, the initial image processing result is directly output. If the result does not meet the threshold, the defect type is located based on the evaluation result, and the corresponding optimization strategy is matched. The image processing flow is adjusted according to the optimization strategy, and the medical image is reprocessed. The steps from result evaluation to flow optimization are repeated until the evaluation result is qualified or the number of iterations reaches the upper limit, and finally the target image processing result is output.

[0060] In this embodiment, after the entire image processing process is completed, in addition to outputting a qualified target image processing result, the key running data during this execution process will be saved to the algorithm knowledge base, which may include, but is not limited to: current medical image features, referring to structured information such as imaging modality, anatomical location, spatial parameters, image quality, and scanning parameters; selected algorithm, referring to the optimal execution algorithm finally adopted; parameter configuration, referring to the algorithm hyperparameters, preprocessing settings, and processing strategy configuration determined after optimization; processing result quality score, referring to the quality score obtained in the last round of evaluation, used to measure the actual performance of the final algorithm and corresponding parameters on this type of medical image; and final processing effect, referring to the final generated segmentation mask, registration image, and key point coordinates and other result data.

[0061] The above-mentioned key operational data during this execution process is saved to the algorithm knowledge base. Instead of relying solely on the inherent indicators obtained from testing on public datasets during the algorithm development phase, the knowledge base is supplemented with real-world usage data from actual clinical scenarios. This makes the recommendation priorities in the algorithm metadata no longer fixed. With the accumulation of a large amount of historical operational data, the priority of each algorithm and the recommendation strategy for algorithm scoring are dynamically adjusted. The algorithm selection criteria are corrected based on the actual performance in real-world scenarios, continuously improving the accuracy of subsequent optimal algorithm selection.

[0062] Furthermore, in the case of adding a new image processing algorithm, in one optional implementation, the new algorithm and its corresponding algorithm metadata are obtained; the new algorithm is deployed to the current algorithm execution environment, and the interface of the new algorithm is standardized to make the new algorithm callable; according to the preset field rules of the algorithm knowledge base, the algorithm metadata of the new algorithm is stored in the algorithm knowledge base to obtain the updated algorithm knowledge base. Accordingly, features can be extracted from medical images, and the optimal execution algorithm can be obtained based on the extracted medical image features, task type, and the updated algorithm knowledge base.

[0063] Among them, "newly added algorithm" refers to medical image processing algorithms that are later added to the system; "current algorithm execution environment" refers to the hardware and software environment of the platform used to run image processing algorithms; and "preset field rules" refers to the pre-defined metadata storage specifications, which stipulate the fixed fields and filling format of each piece of algorithm metadata stored in the algorithm knowledge base.

[0064] In this embodiment, the newly added algorithm to be integrated and its corresponding metadata are first obtained. Then, the new algorithm is deployed to the system's algorithm execution environment. The algorithm is modified according to a unified calling protocol, and its input / output format is standardized to ensure the scheduling module can call the algorithm correctly. Finally, the algorithm metadata is stored in the algorithm knowledge base according to unified fields, and the algorithm knowledge base is updated. This embodiment requires no modification to the main logic of the algorithm scheduling module, making expansion and maintenance simpler.

[0065] As described above, feature extraction is performed on the acquired medical images. Based on the medical image features, the pre-defined task type of the medical image processing task to be performed, and a pre-built algorithm knowledge base, the optimal execution algorithm is obtained. This algorithm is then used to process the medical images to obtain the image processing results. The pre-built algorithm knowledge base allows for direct selection of the optimal execution algorithm, improving the efficiency of the medical image processing process and the reliability of the obtained image processing results.

[0066] Based on the above, Figure 2A flowchart of another adaptive image processing method for a universal mother platform of a positioning surgical robot provided in this application embodiment is shown below. Figure 2 As shown, the method includes: S201, Task type for acquiring medical images and the medical image processing task to be performed.

[0067] The task types include at least one of the following: medical image segmentation task, medical image registration task, medical image key point detection task, and medical image structure recognition task.

[0068] S202, extract features from medical images to obtain medical image features.

[0069] In this embodiment of the application, medical images are analyzed to extract image features such as imaging modality, anatomical location, image quality, and spatial parameters to form structured feature information, namely medical image features.

[0070] S203, based on medical image features and task type, selects a set of candidate algorithms from a pre-built algorithm knowledge base.

[0071] S204: Evaluate the candidate algorithms according to the preset scoring dimensions and sort the obtained scores in descending order.

[0072] The scoring dimensions include at least one of the following: the degree of matching between the candidate algorithm and the medical image features, the algorithm accuracy of the candidate algorithm, the computation time of the candidate algorithm, and the resource consumption of the candidate algorithm during operation.

[0073] S205: Select the candidate algorithm with the highest score as the optimal execution algorithm.

[0074] S206, The medical image is processed based on the optimal execution algorithm to obtain the initial image processing result.

[0075] S207, Determine the corresponding evaluation strategy based on the task type.

[0076] In this embodiment of the application, each type of task has its own specific evaluation criteria, and the corresponding evaluation index is selected based on the current task type.

[0077] S208, evaluate the initial image processing results based on the evaluation strategy to obtain the evaluation results.

[0078] S209, determine whether the evaluation result is greater than or equal to the preset scoring threshold. If yes, proceed to S210; otherwise, proceed to S211.

[0079] S210, output the target image processing result.

[0080] S211, determine whether the total number of iterations has reached the preset maximum number of iterations. If yes, execute S210; otherwise, execute S212.

[0081] In this embodiment of the application, if the target is still not met when the total number of iterations reaches the maximum number of iterations, the image processing result with the highest score during the iteration process is taken as the target image processing result.

[0082] S212, Based on the evaluation results, determine the defect type of the initial image processing result, and match the corresponding optimization strategy according to the defect type.

[0083] In this embodiment of the application, a one-to-one mapping relationship between defect types and optimization strategies is pre-built so as to match the corresponding optimization method according to the identified defect type.

[0084] S213, optimize the image processing flow based on the optimization strategy, and re-execute S206 based on the optimized image processing flow.

[0085] The above Figure 2 The implementation steps of the adaptive image processing method for the universal master platform of the positioning surgical robot of this application are fully illustrated. Based on the same inventive concept, this application also provides a system architecture for implementing the image processing method. Figure 3 This application provides an embodiment of a general-purpose master platform for a positioning surgical robot, as shown in the following diagram. Figure 3 As shown, the system includes: The input module 301 is used to receive intraoperative medical images and surgical task instructions transmitted by each positioning surgical robot, and send the received medical images and task types to the image feature analysis module 302.

[0086] The image feature analysis module 302 performs feature extraction on the received medical images to obtain imaging modality information, anatomical region information, three-dimensional spatial information, imaging quality information, and scanning configuration information. It organizes these into standardized medical image features and transmits the image features and task type to the algorithm scheduling module 303.

[0087] The algorithm scheduling module 303 has a built-in general-purpose mother platform shared algorithm knowledge base. Based on the received medical image features and task type, it selects a set of candidate algorithms, scores and sorts the candidate algorithms according to preset scoring dimensions, selects the optimal execution algorithm, schedules the optimal execution algorithm to complete medical image inference, outputs the initial image processing results and sends them to the result quality evaluation module 304.

[0088] The result quality evaluation module 304 receives the initial image processing result, selects the corresponding evaluation strategy based on the current task type to score it, outputs the quantified evaluation result, and then compares the evaluation result with the preset scoring threshold. If the evaluation result is less than the preset scoring threshold, the score loss information of the evaluation indicator is sent to the feedback optimization module 305; if the evaluation result is qualified, the qualified image processing result is transmitted to the output module 306.

[0089] The feedback optimization module 305 receives evaluation data from the result quality evaluation module, determines the defect type based on the score loss, matches the corresponding optimization strategy based on the defect type, and sends an optimization instruction to the algorithm scheduling module 303. The algorithm scheduling module 303 then adjusts the image processing flow and reprocesses the medical image according to the optimization strategy. The medical image processing results are continuously optimized until the preset termination conditions are met, and the final determined target image processing result is sent to the output module 306.

[0090] The output module 306 receives the qualified target image processing results and sends the final target image processing results to the navigation control unit of the corresponding positioning surgical robot to support the localization of the surgical target. In addition, it can also summarize information such as medical image features, selected algorithms, parameter configurations, quality scores and processing effects during the processing, and send it back to the algorithm scheduling module 303 and store it in the algorithm knowledge base of the general parent platform to continuously optimize the algorithm selection accuracy of all positioning surgical robots on the platform.

[0091] Based on the same inventive concept, this disclosure also provides an adaptive image processing device for a universal mother platform of a positioning surgical robot, as shown in the following embodiment. Since the principle by which this device embodiment solves the problem is similar to that of the above-described method embodiment, the implementation of this device embodiment can refer to the implementation of the above-described method embodiment, and repeated details will not be elaborated further.

[0092] Figure 4 This is a schematic diagram of an adaptive image processing device for a universal mother platform of a positioning surgical robot, as provided in an embodiment of this application. Figure 4 As shown, the device includes: The task acquisition module 401 is used to acquire medical images and the task types of the medical image processing tasks to be performed. The algorithm determination module 402 is used to extract features from medical images and obtain the optimal execution algorithm based on the extracted medical image features, task type and pre-built algorithm knowledge base. The algorithm knowledge base is used to store at least one algorithm and corresponding algorithm metadata. The image processing module 403 is used to process medical images based on the optimal execution algorithm to obtain image processing results.

[0093] As described above, feature extraction is performed on the acquired medical images. Based on the medical image features, the pre-defined task type of the medical image processing task to be performed, and a pre-built algorithm knowledge base, the optimal execution algorithm is obtained. This algorithm is then used to process the medical images to obtain the image processing results. The pre-built algorithm knowledge base allows for direct selection of the optimal execution algorithm, improving the efficiency of the medical image processing process and the reliability of the obtained image processing results.

[0094] In this embodiment of the application, the algorithm determination module 402 is further configured to select a set of candidate algorithms from a pre-built algorithm knowledge base based on medical image features and task type, wherein the set of candidate algorithms includes at least one candidate algorithm that satisfies the medical image processing task to be executed. Candidate algorithms are evaluated according to preset scoring dimensions, and the scores are sorted in descending order. The candidate algorithm with the highest score is selected as the optimal execution algorithm.

[0095] In the embodiments of this application, the task type includes at least one of the following: medical image segmentation task, medical image registration task, medical image key point detection task, and medical image structure recognition task.

[0096] In this application embodiment, the scoring dimensions include at least one of the following: the degree of matching between the candidate algorithm and the medical image features, the algorithm accuracy of the candidate algorithm, the computation time of the candidate algorithm, and the resource consumption of the candidate algorithm during operation.

[0097] In this embodiment of the application, the image processing module 403 is further configured to process medical images based on the optimal execution algorithm to obtain initial image processing results; determine the corresponding evaluation strategy according to the task type, and evaluate the initial image processing results based on the evaluation strategy to obtain evaluation results; compare the evaluation results with a preset scoring threshold, and obtain the target image processing results based on the comparison results.

[0098] In this embodiment, the image processing module 403 is further configured to: use the initial image processing result as the target image processing result when the evaluation result is greater than or equal to a preset scoring threshold; determine the defect type of the initial image processing result based on the evaluation result when the evaluation result is less than the preset scoring threshold, and match the corresponding optimization strategy according to the defect type; optimize the image processing flow based on the optimization strategy, and reprocess the medical image based on the optimized image processing flow to obtain the optimized image processing result; evaluate the optimized image processing result until the preset termination condition is met to obtain the target image processing result.

[0099] In this embodiment of the application, the device further includes: an algorithm addition module 404, used to acquire a new algorithm and its corresponding algorithm metadata; deploy the new algorithm to the current algorithm execution environment, and perform interface standardization configuration on the new algorithm so that the new algorithm is in a callable state; and store the algorithm metadata of the new algorithm in the algorithm knowledge base according to the preset field rules of the algorithm knowledge base to obtain the updated algorithm knowledge base.

[0100] Correspondingly, the algorithm determination module 402 is also used to extract features from medical images and obtain the optimal execution algorithm based on the extracted medical image features, task type, and updated algorithm knowledge base.

[0101] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 5 As shown, the electronic device includes a processor 510 and a memory 520. In one possible form of the electronic device, it may also include an input device 530, an output device 540, and a communication device 550. The number of processors 510 in the electronic device can be one or more. Figure 5 Taking a processor 510 as an example; the processor 510, memory 520, input device 530, output device 540, and communication device 550 in the electronic device can be connected via a bus or other means. Figure 5 Taking the example of a connection between China and Israel via a bus.

[0102] The memory 520, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the adaptive image processing method for the universal mother platform of the positioning surgical robot in this embodiment. The processor 510 executes various functional applications and data processing of the electronic device by running the software programs, instructions, and modules stored in the memory 520, thereby realizing the aforementioned adaptive image processing method for the universal mother platform of the positioning surgical robot.

[0103] The memory 520 may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a given function; the data storage area may store data created based on the use of the electronic device. Furthermore, the memory 520 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 520 may further include memory remotely located relative to the processor 510, which can be connected to the electronic device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0104] Input device 530 can be used to receive network configuration information. Output device 540 may include electronic devices such as a display screen.

[0105] The aforementioned electronic device can be used to execute the adaptive image processing method of a general master platform for arbitrary positioning surgical robots, and has the corresponding functions and beneficial effects.

[0106] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the above-described apparatus and equipment can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0107] Furthermore, embodiments of this application also provide a storage medium containing computer-executable instructions. When executed by a computer processor, the computer-executable instructions are used to perform relevant operations in the adaptive image processing method for a general mother platform of a positioning surgical robot provided in any embodiment of this application, and have corresponding functions and beneficial effects.

[0108] Those skilled in the art will understand that embodiments of this application may be provided as methods, systems, or computer program products.

[0109] Therefore, this application may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, produce implementations of the flowchart... Figure 1 One or more processes and / or boxes Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0110] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory. Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, like read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0111] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0112] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0113] The above specific embodiments have further detailed the purpose, technical solution, and beneficial effects of this application. It should be understood that the above are merely specific embodiments of this application and are not intended to limit the scope of protection of this application. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application for those skilled in the art.

Claims

1. An adaptive image processing method for a universal master platform of a positioning surgical robot, characterized in that, include: The task types for acquiring medical images and the medical image processing tasks to be performed; Feature extraction is performed on the medical image, and the optimal execution algorithm is obtained based on the extracted medical image features, the task type, and a pre-built algorithm knowledge base. The algorithm knowledge base is used to store at least one algorithm and its corresponding algorithm metadata. The medical image is processed based on the optimal execution algorithm to obtain the image processing result.

2. The adaptive image processing method for a universal master platform of a positioning surgical robot according to claim 1, characterized in that, The step of obtaining the optimal execution algorithm based on the extracted medical image features, the task type, and a pre-built algorithm knowledge base includes: Based on medical image features and task type, a set of candidate algorithms is obtained by screening from a pre-built algorithm knowledge base, wherein the set of candidate algorithms includes at least one candidate algorithm that satisfies the medical image processing task to be executed. Candidate algorithms are evaluated according to preset scoring dimensions, and the scores are sorted in descending order. The candidate algorithm with the highest score is selected as the optimal execution algorithm.

3. The adaptive image processing method for a universal master platform of a positioning surgical robot according to claim 2, characterized in that, The task type includes at least one of the following: Medical image segmentation task, medical image registration task, medical image key point detection task, medical image structure recognition task.

4. The adaptive image processing method for a universal master platform of a positioning surgical robot according to claim 2, characterized in that, The scoring dimensions include at least one of the following: The matching degree between the candidate algorithm and the medical image features, the algorithm accuracy of the candidate algorithm, the computation time of the candidate algorithm, and the resource consumption of the candidate algorithm during operation.

5. The adaptive image processing method for a universal master platform of a positioning surgical robot according to claim 1, characterized in that, The process of processing the medical image based on the optimal execution algorithm to obtain the image processing result includes: The medical image is processed based on the optimal execution algorithm to obtain the initial image processing result; The corresponding evaluation strategy is determined according to the task type, and the initial image processing result is evaluated based on the evaluation strategy to obtain the evaluation result; The evaluation result is compared with a preset scoring threshold, and the target image processing result is obtained based on the comparison result.

6. The adaptive image processing method for a universal master platform of a positioning surgical robot according to claim 5, characterized in that, The step of comparing the evaluation result with a preset scoring threshold and obtaining the target image processing result based on the comparison result includes: If the evaluation result is greater than or equal to the preset scoring threshold, the initial image processing result will be used as the target image processing result. If the evaluation result is less than the preset scoring threshold, the defect type of the initial image processing result is determined based on the evaluation result, and the corresponding optimization strategy is matched according to the defect type. The image processing flow is optimized based on the optimization strategy, and the medical image is reprocessed based on the optimized image processing flow to obtain the optimized image processing result. The optimized image processing results are evaluated until the preset termination conditions are met, and the target image processing result is obtained.

7. The adaptive image processing method for a universal master platform of a positioning surgical robot according to claim 1, characterized in that, The method further includes: Retrieve the newly added algorithm and its corresponding algorithm metadata; Deploy the new algorithm into the current algorithm execution environment and standardize the interface configuration of the new algorithm to make the new algorithm callable. According to the preset field rules of the algorithm knowledge base, the algorithm metadata of the newly added algorithm is stored in the algorithm knowledge base to obtain the updated algorithm knowledge base; Accordingly, the step of extracting features from the medical image and obtaining the optimal execution algorithm based on the extracted medical image features, the task type, and a pre-built algorithm knowledge base includes: Feature extraction is performed on the medical image, and the optimal execution algorithm is obtained based on the extracted medical image features, the task type, and the updated algorithm knowledge base.

8. An adaptive image processing device for a universal master platform of a positioning surgical robot, characterized in that, include: The task acquisition module is used to acquire medical images and the task types of the medical image processing tasks to be performed. The algorithm determination module is used to extract features from the medical image and obtain the optimal execution algorithm based on the extracted medical image features, the task type, and a pre-built algorithm knowledge base. The algorithm knowledge base is used to store at least one algorithm and its corresponding algorithm metadata. The image processing module is used to process the medical image based on the optimal execution algorithm to obtain the image processing result.

9. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more computer programs; When the one or more computer programs are executed by the one or more processors, the electronic device implements the adaptive image processing method for a universal master platform of a positioning surgical robot as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the adaptive image processing method for a universal mother platform of a positioning surgical robot as described in any one of claims 1-7.

Images