Machine vision-based chip and method for evaluating flaps after finger replantation

Abstract

The application discloses a machine vision-based finger replantation postoperative flap evaluation intelligent chip and method, relates to the technical field of machine vision, and has the following steps: collecting static visual images and dynamic microcirculation visual images of a finger replantation postoperative flap under standard visual collection conditions to form a flap visual evaluation set, screening effective images in the flap visual evaluation set as flap to-be-evaluated images; obtaining epidermal color and luster features and subcutaneous blood vessel texture features of the flap in the flap to-be-evaluated images, judging whether the static and dynamic flap to-be-evaluated images need to be fused according to the change trend of the epidermal color and luster features and the distribution form of the subcutaneous blood vessel texture features, and the effect is that the static appearance and the dynamic microcirculation dual features of the flap are effectively considered, the problem that a single modal image is difficult to comprehensively reflect the real blood supply state of the flap is solved, and a solid data foundation is laid for subsequent accurate evaluation.

Description

Technical Field

[0001] This invention relates to the field of machine vision technology, and more specifically, to a smart chip and method for evaluating skin flaps after finger replantation based on machine vision. Background Technology

[0002] Finger replantation is a core technique in hand surgery for repairing limb amputations. Accurate postoperative assessment of the blood supply to the skin flap is crucial for ensuring the survival of the replanted finger and reducing the risk of tissue necrosis. With the increasing application of machine vision technology in medicine, visual image-based flap assessment has become a research hotspot. Furthermore, the development of dedicated intelligent chips and the optimization of assessment methods represent a key breakthrough for achieving real-time bedside assessment.

[0003] Currently, in clinical practice and existing technologies, machine vision-based flap assessment methods after finger replantation still face numerous technical bottlenecks and lack dedicated intelligent chips suitable for this scenario. At the methodological level, current visual assessments mostly analyze single static or dynamic microcirculation images, failing to achieve feature fusion decision-making between the two types of images and making it difficult to simultaneously consider the dual core features of epidermal color and subcutaneous vascular texture. Furthermore, the judgment of abnormal features is limited to single-parameter abnormality identification, lacking analysis of synergistic abnormalities involving multiple types of parameters superimposed in the same region. This makes it difficult to accurately distinguish the differentiated impact of single and synergistic abnormalities on blood supply, resulting in insufficient accuracy of assessment results. In addition, the existing assessment process lacks standardized quantitative logic in image fusion, abnormality determination, and risk assessment, relying heavily on empirical threshold settings, making it difficult to form a reusable and scalable assessment system. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the present invention aims to provide a smart chip and method for evaluating skin flaps after finger replantation based on machine vision.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A machine vision-based method for evaluating skin flaps after finger replantation surgery includes the following steps: Static visual images and dynamic microcirculation visual images of the flap after finger replantation under standard visual acquisition conditions were collected to form a flap visual evaluation set. Valid images in the flap visual evaluation set were selected as flap images to be evaluated. The epidermal color features and subcutaneous vascular texture features of the flap in the image to be evaluated are obtained. Based on the changing trend of the epidermal color features and the distribution pattern of the subcutaneous vascular texture features, it is determined whether the static and dynamic images of the flap to be evaluated need to be fused. The corresponding images of the flap to be evaluated are marked as flap fusion processing images. Extract the quantitative feature parameters required for flap evaluation from the flap fusion image, and determine whether there is a cooperative abnormality in the flap in the flap fusion image based on the distribution location of subcutaneous vascular texture features and the type of quantitative feature parameters. If there is no synergistic abnormality in the flap fusion image, the first evaluation result is obtained by judging the impact of a single abnormality on the flap blood supply state based on the numerical range and feature type of the quantified feature parameters. If there are synergistic abnormalities in the flap fusion image, the impact of the synergistic abnormalities on the flap blood supply status is determined based on the correlation degree and feature type of the quantified feature parameters to obtain the second evaluation result. The survival risk level of the flap is determined by combining the results of the first assessment and / or the second assessment.

[0006] Preferably, based on the changing trend of epidermal color features and the distribution pattern of subcutaneous vascular texture features, it is determined whether static and dynamic flap images to be evaluated need to be fused, and the corresponding flap images to be evaluated are marked as flap fusion processing images. This specifically includes the following steps: If the trend of changes in epidermal color features and the distribution pattern of subcutaneous vascular texture features form a complete correlation in a single flap image to be evaluated, then the flap image to be evaluated is directly marked as a flap fusion processing image. If the trend of epidermal color features is reflected in the static flap image to be evaluated, and the distribution pattern of subcutaneous vascular texture features is reflected in the dynamic flap image to be evaluated, then feature regions are extracted from the static and dynamic flap images to be evaluated respectively, forming a static feature region containing epidermal color features and a dynamic feature region containing subcutaneous vascular texture features. Using the microcirculation flow characteristics of the flap in the image to be evaluated as the association anchor point, the feature matching sites of the static feature region and the dynamic feature region are identified. After edge calibration of the feature matching sites, the static feature region and the dynamic feature region are fused to form the flap fusion image.

[0007] Preferably, the microcirculation flow characteristics of the skin flap in the image to be evaluated are used as correlation anchor points to identify feature matching sites between static and dynamic feature regions. After edge calibration of the feature matching sites, the static and dynamic feature regions are fused to form a skin flap fusion image. The specific steps include: Based on the core distribution area of ​​microcirculation flow characteristics, determine whether the number of overlapping matching sites between static and dynamic feature areas reaches a preset threshold. If the number of overlapping matching sites reaches a preset threshold, the overlapping matching site is marked as the core matching area. The static feature area and dynamic feature area are then calibrated and fused to form a flap fusion image based on the core matching area. If the number of overlapping matching sites does not reach the preset threshold, the edge feature points of the static feature region and the dynamic feature region are extracted and paired. If the paired edge feature points can form a continuous flap feature contour, then the region is fused based on the feature contour to form a flap fusion image.

[0008] Preferably, based on the distribution location of subcutaneous vascular texture features and the type of quantified feature parameters, the determination of whether there is a cooperative abnormality in the flap fusion image includes the following steps: If the flap in the fused image has only one type of abnormal quantified feature parameter, then the flap is determined to have no cooperative abnormality representation. If there are at least two types of abnormal quantization feature parameters in the flap fusion image, then based on the distribution location of the subcutaneous vascular texture features, it is determined whether the different types of abnormal quantization feature parameters form an overlapping distribution in the same area of ​​the flap. If different types of abnormal quantitative feature parameters do not form an overlapping distribution in the same area of ​​the flap, it is determined that the flap does not have a synergistic abnormal characterization; if different types of abnormal quantitative feature parameters form an overlapping distribution in the same area of ​​the flap, it is determined that the flap has a synergistic abnormal characterization.

[0009] Preferably, the first assessment result is obtained by determining the impact of a single abnormal feature on the blood supply status of the flap based on the numerical range and feature type of the quantified feature parameters, specifically including the following steps: By comparing the quantitative feature parameter values ​​corresponding to each single abnormality with the flap single feature and blood supply assessment comparison table, the blood supply impact level and impact weight value corresponding to each single abnormality are obtained. The impact on blood supply is categorized into mild, moderate, and severe. Based on the blood supply impact level and impact weight value of each individual abnormality, the impact of a single abnormality on the blood supply status of the flap is calculated to obtain the first assessment result.

[0010] Preferably, based on the blood supply impact level and impact weight value of each individual abnormality, the impact of a single abnormality on the blood supply status of the flap is calculated to obtain the first assessment result, specifically including the following steps: Determine whether there are multiple single abnormal features of the same blood supply level; If there are no multiple single abnormal manifestations with the same blood supply impact level, the blood supply impact level of each single abnormal manifestation and its corresponding impact weight value are directly combined to form the first assessment result. If there are multiple single abnormal manifestations of the same blood supply impact level, the impact weight values ​​of all single abnormal manifestations under that level are summed to obtain the comprehensive weight value of that blood supply impact level. Each blood supply impact level is combined with the corresponding comprehensive weight value to form the first assessment result.

[0011] Preferably, based on the correlation and feature type of the quantified feature parameters, the impact of synergistic abnormality on the flap blood supply status is determined to obtain a second evaluation result, specifically including the following steps: Calculate the correlation degree of quantitative feature parameters of different types of abnormalities within the same overlapping region of the flap, and determine whether the synergistic abnormality is characterized as strongly correlated or weakly correlated. If the synergistic abnormality is strongly correlated, the degree of blockage of the flap microcirculation pathway by the strongly correlated abnormality is used to obtain the first synergistic assessment result of flap blood supply. If the synergistic abnormality is weakly correlated, the degree of influence of the weakly correlated abnormality on the local tissue nutrition supply of the flap is determined to obtain the second synergistic assessment result of flap blood supply. The first and second collaborative evaluation results are integrated to form the second evaluation result.

[0012] Preferably, the first synergistic assessment result of flap blood supply is obtained by determining the degree of blockage of the flap microcirculation pathway by strongly correlated abnormal features, specifically including the following steps: Determine whether strongly correlated abnormal features cause occlusion of the flap microcirculation pathway; If occlusion occurs, obtain the area proportion and location information of the occluded area. Compare the area proportion and location information of the occluded area with the flap microcirculation occlusion and blood supply assessment table to obtain the risk level and intervention recommendations for flap blood supply. Mark the risk level and intervention recommendations for flap blood supply as the first collaborative assessment result. If no occlusion is caused, the influence of strongly correlated abnormal features on the degree of stenosis and blood flow velocity of the flap microcirculation pathway is determined, and two key parameters are obtained: the stenosis ratio of the microcirculation pathway and the blood flow velocity deviation. The stenosis percentage and blood flow velocity deviation values ​​are compared with preset thresholds; If the stenosis percentage and blood flow velocity deviation are both below the preset threshold, it is determined that there is no significant adverse effect on flap blood supply, and the corresponding first collaborative assessment result with no intervention requirement is marked. If either the stenosis percentage or the blood flow velocity deviation is higher than or equal to a preset threshold, it is determined that there is a potential adverse impact on the blood supply of the flap. Based on the abnormal location and characteristic change trend, preventive intervention suggestions and corresponding blood supply risk levels are given, and it is marked as the first collaborative assessment result.

[0013] Preferably, the second synergistic assessment result of flap blood supply is obtained by determining the degree of influence of weakly correlated abnormal features on the local tissue nutrition supply of the flap, specifically including the following steps: The numerical values ​​of each quantitative characteristic parameter corresponding to the weakly correlated abnormal characterization are compared with the flap single feature and blood supply assessment comparison table to obtain the blood supply impact level corresponding to each abnormal characterization. The degree of damage to local tissue nutrition supply is determined based on the superposition result of each blood supply impact level, and this degree of damage is marked as the second collaborative assessment result.

[0014] A machine vision-based intelligent chip for flap evaluation after finger replantation surgery includes: an image visual acquisition unit, an image feature acquisition unit, a feature parameter extraction unit, a feature judgment unit 1, a feature judgment unit 2, and a grade judgment unit. The image visual acquisition unit is used to acquire static visual images and dynamic microcirculation visual images of the flap after finger replantation under standard visual acquisition conditions to form a flap visual evaluation set, and to select effective images in the flap visual evaluation set as flap images to be evaluated. The image feature acquisition unit is used to acquire the epidermal color features and subcutaneous vascular texture features of the flap in the flap to be evaluated image. Based on the changing trend of the epidermal color features and the distribution pattern of the subcutaneous vascular texture features, it determines whether it is necessary to perform fusion processing on the static and dynamic flap to be evaluated images and marks the corresponding flap to be evaluated image as the flap fusion processing image. The feature parameter extraction unit is used to extract the quantitative feature parameters required for flap evaluation in the flap fusion image. Based on the distribution location of subcutaneous vascular texture features and the type of quantitative feature parameters, it determines whether there is a cooperative abnormality in the flap in the flap fusion image. The feature judgment unit is used to determine that there is no synergistic abnormal characterization of the flap in the flap fusion processing image. Based on the numerical range and feature type of the quantified feature parameters, the influence of a single abnormal characterization on the blood supply status of the flap is judged to obtain the first evaluation result. The second feature judgment unit is used to judge whether there is a cooperative abnormality in the flap in the flap fusion processing image. Based on the correlation degree and feature type of the quantified feature parameters, the influence of the cooperative abnormality on the blood supply status of the flap is judged to obtain the second evaluation result. The rating judgment unit is used to combine the first assessment result and / or the second assessment result to comprehensively determine the survival risk level of the flap.

[0015] Compared with existing technologies, this invention has the following beneficial effects: By extracting epidermal color features and subcutaneous vascular texture features, and making fusion decisions based on the trend and distribution pattern of feature changes, it effectively takes into account both the static appearance and dynamic microcirculation characteristics of the flap, solving the problem that a single modality image cannot fully reflect the true blood supply status of the flap, and laying a solid data foundation for subsequent accurate assessment; by judging whether different types of quantitative feature parameters form an overlapping distribution in the same area of ​​the flap, it accurately identifies synergistic abnormalities, clearly distinguishes the differential impact of single abnormalities and synergistic abnormalities on blood supply, effectively improves the ability to identify complex blood supply disorders, and avoids errors caused by missed synergistic abnormalities. The assessment bias is reduced; by quantifying the numerical range of characteristic parameters, calculating correlation, and superimposing influence weights, empirical judgment is transformed into a quantifiable and reproducible assessment process, forming a reusable and scalable standardized assessment system, which significantly improves the objectivity and consistency of assessment results; by distinguishing between single abnormalities and synergistic abnormalities, first and second assessment results are generated respectively, and the survival risk level of flaps is comprehensively determined, which can provide intervention suggestions for clinicians. For abnormalities with no obvious impact, intervention-free monitoring can be achieved; for potential risks, preventive intervention can be provided; and for high-risk situations, timely guidance for emergency treatment can be provided, effectively improving the success rate of treatment and the quality of patient prognosis after finger replantation. Attached Figure Description

[0016] Figure 1 A schematic diagram illustrating the steps of a machine vision-based flap evaluation method for replantation of severed fingers in this invention embodiment; Figure 2 A schematic diagram of the steps in forming the first evaluation result in a machine vision-based flap evaluation method for replantation of severed fingers is provided for an embodiment of the present invention. Figure 3 This invention provides a schematic diagram of a smart chip for evaluating skin flaps after finger replantation based on machine vision. Detailed Implementation

[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0018] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0019] Secondly, the term "an embodiment" or "embodiment" as used 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 different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0020] Reference Figures 1-3 As shown.

[0021] This embodiment further illustrates the intelligent chip and method for evaluating skin flaps after finger replantation based on machine vision proposed in this invention.

[0022] A machine vision-based method for evaluating skin flaps after finger replantation surgery includes the following steps: Static visual images and dynamic microcirculation visual images of the flap after finger replantation under standard visual acquisition conditions were collected to form a flap visual evaluation set. Valid images in the flap visual evaluation set were selected as flap images to be evaluated. The epidermal color features and subcutaneous vascular texture features of the flap in the image to be evaluated are obtained. Based on the changing trend of the epidermal color features and the distribution pattern of the subcutaneous vascular texture features, it is determined whether the static and dynamic images of the flap to be evaluated need to be fused. The corresponding images of the flap to be evaluated are marked as flap fusion processing images.

[0023] First, images of the skin flap after finger replantation are acquired. Simultaneously, static visual images clearly showing the flap's surface morphology and dynamic microcirculation visual images capturing the blood flow within the flap are acquired. These static and dynamic microcirculation visual images together constitute the flap visual evaluation set. After image acquisition, the system performs a validity screening of all images in the evaluation set, removing invalid images due to factors such as shooting angle deviation, abnormal light reflection, or image blurring. Only images with clear image quality, intact flap areas, and a true reflection of the flap's condition are retained as images to be evaluated in subsequent analysis.

[0024] After image acquisition, depth features are extracted from the selected flap images to be evaluated, focusing on the epidermal color and subcutaneous vascular texture features. Epidermal color features directly reflect the overall basic blood supply status of the flap, while subcutaneous vascular texture features accurately reflect the distribution and operation of local microcirculation within the flap. After feature extraction, the system comprehensively analyzes the epidermal color and subcutaneous vascular texture features, combining the changing trends of the epidermal color features and the distribution pattern of the subcutaneous vascular texture features to determine whether static visual images and dynamic microcirculation visual images need to be fused. For example, when the epidermal color shows uneven changes and the subcutaneous vascular texture shows a locally sparse pattern, a single type of image cannot fully represent the true state of the flap. In this case, the system will determine that image fusion is required and mark the flap images requiring fusion as flap fusion processing images. If both the epidermal color and subcutaneous vascular texture features are stable and clear, a single image can meet the analysis requirements, and fusion processing is not necessary.

[0025] After image fusion and labeling, the system extracts various quantitative feature parameters for flap evaluation from the labeled flap fusion images. These quantitative feature parameters are a digital representation of the flap's state, enabling precise description of its characteristics. Subsequently, the system combines the distribution of subcutaneous vascular texture features within the flap region with the specific types of the extracted quantitative feature parameters to determine synergistic anomalies. For example, if the subcutaneous vascular texture is interrupted at the distal end of the flap, and the corresponding blood flow velocity and vascular density quantitative parameters show abnormal fluctuations, the system determines that the flap in the fusion image exhibits synergistic anomalies. If only a single quantitative parameter shows a slight deviation, and the distribution of subcutaneous vascular texture is normal, then no synergistic anomaly is determined. This allows for accurate identification of multi-dimensional blood supply abnormalities in the flap, providing a core basis for subsequent flap survival risk assessment.

[0026] Extract the quantitative feature parameters required for flap evaluation from the flap fusion image. Based on the distribution location of subcutaneous vascular texture features and the type of quantitative feature parameters, determine whether there is a cooperative abnormality in the flap in the flap fusion image.

[0027] The quantitative characteristic parameters include color deviation value, vascular texture branch density, and microcirculation flow velocity value.

[0028] If there is no synergistic abnormality in the flap fusion image, the first evaluation result is obtained by judging the impact of a single abnormality on the flap blood supply state based on the numerical range and feature type of the quantified feature parameters. If there are synergistic abnormalities in the flap fusion image, the impact of the synergistic abnormalities on the flap blood supply status is determined based on the correlation degree and feature type of the quantified feature parameters to obtain the second evaluation result. The survival risk level of the flap is determined by combining the results of the first assessment and / or the second assessment.

[0029] When no synergistic anomalous features are present in the flap fusion image, the system will determine the independent impact of each anomalous feature on the flap's blood supply based on the numerical range and feature type of the extracted quantitative feature parameters. For example, if only the epidermal color parameter is slightly darker, while other quantitative parameters are within the normal range, the system will determine that the impact of this single anomalous feature on the flap's blood supply is low, thus generating a first assessment result. This result focuses on the impact of a single factor, providing a basis for risk assessment.

[0030] When synergistic anomalous features are present in the flap fusion image, the system switches to the synergistic anomalous feature evaluation logic. In this case, the system focuses on analyzing the correlation between various quantitative feature parameters and the combined effects of different feature types. For example, when the distribution of subcutaneous vascular texture is significantly disrupted simultaneously with blood flow velocity and vascular density parameters, the system determines that there is a strong correlation between these anomalous features, collectively exacerbating the risk of flap blood supply disorders. This results in a second evaluation result, which focuses more on the degree of risk under the synergistic effect of multiple factors and can more accurately reflect the complex blood supply problems faced by the flap.

[0031] After generating either the first or second assessment result, the system will make a comprehensive judgment based on the actual situation. If only the first assessment result is generated, the system will directly determine the survival risk level of the flap based on the degree of influence of a single abnormal feature in that result. If a second assessment result is generated, or both the first and second assessment results are generated simultaneously, the system will weight and integrate these results, prioritizing the influence weight of synergistic abnormal features, and then combining the supplementary role of the single abnormal feature to ultimately determine the survival risk level of the flap. For example, when the second assessment result shows a highly correlated synergistic abnormality, while the first assessment result shows that the single abnormality has a minor impact, the system will use the second assessment result as the core to comprehensively determine that the flap is at a high survival risk level, providing clear decision-making guidance for clinical intervention.

[0032] Based on the changing trends of epidermal color features and the distribution patterns of subcutaneous vascular texture features, it is determined whether static and dynamic flap images to be evaluated need to be fused. The corresponding flap images to be evaluated are then marked as flap fusion processing images. The specific steps include: If the trend of changes in epidermal color features and the distribution pattern of subcutaneous vascular texture features form a complete correlation in a single flap image to be evaluated, then the flap image to be evaluated is directly marked as a flap fusion processing image. If the trend of epidermal color features is reflected in the static flap image to be evaluated, and the distribution pattern of subcutaneous vascular texture features is reflected in the dynamic flap image to be evaluated, then feature regions are extracted from the static and dynamic flap images to be evaluated respectively, forming a static feature region containing epidermal color features and a dynamic feature region containing subcutaneous vascular texture features. Using the microcirculation flow characteristics of the flap in the image to be evaluated as the association anchor point, the feature matching sites of the static feature region and the dynamic feature region are identified. After edge calibration of the feature matching sites, the static feature region and the dynamic feature region are fused to form the flap fusion image.

[0033] After extracting the epidermal color features and subcutaneous vascular texture features from the image to be evaluated for the flap, the system first determines the distribution carriers of these two types of features to classify different processing paths. The first processing path is suitable when the epidermal color features and subcutaneous vascular texture features form a complete association in a single image to be evaluated for the flap. The changing trend of epidermal color features refers to the gradual change pattern, uniformity, and overall direction of color within the flap area. For example, the color depth changes and continuous transition of redness from the center to the edge of the flap. These changes can intuitively reflect the local blood supply gradient of the flap. The distribution pattern of subcutaneous vascular texture features includes the direction of blood vessels, the density of branches, network integrity, and relative position to the flap boundary, which is an intuitive manifestation of the microcirculation status. When the epidermal color features and subcutaneous vascular texture features are clearly present in the same image to be evaluated for the flap, and form a complete association in spatial location and logic, the system will directly mark the image as a flap fusion processing image. For example, the static flap image to be evaluated, thanks to its high-resolution imaging capabilities, clearly shows the color change trend of the flap epidermis from a rosy center to a light red edge, and also fully displays the distribution pattern of subcutaneous vascular texture, which starts from the main blood vessels, branches out evenly in a tree-like manner, and extends to the edge of the flap. The area of ​​epidermal color change and the distribution range of vascular texture are perfectly matched. The epidermal color features and the subcutaneous vascular texture features form a close and complete correlation. At this time, no additional image operations are required, and the labeling of the fused image can be completed directly.

[0034] The second processing approach addresses situations where epidermal color features and subcutaneous vascular texture features exist in different image types. Specifically, the trend of epidermal color features is only reflected in static images of the flap being evaluated, while the distribution of subcutaneous vascular texture features is only shown in dynamic images. The imaging characteristics of static images allow for precise capture of detailed changes in epidermal color. The trend of epidermal color features can be reflected through the distribution of color pixel values ​​and the gradient of color differences between regions. For example, in static images, the color pixel values ​​are higher at the proximal end of the flap and lower at the distal end, forming a clear linear trend. This feature is an important basis for judging the overall blood supply fullness of the flap. Dynamic images, through the advantage of continuous frame imaging, can filter out interference from epidermal tissue and clearly present the distribution of subcutaneous vascular texture. For example, in dynamic images, the main trunk direction, branch node positions, and the sparse or dense state of vascular texture in local areas can be clearly seen. This feature is a core basis for analyzing microcirculation patency. To address this feature separation issue, the system first performs feature region extraction. It accurately locates and extracts the region containing the complete epidermal color change trend from the static flap image to be evaluated, as the static feature region, ensuring that this region covers all key parts of the flap's color change. At the same time, it extracts the region containing the distribution pattern of all subcutaneous vascular textures from the dynamic flap image to be evaluated, as the dynamic feature region, ensuring that the main trunk, branches, and subtle edge textures of the vascular textures are completely included.

[0035] After feature region extraction, the system uses the microcirculation flow characteristics of the skin flap as an anchor point. These microcirculation flow characteristics are the core link between static epidermal color and dynamic vascular texture, as the microcirculation flow state of the flap directly determines the appearance of epidermal color and forms a causal relationship with the distribution pattern of subcutaneous vascular texture. Based on this anchor point, the system identifies feature matching sites between static and dynamic feature regions. These sites are precise spatial correspondences between the two types of features. For example, abrupt changes in epidermal color within the static feature region correspond to branch nodes of subcutaneous vascular texture within the dynamic feature region; the core area with the most vibrant red color in the static feature region corresponds to the core distribution of the main subcutaneous blood vessels within the dynamic feature region. After identifying the feature matching sites, the system performs edge calibration on these sites. By adjusting the spatial position of the dynamic or static feature regions, all matching sites achieve precise overlap, eliminating positional deviations caused by imaging angles or slight flap displacement, ensuring complete spatial alignment of the two types of features. After edge calibration, the system will fuse the calibrated static feature regions with the dynamic feature regions, so that the fused image not only fully preserves the trend of epidermal color change in the static image, but also clearly presents the distribution pattern of subcutaneous blood vessel texture in the dynamic image, and finally forms a flap fusion image, providing complete image data containing dual core features for subsequent judgment of flap abnormality.

[0036] Using the microcirculation flow characteristics of the skin flap in the image to be evaluated as correlation anchor points, feature matching sites between static and dynamic feature regions are identified. After edge calibration of the feature matching sites, the static and dynamic feature regions are fused to form a skin flap fusion image. The specific steps include: Based on the core distribution area of ​​microcirculation flow characteristics, determine whether the number of overlapping matching sites between static and dynamic feature areas reaches a preset threshold. If the number of overlapping matching sites reaches a preset threshold, the overlapping matching site is marked as the core matching area. The static feature area and dynamic feature area are then calibrated and fused to form a flap fusion image based on the core matching area. If the number of overlapping matching sites does not reach the preset threshold, the edge feature points of the static feature region and the dynamic feature region are extracted and paired. If the paired edge feature points can form a continuous flap feature contour, then the region is fused based on the feature contour to form a flap fusion image.

[0037] Microcirculation flow characteristics are key to connecting epidermal color features and subcutaneous vascular texture features. The core distribution area is where blood perfusion is most active and vascular texture is densest within the skin flap, and it is also the area with the strongest correlation between epidermal color features and subcutaneous vascular texture features. Therefore, the system uses this core distribution area as the initial benchmark for the entire matching and fusion process. After establishing the benchmark, the system first counts the number of overlapping matching sites between static and dynamic feature regions and compares this number with a preset threshold. This step is to determine the degree of matching between the two types of feature regions in the core correlation area, thereby selecting the corresponding fusion strategy.

[0038] When the number of overlapping matching sites reaches a preset threshold, it indicates that a stable and sufficient correspondence has been formed between the static and dynamic feature regions in the core distribution area of ​​microcirculation flow features. At this point, the system marks these overlapping matching sites as the core matching area. The core matching area concentrates the points with the strongest correlation and most critical information between the two types of feature regions. For example, the cluster of pixels with the most rosy skin color in the static feature region overlaps with the cluster of nodes with the densest subcutaneous vascular texture in the dynamic feature region, thus jointly constituting the core matching area. The system uses this core matching area as a reference to perform position calibration on the static and dynamic feature regions. By fine-tuning the spatial coordinates, angles, and scaling ratios of the regions, all sites within the core matching area are perfectly aligned, ensuring that the spatial positions of the two types of features are completely aligned in the core region. After position calibration, the system fuses the two regions, so that the fused image clearly retains the trend of skin color change in the core region and accurately presents the distribution pattern of subcutaneous vascular texture, ultimately forming a flap fusion image. For example, the central region of the skin flap after finger replantation is the core distribution area of ​​microcirculation flow characteristics. In the static feature area, the epidermis of this area has a uniform red color, while in the dynamic feature area, the subcutaneous vascular texture of this area is densely distributed in a network. The number of overlapping matching sites formed by the two far exceeds the preset threshold. After the system marks this area as the core matching area, the generated image achieves accurate superposition of the two types of features in the central core area after position calibration and fusion.

[0039] If the number of overlapping matching sites does not reach the preset threshold, it indicates that the matching degree between the two types of feature regions in the core distribution area of ​​microcirculation flow features is insufficient, and a stable core matching area cannot be formed. At this time, the system will switch the fusion path and turn to edge feature point matching and fusion based on the overall contour of the flap. The system will first extract the edge feature points of the static feature region and the dynamic feature region. The edge feature points of the static feature region originate from the boundary contour of the skin color of the flap, such as the boundary point where the skin color transitions from rosy to the pale of the surrounding normal skin. These boundary points outline the actual physical contour of the flap. The edge feature points of the dynamic feature region originate from the distribution boundary of the subcutaneous vascular texture, such as the farthest node of the extension of the subcutaneous vascular texture. These points outline the coverage contour of the flap microcirculation. After extraction, the system will pair the two sets of edge feature points one by one. The core logic of pairing is to find the corresponding point with the best match in position and attributes between the two types of edge feature points based on the spatial coordinates, gradient direction and distance from the geometric center of the flap. After pairing, the system checks whether the paired edge feature points can be connected to form a continuous flap feature contour. This continuous contour needs to completely cover the entire flap area, conforming to both the physical boundaries of the epidermis and the coverage of the microcirculation. For example, after finger replantation, the core area of ​​the flap may have indistinct microcirculation flow characteristics due to postoperative swelling, and the number of overlapping matching points in the core distribution area may not reach the threshold. In this case, the system extracts the edge junctions of the flap epidermis within the static feature area and the farthest nodes of the subcutaneous vascular texture within the dynamic feature area. After pairing, these points are connected to form a continuous contour surrounding the entire flap, perfectly fitting all boundaries of the fingertip and lateral edges of the flap. After confirming the continuity of the contour, the system uses this flap feature contour as a unified benchmark to align and fuse the static and dynamic feature areas as a whole. The static epidermal color features and the dynamic subcutaneous vascular texture features are completely superimposed according to the contour benchmark, ultimately forming a flap fusion image. This ensures that even if the core area matching is insufficient, the two types of features can be effectively integrated through the overall contour, providing complete image data for subsequent judgment of flap abnormalities.

[0040] Based on the distribution location of subcutaneous vascular texture features and the type of quantified feature parameters, the determination of whether there are cooperative abnormalities in the flap fusion image includes the following steps: If the flap in the fused image has only one type of abnormal quantified feature parameter, then the flap is determined to have no cooperative abnormality representation. If there are at least two types of abnormal quantization feature parameters in the flap fusion image, then based on the distribution location of the subcutaneous vascular texture features, it is determined whether the different types of abnormal quantization feature parameters form an overlapping distribution in the same area of ​​the flap. If different types of abnormal quantitative feature parameters do not form an overlapping distribution in the same area of ​​the flap, it is determined that the flap does not have a synergistic abnormal characterization; if different types of abnormal quantitative feature parameters form an overlapping distribution in the same area of ​​the flap, it is determined that the flap has a synergistic abnormal characterization.

[0041] The system first performs preliminary statistical analysis on the types and abnormal states of the quantitative feature parameters extracted from the flap fusion image. The types of quantitative feature parameters cover multiple dimensions of flap blood supply assessment, such as the mean hue and saturation standard deviation related to epidermal color, vascular density and blood flow velocity related to subcutaneous vessels, and the oxygenation index related to tissue perfusion. Each type of parameter reflects the physiological state of the flap from different perspectives. When the system detects an abnormality in only one type of quantitative feature parameter in the flap fusion image—for example, only the mean hue of the epidermal color is below the normal range, while all other quantitative feature parameters are within the normal range—the system determines that the flap does not exhibit synergistic abnormalities. This is because a single type of parameter abnormality usually only reflects a local or single-dimensional problem and does not constitute a multi-factor synergistic blood supply disorder.

[0042] When the system detects anomalies in at least two types of quantified feature parameters in the flap fusion image, it proceeds to the next step of spatial distribution determination. Combining this with the distribution location of subcutaneous vascular texture features, it determines whether these abnormal quantified feature parameters overlap within the same region of the flap. The distribution location of subcutaneous vascular texture features is crucial for determining regional correlation, clearly indicating the vascular network coverage in different regions within the flap. For example, the vascular texture is dense in the proximal region, sparse in the distal region, and branching at the edges. The system spatially maps the spatial distribution range of each abnormal quantified feature parameter to the distribution location of the subcutaneous vascular texture features. For instance, it compares regions with abnormally reduced blood flow velocity with regions with sparse vascular texture, and regions with abnormally low oxygenation index with regions with interrupted vascular texture.

[0043] If, after spatial mapping, it is found that different types of abnormal quantitative characteristic parameters do not overlap in the same area of ​​the flap—for example, an area with abnormally reduced blood flow velocity is located proximally on the flap, while an area with abnormally low oxygenation index is located distally, and the corresponding subcutaneous vascular texture distributions in these two areas are independent—the system will determine that the flap does not exhibit synergistic abnormal characteristics. This is because the abnormal parameters in different areas lack spatial correlation and do not constitute a synergistic vascular obstruction. Conversely, if, after spatial mapping, it is found that different types of abnormal quantitative characteristic parameters overlap in the same area of ​​the flap—for example, abnormally reduced blood flow velocity, abnormally reduced vascular density, and abnormally low oxygenation index are all concentrated in the same area distal to the flap, and this area happens to correspond to the distribution location of the interrupted subcutaneous vascular texture—the system will determine that the flap exhibits synergistic abnormal characteristics. This is because multidimensional vascular abnormalities overlap in the same area and highly coincide with the pathological distribution of vascular textures, indicating that the flap faces a severe risk of vascular obstruction due to the synergistic effect of multiple factors in this area.

[0044] For example, in the processed images of skin flap fusion after finger replantation, the system detected abnormalities in three quantitative characteristic parameters: vascular density, blood flow velocity, and oxygenation index. After spatial mapping, it was found that the abnormal areas of these three parameters were concentrated at the distal end of the fingertip of the skin flap, and the subcutaneous vascular texture features in this area showed obvious interrupted distribution. At this time, the system would determine that the skin flap has a synergistic abnormality, suggesting that the clinic should focus on intervention in this area to avoid further deterioration of blood supply disorders.

[0045] Based on the numerical range and feature type of the quantified feature parameters, the impact of a single abnormal feature on the blood supply status of the flap is determined to obtain the first assessment result, which specifically includes the following steps: By comparing the quantitative feature parameter values ​​corresponding to each single abnormality with the flap single feature and blood supply assessment comparison table, the blood supply impact level and impact weight value corresponding to each single abnormality are obtained. The impact on blood supply is categorized into mild, moderate, and severe. Based on the blood supply impact level and impact weight value of each individual abnormality, the impact of a single abnormality on the blood supply status of the flap is calculated to obtain the first assessment result.

[0046] First, the system focuses on each individual abnormality identified in the flap fusion image and extracts its corresponding quantitative feature parameter values. These quantitative feature parameters correspond one-to-one with the abnormality, and different types of feature parameters accurately reflect different dimensions of the flap's physiological state. For example, abnormalities in epidermal color correspond to the mean hue parameter, abnormalities in subcutaneous vascular texture correspond to the vascular density parameter, and abnormalities in microcirculation flow correspond to the blood flow velocity parameter. Then, the system compares these specific parameter values ​​one by one with a pre-constructed flap single-feature and blood supply assessment comparison table based on extensive clinical monitoring data of flaps after finger replantation. This comparison table defines clear numerical ranges for each type of quantitative feature parameter, with each range corresponding to a unique blood supply impact level and a fixed impact weight value, ensuring the uniformity and objectivity of the assessment standards.

[0047] After completing the control group, the system clearly categorized the impact on blood supply into three levels: mild, moderate, and severe. Each level corresponds to a different degree of damage to the flap's blood supply, and is assigned a differentiated impact weight value. Mild impact corresponds to a weight value of 0.3, applicable to situations where the quantitative parameter values ​​slightly deviate from the normal range, indicating that the abnormality only causes a temporary and minor interference with the flap's blood supply. Moderate impact corresponds to a weight value of 0.6, applicable to situations where the quantitative parameter values ​​significantly deviate from the normal range, indicating that the abnormality has caused a persistent interference with the local blood supply of the flap, requiring close monitoring. Severe impact corresponds to a weight value of 0.9, applicable to situations where the quantitative parameter values ​​severely deviate from the normal range, indicating that the abnormality has directly threatened the patency of the local blood supply to the flap, posing a high risk of tissue necrosis. For example, regarding the quantitative characteristic parameter of vascular density, the control table sets the normal value range to 120 to 150 vessels per square millimeter. If the vascular density value corresponding to a single abnormal feature is 112 vessels per square millimeter, it will be judged as a mild impact after comparison, with an impact weight value of 0.3; if the value is 85 vessels per square millimeter, it will be judged as a moderate impact, with an impact weight value of 0.6; and if the value is 40 vessels per square millimeter, it will be judged as a severe impact, with an impact weight value of 0.9.

[0048] After matching the blood supply impact level and weight values ​​for all individual abnormalities, a weighted calculation is performed to obtain the comprehensive impact of each individual abnormality on the flap's blood supply status, which is the first assessment result. The weighted summation formula is as follows: Where S represents the overall impact value of the first assessment result, and W i G represents the influence weight value corresponding to the i-th single anomaly representation. iThis represents the quantified value of the blood supply impact level corresponding to the i-th single abnormal feature, where a mild impact is quantified as 1, a moderate impact as 2, and a severe impact as 3. For example, in a flap fusion image, there are two single abnormal features. The first abnormal feature is a low mean epidermal color tone, with a value of 13, which is considered a mild impact after comparison (W1=0.3, G1=1). The second abnormal feature is a slow blood flow velocity, with a value of 60% of the normal range, which is considered a moderate impact after comparison (W2=0.6, G2=2). Substituting into the formula, we get S=0.3×1+0.6×2=1.5. This comprehensive impact value of 1.5 is the first assessment result, clearly quantifying the overall impact of these two single abnormal features on the flap's blood supply status. If the flap has only one single abnormal feature, such as a vessel density of 40 vessels per square millimeter, it is judged as a severe impact (W1=0.9, G1=3), then S=0.9×3=2.7, which is the corresponding first assessment result. Through this quantitative calculation method, the system can transform qualitative abnormalities into quantitative assessment results, providing accurate and objective data support for the comprehensive determination of the flap survival risk level.

[0049] Based on the blood supply impact level and impact weight value of each individual abnormality, the impact of each individual abnormality on the blood supply status of the flap is calculated to obtain the first assessment result, which specifically includes the following steps: Determine whether there are multiple single abnormal features of the same blood supply level; If there are no multiple single abnormal manifestations with the same blood supply impact level, the blood supply impact level of each single abnormal manifestation and its corresponding impact weight value are directly combined to form the first assessment result. If there are multiple single abnormal manifestations of the same blood supply impact level, the impact weight values ​​of all single abnormal manifestations under that level are summed to obtain the comprehensive weight value of that blood supply impact level. Each blood supply impact level is combined with the corresponding comprehensive weight value to form the first assessment result.

[0050] First, the system categorizes and statistically analyzes the blood supply impact levels of all single abnormal features to determine if multiple single abnormal features of the same blood supply impact level exist. Since multiple abnormal features of the same level can have a cumulative effect on flap blood supply, a comprehensive weighting value is needed to reflect this cumulative effect. For example, if a flap simultaneously exhibits two single abnormal features—darkened skin color and low local temperature—and both are classified as mild impacts, the system will determine that multiple single abnormal features of the same blood supply impact level exist. If the flap's single abnormal features belong to mild, moderate, and severe impacts, and only one abnormal feature exists at each level, then it is determined that multiple abnormal features of the same level do not exist.

[0051] When multiple single abnormalities of the same blood supply impact level do not exist, the system will directly combine them to form the first assessment result. In this case, the blood supply impact level of each single abnormality corresponds one-to-one with its corresponding impact weight value. The system directly associates the level with the weight value, forming a clear assessment list that clarifies the independent impact of each single abnormality on the flap's blood supply. For example, if the flap has three single abnormalities: mild epidermal color abnormality (weight 0.3), moderate vascular density abnormality (weight 0.6), and severe blood flow velocity abnormality (weight 0.9), and only one abnormality exists under each level, the system will directly combine mild impact with 0.3, moderate impact with 0.6, and severe impact with 0.9 to form the first assessment result, clearly presenting the independent impact degree of each single abnormality.

[0052] When multiple single abnormal features of the same blood supply impact level exist, the system uses a comprehensive weighting calculation to form the first assessment result. In this case, the system first sums the impact weights of all single abnormal features under that level to obtain the comprehensive weight value for that blood supply impact level. This comprehensive weight value reflects the cumulative effect of multiple abnormal features of the same level. For example, if a skin flap has three single abnormal features, all belonging to mild impact, with corresponding impact weights of 0.3, 0.25, and 0.35 respectively, the system will sum these three weights to obtain the comprehensive weight value for mild impact: 0.3 + 0.25 + 0.35 = 0.9. Then, the mild impact is combined with the comprehensive weight value of 0.9. If other levels of single abnormal features also exist, such as a moderate impact abnormal feature with a weight value of 0.6, the system will directly combine the moderate impact with 0.6. Finally, the first assessment result is formed by combining the mild impact with 0.9 and the moderate impact with 0.6, reflecting both the cumulative effect of multiple abnormalities under mild impact and clearly presenting the independent impact of moderate impact. By employing differentiated computational strategies, we can more accurately quantify the comprehensive impact of a single abnormality on the blood supply status of the flap, providing a more reliable basis for determining the subsequent flap survival risk level.

[0053] Based on the correlation and feature type of the quantitative feature parameters, the impact of synergistic abnormality on the flap blood supply status is determined to obtain the second assessment result, which specifically includes the following steps: Calculate the correlation degree of quantitative feature parameters of different types of abnormalities within the same overlapping region of the flap, and determine whether the synergistic abnormality is characterized as strongly correlated or weakly correlated. If the synergistic abnormality is strongly correlated, the degree of blockage of the flap microcirculation pathway by the strongly correlated abnormality is used to obtain the first synergistic assessment result of flap blood supply. If the synergistic abnormality is weakly correlated, the degree of influence of the weakly correlated abnormality on the local tissue nutrition supply of the flap is determined to obtain the second synergistic assessment result of flap blood supply. The first and second collaborative evaluation results are integrated to form the second evaluation result.

[0054] First, the system focuses on the same region within the flap where abnormal quantitative characteristic parameters are superimposed and distributed. It calculates the correlation between different types of abnormal quantitative characteristic parameters within this region to determine whether the synergistic abnormality is strongly or weakly correlated. Different types of quantitative characteristic parameters correspond to different dimensions of flap blood supply. For example, blood flow velocity reflects the flow efficiency of microcirculation, vascular density reflects the coverage of the vascular network, and oxygenation index reflects the oxygen supply status of the tissue. The system uses statistical analysis to calculate the synchronicity of changes in these parameters within the same superimposed region. If different types of abnormal parameters are highly synchronized in spatial distribution and numerical fluctuations—for example, when blood flow velocity decreases, vascular density and oxygenation index also decrease significantly and synchronously, and the trends of all three are completely consistent—it is determined to be a strongly correlated synergistic abnormality. If different types of abnormal parameters show significantly different trends within the same region—for example, blood flow velocity decreases significantly, but vascular density only decreases slightly, and oxygenation index does not change significantly—it is determined to be a weakly correlated synergistic abnormality.

[0055] When the synergistic anomaly is characterized as strongly correlated, the system focuses on assessing the degree of blockage of the flap's microcirculation pathway, generating the first synergistic assessment result of flap blood supply. A strongly correlated anomaly indicates that blood supply parameters of different dimensions form a synergistic effect in the same area, which directly exacerbates the risk of microcirculation pathway blockage. The system quantifies this using an obstruction degree index B, calculated using the following formula: Where n is the number of anomaly quantization feature parameter types involved in the calculation, N i N represents the actual value of the i-th parameter. i,正常 Here, S represents the normal reference value for this parameter, and S is the proportion of the abnormal area to the total area of ​​the flap. For example, if blood flow velocity, vascular density, and oxygenation index are all simultaneously severely abnormal in the distal region of the flap, with Ni being 40%, 35%, and 30% of the normal reference values, respectively, and the abnormal area covers 35% of the flap area, then the degree of obstruction index is: B = 1 / 3 × [(0.6 + 0.65 + 0.7) × 0.35] ≈ 0.2275, corresponding to severe obstruction, and the first collaborative assessment result is high risk; if it is only moderately abnormal and the coverage area is 15%, then B is approximately 0.08, corresponding to mild obstruction, and the first collaborative assessment result is low risk.

[0056] When the synergistic anomaly is characterized as weakly correlated, the system shifts to assessing its impact on the local tissue nutrient supply of the flap, generating a second synergistic assessment result for flap blood supply. Although weakly correlated anomalies do not constitute a completely synergistic pathway blockage, abnormalities in different types of parameters weaken the nutrient supply to the local tissue from different dimensions. The system quantifies this using a nutrient supply index N, calculated as follows: , where w i The weights for the influence of the i-th parameter on nutrient supply are as follows: blood flow velocity has a weight of 0.5, vascular density has a weight of 0.3, and oxygenation index has a weight of 0.2. For example, if blood flow velocity is 60% of the normal reference value, vascular density is 80% of the normal reference value, and oxygenation index is 90% of the normal reference value, then the nutrient supply index N = 0.5 × 0.4 + 0.3 × 0.2 + 0.2 × 0.1 = 0.28, corresponding to moderate nutrient supply impairment, and the second co-assessment result is medium risk.

[0057] Finally, the system integrates the first and second synergistic assessment results to form a second assessment result. The integration process uses a weighted summation method, with a weight of 0.7 for strongly correlated assessment results and 0.3 for weakly correlated results. This is because strongly correlated abnormalities directly threaten the patency of microcirculatory pathways and have a more critical impact on flap survival. For example, if the first synergistic assessment result is high risk (quantitative value 3) and the second synergistic assessment result is medium risk (quantitative value 2), then the second assessment result is 3 × 0.7 + 2 × 0.3 = 2.7. This value comprehensively reflects the combined impact of synergistic abnormalities on flap blood supply, providing a precise and objective basis for clinically assessing the severity of synergistic blood supply disorders.

[0058] The first synergistic assessment of flap blood supply is obtained by determining the degree of blockage of the flap microcirculation pathway by strongly correlated abnormal features, which specifically includes the following steps: Determine whether strongly correlated abnormal features cause occlusion of the flap microcirculation pathway; If occlusion occurs, obtain the area proportion and location information of the occluded area. Compare the area proportion and location information of the occluded area with the flap microcirculation occlusion and blood supply assessment table to obtain the risk level and intervention recommendations for flap blood supply. Mark the risk level and intervention recommendations for flap blood supply as the first collaborative assessment result. If no occlusion is caused, the influence of strongly correlated abnormal features on the degree of stenosis and blood flow velocity of the flap microcirculation pathway is determined, and two key parameters are obtained: the stenosis ratio of the microcirculation pathway and the blood flow velocity deviation. The stenosis percentage and blood flow velocity deviation values ​​are compared with preset thresholds; If the stenosis percentage and blood flow velocity deviation are both below the preset threshold, it is determined that there is no significant adverse effect on flap blood supply, and the corresponding first collaborative assessment result with no intervention requirement is marked. If either the stenosis percentage or the blood flow velocity deviation is higher than or equal to a preset threshold, it is determined that there is a potential adverse impact on the blood supply of the flap. Based on the abnormal location and characteristic change trend, preventive intervention suggestions and corresponding blood supply risk levels are given, and it is marked as the first collaborative assessment result.

[0059] First, the system determines whether strongly correlated abnormal features cause occlusion of the flap microcirculation pathway based on the spatial distribution and numerical deviation of abnormal quantitative feature parameters. If different types of abnormal parameters cause complete blockage in a certain area of ​​the flap, such as zero vascular density parameter, zero blood flow velocity parameter, and no blood perfusion signal in the area, it is determined that microcirculation pathway occlusion has been caused. If the abnormal parameters only cause reduced blood flow and narrowing of blood vessels, but there is still a continuous blood perfusion channel, it is determined that no occlusion has been caused, and the process of assessing the degree of stenosis begins.

[0060] If the system determines that microcirculatory pathway occlusion has occurred, it obtains the area percentage and location of the occlusion. The area percentage refers to the ratio of the occluded area to the total flap area, while the location information specifies whether the occlusion occurs in a critical location—proximal, middle, or distal—of the flap. The impact of occlusion at different locations on blood supply varies significantly; for example, distal occlusion directly affects blood supply to the fingertip, posing a higher risk. The system then compares the area percentage and location information with a pre-defined flap microcirculatory occlusion and blood supply assessment table. This table, built based on extensive clinical data, links different area percentages and locations of occlusion with corresponding blood supply risk levels and intervention recommendations. For example, if the occlusion area percentage is 25% and the occlusion location is distal to the flap, it would be classified as high-risk after comparing with the assessment table, and the corresponding intervention recommendation would be immediate vascular exploration and recanalization surgery. The system would then mark this risk level and intervention recommendation as the first collaborative assessment result.

[0061] If the system determines that no microcirculatory pathway occlusion has occurred, it then assesses the impact of strongly correlated abnormal characteristics on the degree of stenosis and blood flow velocity in the microcirculatory pathway, focusing on obtaining the stenosis percentage and blood flow velocity deviation. The stenosis percentage is the ratio of the cross-sectional area of ​​the stenotic region to the normal cross-sectional area, reflecting the degree of pathway stenosis; the blood flow velocity deviation is the difference between the blood flow velocity in the abnormal region and the normal blood flow velocity, reflecting the degree of impaired blood flow efficiency. The system compares the obtained stenosis percentage and blood flow velocity deviation values ​​with preset thresholds, which are critical values ​​set based on clinical experience; for example, the stenosis percentage threshold is 30%, and the blood flow velocity deviation threshold is 40% of the normal velocity.

[0062] If both the stenosis percentage and blood flow velocity deviation are below preset thresholds, for example, a stenosis percentage of 20% and a blood flow velocity deviation of 30% of normal velocity, the system will determine that the strongly correlated abnormality has no significant adverse effect on flap blood supply. In this case, the system will mark the corresponding first co-assessment result as requiring no intervention, indicating that only routine monitoring is needed clinically. If either the stenosis percentage or the blood flow velocity deviation is higher than or equal to the preset threshold, for example, a stenosis percentage of 35% and a blood flow velocity deviation of 50% of normal velocity, the system will determine that the strongly correlated abnormality has a potential adverse effect on flap blood supply. In this case, the system will combine the location of the abnormality and the trend of its characteristic changes to provide preventive intervention recommendations and corresponding blood supply risk levels. For example, it may recommend anticoagulation therapy and close monitoring, with a risk level of medium risk. The system will mark this recommendation and risk level as the first co-assessment result, providing precise guidance for early clinical intervention and preventing the worsening of blood supply disorders.

[0063] The second synergistic assessment of flap blood supply is obtained by determining the extent to which weakly correlated abnormal features affect the local tissue nutrition supply of the flap. This includes the following steps: The numerical values ​​of each quantitative characteristic parameter corresponding to the weakly correlated abnormal characterization are compared with the flap single feature and blood supply assessment comparison table to obtain the blood supply impact level corresponding to each abnormal characterization. The degree of damage to local tissue nutrition supply is determined based on the superposition result of each blood supply impact level, and this degree of damage is marked as the second collaborative assessment result.

[0064] First, the system extracts the numerical values ​​of various quantitative characteristic parameters corresponding to weakly correlated abnormalities. These include parameters such as epidermal color, vascular density, blood flow velocity, and oxygenation index; each parameter reflects the nutritional supply potential of the local tissue from different perspectives. Then, the system compares these parameter values ​​one by one against a pre-defined flap single-feature and blood supply assessment comparison table. This comparison table is a standardized reference system built based on extensive clinical data. It links the numerical range of each type of quantitative characteristic parameter with the corresponding blood supply impact level, which is divided into three levels: mild impact, moderate impact, and severe impact, corresponding to slight interference, significant restriction, and severe deficiency of nutritional supply, respectively.

[0065] After completing the comparison, the system obtains the blood supply impact level corresponding to each abnormal manifestation. The system then sums these blood supply impact levels to determine the degree of damage to local tissue nutrient supply. The summation logic is based on a comprehensive judgment of the impact weight of each level on nutrient supply. For example, a mild impact has a weight of 1, a moderate impact has a weight of 2, and a severe impact has a weight of 3. The system sums the weight values ​​corresponding to the impact levels of all abnormal manifestations to obtain a comprehensive impact value, and then determines the final degree of damage based on the range in which this value falls.

[0066] For example, weakly correlated abnormalities include three quantitative parameters: epidermal color corresponds to mild impact, vascular density corresponds to moderate impact, and oxygenation index corresponds to mild impact. The corresponding impact weights are 1, 2, and 1, respectively, resulting in a total impact value of 1+2+1=4. The system's preset damage range is: a total impact value less than or equal to 2 indicates mild damage, 3 to 4 indicates moderate damage, and greater than 4 indicates severe damage. Therefore, in this case, the total impact value of 4 was determined to be moderate damage, and the system marked moderate damage as the second collaborative assessment result. This clearly reflects that although weakly correlated abnormalities do not directly block microcirculation pathways, through the superposition of multi-dimensional effects, they significantly restrict the nutrient supply to local tissues, providing precise quantitative evidence for clinical assessment and intervention.

[0067] A machine vision-based intelligent chip for flap evaluation after finger replantation surgery includes: an image visual acquisition unit, an image feature acquisition unit, a feature parameter extraction unit, a feature judgment unit 1, a feature judgment unit 2, and a grade judgment unit. The image visual acquisition unit is used to acquire static visual images and dynamic microcirculation visual images of the flap after finger replantation under standard visual acquisition conditions to form a flap visual evaluation set, and to select effective images in the flap visual evaluation set as flap images to be evaluated. The image feature acquisition unit is used to acquire the epidermal color features and subcutaneous vascular texture features of the flap in the flap to be evaluated image. Based on the changing trend of the epidermal color features and the distribution pattern of the subcutaneous vascular texture features, it determines whether it is necessary to perform fusion processing on the static and dynamic flap to be evaluated images and marks the corresponding flap to be evaluated image as the flap fusion processing image. The feature parameter extraction unit is used to extract the quantitative feature parameters required for flap evaluation in the flap fusion image. Based on the distribution location of subcutaneous vascular texture features and the type of quantitative feature parameters, it determines whether there is a cooperative abnormality in the flap in the flap fusion image. The feature judgment unit is used to determine that there is no synergistic abnormal characterization of the flap in the flap fusion processing image. Based on the numerical range and feature type of the quantified feature parameters, the influence of a single abnormal characterization on the blood supply status of the flap is judged to obtain the first evaluation result. The second feature judgment unit is used to judge whether there is a cooperative abnormality in the flap in the flap fusion processing image. Based on the correlation degree and feature type of the quantified feature parameters, the influence of the cooperative abnormality on the blood supply status of the flap is judged to obtain the second evaluation result. The rating judgment unit is used to combine the first assessment result and / or the second assessment result to comprehensively determine the survival risk level of the flap.

[0068] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for post-replantation flap evaluation of severed finger based on machine vision, characterized in that, The method includes the following steps: Static visual images and dynamic microcirculation visual images of the flap after finger replantation under standard visual acquisition conditions were collected to form a flap visual evaluation set. Valid images in the flap visual evaluation set were selected as flap images to be evaluated. The epidermal color features and subcutaneous vascular texture features of the flap in the image to be evaluated are obtained. Based on the changing trend of the epidermal color features and the distribution pattern of the subcutaneous vascular texture features, it is determined whether the static and dynamic images of the flap to be evaluated need to be fused. The corresponding images of the flap to be evaluated are marked as flap fusion processing images. Extract the quantitative feature parameters required for flap evaluation from the flap fusion image, and determine whether there is a cooperative abnormality in the flap in the flap fusion image based on the distribution location of subcutaneous vascular texture features and the type of quantitative feature parameters. If there is no synergistic abnormality in the flap fusion image, the first evaluation result is obtained by judging the impact of a single abnormality on the flap blood supply state based on the numerical range and feature type of the quantified feature parameters. If there are synergistic abnormalities in the flap fusion image, the impact of the synergistic abnormalities on the flap blood supply status is determined based on the correlation degree and feature type of the quantified feature parameters to obtain the second evaluation result. The survival risk level of the flap is determined by combining the results of the first assessment and / or the second assessment.

2. The machine vision-based post-replantation flap evaluation method according to claim 1, wherein, Based on the changing trends of epidermal color features and the distribution patterns of subcutaneous vascular texture features, it is determined whether static and dynamic flap images to be evaluated need to be fused. The corresponding flap images to be evaluated are then marked as flap fusion processing images. The specific steps include: If the trend of changes in epidermal color features and the distribution pattern of subcutaneous vascular texture features form a complete correlation in a single flap image to be evaluated, then the flap image to be evaluated is directly marked as a flap fusion processing image. If the trend of epidermal color features is reflected in the static flap image to be evaluated, and the distribution pattern of subcutaneous vascular texture features is reflected in the dynamic flap image to be evaluated, then feature regions are extracted from the static and dynamic flap images to be evaluated respectively, forming a static feature region containing epidermal color features and a dynamic feature region containing subcutaneous vascular texture features. Using the microcirculation flow characteristics of the flap in the image to be evaluated as the association anchor point, the feature matching sites of the static feature region and the dynamic feature region are identified. After edge calibration of the feature matching sites, the static feature region and the dynamic feature region are fused to form the flap fusion image.

3. The machine vision-based post-replantation flap evaluation method according to claim 2, wherein, Using the microcirculation flow characteristics of the skin flap in the image to be evaluated as correlation anchor points, feature matching sites between static and dynamic feature regions are identified. After edge calibration of the feature matching sites, the static and dynamic feature regions are fused to form a skin flap fusion image. The specific steps include: Based on the core distribution area of ​​microcirculation flow characteristics, determine whether the number of overlapping matching sites between static and dynamic feature areas reaches a preset threshold. If the number of overlapping matching sites reaches a preset threshold, the overlapping matching site is marked as the core matching area. The static feature area and dynamic feature area are then calibrated and fused to form a flap fusion image based on the core matching area. If the number of overlapping matching sites does not reach the preset threshold, the edge feature points of the static feature region and the dynamic feature region are extracted and paired. If the paired edge feature points can form a continuous flap feature contour, then the region is fused based on the feature contour to form a flap fusion image.

4. The machine vision-based post-replantation flap evaluation method according to claim 3, wherein, Based on the distribution location of subcutaneous vascular texture features and the type of quantified feature parameters, the determination of whether there are cooperative abnormalities in the flap fusion image includes the following steps: If the flap in the fused image has only one type of abnormal quantified feature parameter, then the flap is determined to have no cooperative abnormality representation. If there are at least two types of abnormal quantization feature parameters in the flap fusion image, then based on the distribution location of the subcutaneous vascular texture features, it is determined whether the different types of abnormal quantization feature parameters form an overlapping distribution in the same area of ​​the flap. If different types of abnormal quantitative feature parameters do not form an overlapping distribution in the same area of ​​the flap, it is determined that the flap does not have a synergistic abnormal characterization; if different types of abnormal quantitative feature parameters form an overlapping distribution in the same area of ​​the flap, it is determined that the flap has a synergistic abnormal characterization.

5. The machine vision-based post-replantation flap evaluation method according to claim 4, wherein, Based on the numerical range and feature type of the quantified feature parameters, the impact of a single abnormal feature on the blood supply status of the flap is determined to obtain the first assessment result, which specifically includes the following steps: By comparing the quantitative feature parameter values ​​corresponding to each single abnormality with the flap single feature and blood supply assessment comparison table, the blood supply impact level and impact weight value corresponding to each single abnormality are obtained. The impact on blood supply is categorized into mild, moderate, and severe. Based on the blood supply impact level and impact weight value of each individual abnormality, the impact of a single abnormality on the blood supply status of the flap is calculated to obtain the first assessment result.

6. The machine vision-based post-replantation flap evaluation method according to claim 5, wherein, Based on the blood supply impact level and impact weight value of each individual abnormality, the impact of each individual abnormality on the blood supply status of the flap is calculated to obtain the first assessment result, which specifically includes the following steps: Determine whether there are multiple single abnormal features of the same blood supply level; If there are no multiple single abnormal manifestations with the same blood supply impact level, the blood supply impact level of each single abnormal manifestation and its corresponding impact weight value are directly combined to form the first assessment result. If there are multiple single abnormal manifestations of the same blood supply impact level, the impact weight values ​​of all single abnormal manifestations under that level are summed to obtain the comprehensive weight value of that blood supply impact level. Each blood supply impact level is combined with the corresponding comprehensive weight value to form the first assessment result.

7. The machine vision-based post-replantation flap evaluation method according to claim 6, wherein, Based on the correlation and feature type of the quantitative feature parameters, the impact of synergistic abnormality on the flap blood supply status is determined to obtain the second assessment result, which specifically includes the following steps: Calculate the correlation degree of quantitative feature parameters of different types of abnormalities within the same overlapping region of the flap, and determine whether the synergistic abnormality is characterized as strongly correlated or weakly correlated. If the synergistic abnormality is strongly correlated, the degree of blockage of the flap microcirculation pathway by the strongly correlated abnormality is used to obtain the first synergistic assessment result of flap blood supply. If the synergistic abnormality is weakly correlated, the degree of influence of the weakly correlated abnormality on the local tissue nutrition supply of the flap is determined to obtain the second synergistic assessment result of flap blood supply. The first and second collaborative evaluation results are integrated to form the second evaluation result.

8. The machine vision-based post-replantation flap evaluation method according to claim 7, wherein, The first synergistic assessment of flap blood supply is obtained by determining the degree of blockage of the flap microcirculation pathway by strongly correlated abnormal features, which specifically includes the following steps: Determine whether strongly correlated abnormal features cause occlusion of the flap microcirculation pathway; If occlusion occurs, obtain the area proportion and location information of the occluded area. Compare the area proportion and location information of the occluded area with the flap microcirculation occlusion and blood supply assessment table to obtain the risk level and intervention recommendations for flap blood supply. Mark the risk level and intervention recommendations for flap blood supply as the first collaborative assessment result. If no occlusion is caused, the influence of strongly correlated abnormal features on the degree of stenosis and blood flow velocity of the flap microcirculation pathway is determined, and two key parameters are obtained: the stenosis ratio of the microcirculation pathway and the blood flow velocity deviation. The stenosis percentage and blood flow velocity deviation values ​​are compared with preset thresholds; If the stenosis percentage and blood flow velocity deviation are both below the preset threshold, it is determined that there is no significant adverse effect on flap blood supply, and the corresponding first collaborative assessment result with no intervention requirement is marked. If either the stenosis percentage or the blood flow velocity deviation is higher than or equal to a preset threshold, it is determined that there is a potential adverse impact on the blood supply of the flap. Based on the abnormal location and characteristic change trend, preventive intervention suggestions and corresponding blood supply risk levels are given, and it is marked as the first collaborative assessment result.

9. The machine vision-based post-replantation flap evaluation method according to claim 8, wherein, The second synergistic assessment of flap blood supply is obtained by determining the extent to which weakly correlated abnormal features affect the local tissue nutrition supply of the flap. This includes the following steps: The numerical values ​​of each quantitative characteristic parameter corresponding to the weakly correlated abnormal characterization are compared with the flap single feature and blood supply assessment comparison table to obtain the blood supply impact level corresponding to each abnormal characterization. The degree of damage to local tissue nutrition supply is determined based on the superposition result of each blood supply impact level, and this degree of damage is marked as the second collaborative assessment result.

10. The intelligent chip for post-replantation of severed finger skin flap evaluation based on machine vision is applied to the method for post-replantation of severed finger skin flap evaluation based on machine vision as claimed in claims 1-9, characterized in that, The intelligent chip includes: an image visual acquisition unit, an image feature acquisition unit, a feature parameter extraction unit, a feature judgment unit one, a feature judgment unit two, and a level judgment unit; The image visual acquisition unit is used to acquire static visual images and dynamic microcirculation visual images of the flap after finger replantation under standard visual acquisition conditions to form a flap visual evaluation set, and to select effective images in the flap visual evaluation set as flap images to be evaluated. The image feature acquisition unit is used to acquire the epidermal color features and subcutaneous vascular texture features of the flap in the flap to be evaluated image. Based on the changing trend of the epidermal color features and the distribution pattern of the subcutaneous vascular texture features, it determines whether it is necessary to perform fusion processing on the static and dynamic flap to be evaluated images and marks the corresponding flap to be evaluated image as the flap fusion processing image. The feature parameter extraction unit is used to extract the quantitative feature parameters required for flap evaluation in the flap fusion image. Based on the distribution location of subcutaneous vascular texture features and the type of quantitative feature parameters, it determines whether there is a cooperative abnormality in the flap in the flap fusion image. The feature judgment unit is used to determine that there is no synergistic abnormal characterization of the flap in the flap fusion processing image. Based on the numerical range and feature type of the quantified feature parameters, the influence of a single abnormal characterization on the blood supply status of the flap is judged to obtain the first evaluation result. The second feature judgment unit is used to judge whether there is a cooperative abnormality in the flap in the flap fusion processing image. Based on the correlation degree and feature type of the quantified feature parameters, the influence of the cooperative abnormality on the blood supply status of the flap is judged to obtain the second evaluation result. The rating judgment unit is used to combine the first assessment result and / or the second assessment result to comprehensively determine the survival risk level of the flap.

Images