Early-stage marginal osteophyte leakage-oriented knee osteoarthritis local sign enhanced determination method and system

Abstract

This invention relates to the field of medical image processing technology, and discloses a method and system for determining the enhancement of local signs in knee osteoarthritis that addresses the missed detection of early marginal osteophytes. The method includes: acquiring routine outpatient anteroposterior X-ray images of the knee joint; performing bone segmentation and boundary continuity optimization; fitting a reference contour based on the continuous cortical bone contour set; constructing a local outward-protruding enhancement response; screening candidate abnormal arc segments; determining the enhancement of marginal osteophytes in the candidate abnormal arc segments; and outputting a structured determination result. Compared to existing methods that mainly rely on joint space changes or single grayscale features for interpretation, especially when early marginal osteophytes are short, this invention addresses the technical problem of failing to stably identify small outward-protruding osteophytes. By constructing a reference contour and introducing normal offset and local curvature to jointly characterize the degree of bone edge protrusion, this invention achieves stable enhancement determination of early small marginal osteophytes.

Description

Technical Field

[0001] This invention relates to the field of medical image processing technology, and in particular to a method and system for enhancing the assessment of local signs in knee osteoarthritis that addresses the early omission of marginal osteophytes. Background Technology

[0002] Currently, routine outpatient knee X-ray images remain the most commonly used source of imaging data for early screening and auxiliary interpretation of knee osteoarthritis. Clinically, knee osteoarthritis is usually graded by observing signs such as joint space narrowing, subchondral bone sclerosis, and osteophyte formation. However, in the early stages of KL grade 0 to 1, the lesion often presents as a slight outward protrusion at the edge of the femoral condyle or tibial plateau. This protrusion is short, has incomplete boundaries, and low grayscale contrast, appearing as a weak contrast feature on routine imaging.

[0003] Existing technologies often rely on overall structural changes or single grayscale differences for interpretation. For example, they extract bone structure boundaries through global threshold segmentation or obtain cortical bone contours using simple edge detection operators. Some methods also use convolutional neural networks to segment bone and joint regions before classification. However, these methods typically focus more on obvious signs of degeneration and lack targeted representation methods for localized protruding structures close to the cortical bone edge that are subtle in shape and change slowly, making it difficult to reliably identify early, small marginal osteophytes. Furthermore, during actual outpatient imaging, differences in patient positioning, slight knee joint rotation, and unstable imaging conditions can easily lead to projection distortion or grayscale fluctuations in the bone edge contour, further obscuring small osteophytes or misjudging them as normal bone edge undulations.

[0004] For example, when the gray-scale distribution at the edge of the bone cortex is uneven or there is strong interference from overlapping soft tissues, detection methods based on fixed thresholds or single edge intensity are difficult to distinguish between real outward protrusions and noise disturbances. When there is a certain pose deviation in the image, the local curvature change of the bone edge may come from projection factors rather than structural abnormalities. Existing methods cannot effectively distinguish such cases, resulting in prominent problems of missed or false judgments.

[0005] Therefore, there is an urgent need for a method that can still enhance and reliably determine the signs of local bone protrusion even under conditions of weak osteophyte imaging, image noise interference, and posture deviation, in order to improve the accuracy and consistency of early screening for knee osteoarthritis. Summary of the Invention

[0006] To address the aforementioned technical shortcomings, the purpose of this invention is to propose an enhanced method for identifying local signs of knee osteoarthritis that addresses the missed detection of early marginal osteophytes. This method aims to solve the technical problem that existing methods, which mainly rely on changes in joint space or single grayscale features for interpretation, are unable to reliably identify small, protruding osteophytes, especially when the length of early marginal osteophytes is short.

[0007] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: The present invention provides a method for enhancing the determination of local signs of knee osteoarthritis for early detection of missed marginal osteophytes.

[0008] The enhanced assessment method for local signs of knee osteoarthritis aimed at early detection of marginal osteophytes includes:

[0009] Step S10: Obtain a routine outpatient knee joint anteroposterior X-ray image, perform local standardization processing based on the routine outpatient knee joint anteroposterior X-ray image, and output a local standardization parameter set;

[0010] Step S20: Based on the local standardized parameter set, the target bone edge is extracted using the bone tissue segmentation and boundary continuity optimization method, and a continuous bone cortex contour set is output;

[0011] Step S30: Based on the continuous cortical bone contour set, a reference contour fitting and differential enhancement method is used to perform local protrusion preprocessing, and output the local protrusion enhancement response set;

[0012] Step S40: Based on the local protrusion enhancement response set, construct the marginal osteophyte risk parameters and output the marginal osteophyte risk parameter set;

[0013] Step S50: Based on the marginal osteophyte risk parameter set, a cost-sensitive classification method is used to determine the enhancement of marginal osteophytes, and the determination result of the enhancement of marginal osteophytes is output.

[0014] Preferably, step S10, which involves performing local standardization processing on the routine outpatient knee joint anteroposterior X-ray image and outputting a local standardized parameter set, specifically includes:

[0015] Step S101: Perform percentile grayscale normalization on the routine outpatient knee joint anteroposterior X-ray image to obtain a standard grayscale image;

[0016] Step S102: Based on the standard grayscale image, extract the femoral distal positioning region, tibial plateau positioning region, tibial spine positioning region and fibular head positioning region to construct a local anatomical coordinate system for the knee joint;

[0017] Step S103: Generate local scale parameters, posture deviation parameters, and mode selection parameters based on the local anatomical coordinate system of the knee joint to form the local standardized parameter set;

[0018] Wherein, the percentile grayscale normalization process satisfies:

[0019] ;

[0020] in, Represents the original grayscale value. This represents the grayscale value at the 100th percentile. This represents the 99th percentile gray value. This represents the normalized grayscale value. This indicates a bounded truncation function, used to restrict input values ​​to a preset range.

[0021] Preferably, step S20, which involves extracting the target bone edge using a bone tissue segmentation and boundary continuity optimization method based on the local standardized parameter set, and outputting a continuous bone cortical contour set, specifically includes:

[0022] Step S201: Based on the local standardized parameter set, a convolutional neural network segmentation method is used to segment the bone body of the distal femur region and the proximal tibia region to obtain the bone body segmentation result;

[0023] Step S202: Based on the bone segmentation results, predict the probability of the bone cortex boundary and use an active contour model to optimize the boundary continuity to obtain the initial bone cortex contour.

[0024] Step S203: Based on the initial cortical bone contour, perform continuity checks on adjacent arc segments and correct abnormal fractures to obtain the continuous cortical bone contour set.

[0025] Preferably, step S30, which involves performing local protrusion preprocessing based on the continuous cortical bone contour set using a reference contour fitting and differential enhancement method, and outputting a local protrusion enhancement response set, specifically includes:

[0026] Step S301: Establish arc length parameters for each bone edge arc segment in the continuous bone cortex contour set, and construct a reference contour using a cubic spline fitting method;

[0027] Step S302: Obtain the actual bone edge contour, and generate local difference feature parameters based on the normal offset, local curvature and gray-level gradient difference between the reference contour and the actual bone edge contour;

[0028] Step S303: Based on the local differential feature parameters, perform outward protrusion enhancement and physiological gradual contour suppression to obtain the local outward protrusion enhancement response set.

[0029] Preferably, the normal offset of the actual bone edge contour relative to the reference contour satisfies:

[0030] ;

[0031] in, Indicates the reference profile in the arc length parameter The coordinates of the point, This indicates the actual bone edge contour in the arc length parameter. The coordinates of the points are all represented using two-dimensional plane coordinates. Indicates the reference profile in the arc length parameter The outward normal unit vector at a given location satisfies:

[0032] ;

[0033] in, , These represent the reference profile in terms of arc length parameter. The first derivative at that point;

[0034] This represents the transpose of the outward normal unit vector;

[0035] This represents the normal offset of the actual bone edge contour relative to the reference contour, when... When it indicates an outward protrusion, This indicates that no outward protrusion has occurred;

[0036] The local curvature satisfies:

[0037] ;

[0038] in, Represents arc length parameter Local curvature at that point;

[0039] , These represent the actual bone edge contour in terms of arc length parameters. The second derivative at point .

[0040] Preferably, step S40, which involves constructing marginal osteophyte risk parameters based on the local protrusion enhancement response set and outputting the marginal osteophyte risk parameter set, specifically includes:

[0041] Step S401: Construct artifact constraint regions for the patellar adjacent overlapping region, soft tissue high-contrast edge region, and bone edge overlapping region based on the continuous bone cortex contour set;

[0042] Step S402: Based on the artifact constraint region, perform artifact scoring and dual threshold screening on the candidate abnormal arc segments in the local outward enhancement response set, and output the candidate abnormal arc segments;

[0043] Step S403: Extract the outward protrusion amplitude parameter, outward protrusion length parameter, contour deviation parameter, grayscale change parameter, artifact scoring parameter and image quality parameter from the candidate abnormal arc segment to form the marginal osteophyte risk parameter set.

[0044] Preferably, step S50, which involves using a cost-sensitive classification method to determine marginal osteophyte enhancement based on the marginal osteophyte risk parameter set and outputting the marginal osteophyte enhancement determination result, specifically includes:

[0045] Step S501: Construct a judgment feature vector for candidate abnormal arc segments based on the marginal osteophyte risk parameter set, and generate the positive probability of marginal osteophytes;

[0046] Step S502: Introduce an adaptive judgment threshold and compare the positive probability of marginal osteophytes with the adaptive judgment threshold to obtain the positive judgment result of marginal osteophytes;

[0047] Step S503: Generate a structured judgment result based on the positive determination result of marginal osteophytes, including candidate location, degree of protrusion, degree of contour abnormality, and judgment confidence.

[0048] Wherein, the adaptive determination threshold satisfies:

[0049] ;

[0050] in, This represents the adaptive threshold for determining the probability of marginal osteocheilitis. This represents the base threshold, which is a fixed threshold obtained through statistical analysis of training samples. , and These represent weighting coefficients, which are used to adjust the degree of influence of image noise, pose deviation, and image quality on the decision threshold.

[0051] This invention also provides a local sign enhancement assessment system for knee osteoarthritis aimed at early detection of marginal osteophytes, including:

[0052] The standardized parameter construction module is used to acquire routine outpatient knee joint anteroposterior X-ray images, perform local standardization processing based on the routine outpatient knee joint anteroposterior X-ray images, and output a local standardized parameter set.

[0053] The bone edge extraction module is used to perform target bone edge extraction based on the local standardized parameter set using bone tissue segmentation and boundary continuity optimization methods, and output a continuous bone cortex contour set;

[0054] The external protrusion enhancement module is used to perform local external protrusion preprocessing based on the continuous cortical bone contour set using a reference contour fitting and differential enhancement method, and output a local external protrusion enhancement response set;

[0055] The risk construction module is used to construct risk parameters for marginal osteophytes based on the local protrusion enhancement response set and output the risk parameter set for marginal osteophytes.

[0056] The enhancement determination module is used to perform edge osteophyte enhancement determination based on the edge osteophyte risk parameter set using a cost-sensitive classification method, and output the edge osteophyte enhancement determination result.

[0057] The present invention also provides a device for enhancing the determination of local signs of knee osteoarthritis for early-stage marginal osteophyte omission, comprising: a memory, a processor, and a program for enhancing the determination of local signs of knee osteoarthritis for early-stage marginal osteophyte omission stored in the memory and executable on the processor. When the program for enhancing the determination of local signs of knee osteoarthritis for early-stage marginal osteophyte omission is executed by the processor, the method for enhancing the determination of local signs of knee osteoarthritis for early-stage marginal osteophyte omission is realized.

[0058] The present invention also provides a computer program product, including a local sign enhancement determination program for knee osteoarthritis with early marginal osteophyte omission, wherein the local sign enhancement determination program for knee osteoarthritis with early marginal osteophyte omission is executed by a processor to implement the local sign enhancement determination method for knee osteoarthritis with early marginal osteophyte omission.

[0059] The beneficial effects of this invention are as follows: By constructing a reference contour and introducing a joint characterization method of normal offset and local curvature, this invention can quantitatively characterize the tiny outward protrusions at the edge of the bone cortex. At the same time, it can enhance the processing by combining the gray-level gradient difference between the inside and outside of the bone cortex. This allows for the effective differentiation between early marginal osteophytes and normal bone edge undulations even when the osteophyte length is short and the gray-level change is weak, thereby significantly reducing the probability of missed detection in the early stage.

[0060] This invention constructs an adaptive judgment threshold by introducing image noise parameters, posture deviation parameters, and image quality parameters during the judgment process. This allows the judgment results to dynamically adjust the judgment criteria based on the imaging quality and positional differences of routine outpatient X-ray images. In cases of low-quality images or large posture deviations, the judgment strictness is automatically increased, thereby reducing misjudgments and omissions and improving the stability and consistency of the overall judgment results. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0062] Figure 1 This is a flowchart illustrating the first embodiment of the method for enhancing the assessment of local signs of knee osteoarthritis in the context of early-stage marginal osteophyte diagnoses according to the present invention.

[0063] Figure 2 This is a schematic diagram of the original local bone edge scene of the first embodiment of the method for enhancing the determination of local signs of knee osteoarthritis for early omission of marginal osteophytes according to the present invention.

[0064] Figure 3 This is a schematic diagram of the normal offset distribution of the first embodiment of the method for enhancing the determination of local signs of knee osteoarthritis for early marginal osteophyte omission.

[0065] Figure 4 This is a schematic diagram of the curvature abnormality and grayscale gradient difference in the first embodiment of the method for enhancing the local signs of knee osteoarthritis in the early stage of missed osteophyte detection of the present invention.

[0066] Figure 5 This is a schematic diagram showing the enhanced local protrusion signs of the first embodiment of the method for enhancing the local signs of knee osteoarthritis in the early stage of missed diagnosis of marginal osteophytes according to the present invention. Detailed Implementation

[0067] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0068] Example 1: As Figure 1 The diagram shown is a flowchart illustrating the first embodiment of the method for enhancing the assessment of local signs of knee osteoarthritis for early detection of marginal osteophytes, according to the present invention.

[0069] In the first embodiment, the method for enhancing the assessment of local signs of knee osteoarthritis aimed at early detection of marginal osteophytes includes:

[0070] Step S10: Obtain a routine outpatient knee joint anteroposterior X-ray image, perform local standardization processing based on the routine outpatient knee joint anteroposterior X-ray image, and output a local standardization parameter set.

[0071] It should be noted that the aforementioned routine outpatient knee X-ray image refers to a medical image obtained by taking an anteroposterior view of the knee joint of the subject in a primary hospital, physical examination center, or orthopedic clinic using digital X-ray imaging equipment. The image range must at least cover the distal femur, tibial plateau, tibial spine, proximal tibial cortical margin, and the area adjacent to the fibular head. The local standardization processing does not involve simply scaling or enhancing the overall grayscale of the entire image. Instead, it establishes a local processing range around the femoral condyle margin and tibial plateau margin, where early marginal osteophytes are prevalent, and performs grayscale standardization, local scale unification, and posture deviation quantification within this local processing range. Specifically, the original X-ray image can first be normalized to a percentile gray level, mapping the original gray values ​​that are greatly affected by exposure differences to a stable gray level range; then, the femoral distal end positioning region, tibial plateau positioning region, tibial spine positioning region, and fibular head positioning region can be extracted from the standard gray level image; subsequently, a local anatomical coordinate system of the knee joint can be established based on the relative position between the femoral distal end and the tibial plateau, and local scale parameters, posture deviation parameters, and mode selection parameters can be generated, thereby forming the local standardized parameter set.

[0072] Understandably, the local standardized parameter set is used to provide a unified reference for subsequent steps. Since the acquisition conditions for routine outpatient knee X-ray images are not entirely consistent, different equipment exposure parameters, detector responses, patient body type, standing posture, and slight knee rotation can all cause changes in the appearance of the bone margin. For images of mid-to-late stage osteoarthritis, obvious osteophytes or significant narrowing of the joint space are relatively easy to observe even at different image qualities; however, for early samples in the transitional stage from KL 0 to 1, marginal osteophytes typically appear as short, protruding areas attached to the outer edge of the bone cortex, with insignificant local grayscale changes and relatively short protrusion length. If local scale and posture constraints are not established first, subsequent judgments based solely on the original pixel distance or original grayscale difference can easily lead to interpreting protrusions of the same size at different magnifications as different degrees of severity, or mistaking projection expansion caused by slight rotation for osteophytes. Therefore, step S10 uses the local standardized parameter set to confine the subsequent identification objects within a stable and comparable local coordinate framework.

[0073] It should be understood that, compared to traditional full-image grayscale equalization, ordinary histogram enhancement, or full-image scaling, the improvement in this step lies in shifting the starting point of image processing from "visual enhancement of the entire image" to "constructing the basis for local determination of the target bone edge." Traditional full-image enhancement typically improves the visual contrast of the entire image, but it does not distinguish between the cortical bone edge, soft tissue areas, and background noise areas, easily amplifying noise points and soft tissue shadows while enhancing bone edge details. Traditional scaling only changes the image pixel size and cannot eliminate the influence of differences in femoral condyle width and shooting distance among different patients on the subsequent calculation of the lateral protrusion amplitude. This step uses anatomical regions such as the distal femur and tibial plateau as a basis to generate scale parameters and pose deviation parameters directly related to local bone edge determination. This allows subsequent bone edge extraction, reference contour fitting, and positive probability determination to all use the same set of standardized parameters, thereby improving the stability of early weakly visualized osteophyte identification.

[0074] For example, in an anteroposterior X-ray image of the knee joint collected at a primary care clinic, the femoral condyle region has a higher grayscale value, while the medial edge of the tibial plateau has a lower grayscale value, and there is also local dark noise in the soft tissue background. If a fixed grayscale threshold is directly used for edge extraction, the femoral edge may be over-expanded, while small protrusions near the edge of the medial tibial plateau may be missed. After adopting this step, percentile grayscale normalization can be used to reduce the impact of abnormal bright spots and dark noise on the overall grayscale range; then, the transverse distance between the medial and lateral femoral condyles is extracted as a local scale parameter, for example, the femoral condyle width is 180 pixels; at the same time, the angular deviation between the femoral reference axis and the tibial reference axis is calculated, for example, 2.5°, and the overlap offset of the fibular head relative to the lateral edge of the tibia is calculated, for example, 12 pixels. The resulting local normalized parameter set not only contains the standard grayscale image, but also local scale, pose deviation, and subsequent mode selection information, so that subsequent steps can judge early osteophytes according to the "proportion of outward protrusion relative to the width of the femoral condyle" rather than simply pixel values.

[0075] Step S20: Based on the local standardized parameter set, the target bone edge is extracted using the bone tissue segmentation and boundary continuity optimization method, and a continuous bone cortex contour set is output.

[0076] It should be noted that the target bone edge extraction refers to extracting the cortical bone boundary related to the early determination of marginal osteophytes from routine outpatient knee X-ray images under the constraints of the local normalized parameter set output in step S10. The target bone edge includes at least the edge of the medial femoral condyle, the edge of the lateral femoral condyle, the edge of the medial tibial plateau, and the edge of the lateral tibial plateau. In specific implementation, the search range of the distal femoral region and the proximal tibial region can be determined first based on the local normalized parameter set; then, the bone body region of the distal femoral region and the proximal tibial region can be segmented using a convolutional neural network segmentation method to obtain the bone body segmentation result; subsequently, the probability of the cortical bone boundary is predicted in the boundary neighborhood of the bone body segmentation result; finally, the boundary continuity of the region with a high boundary probability is optimized using an active contour model, and local breaks, local burrs, or short-distance misalignments are corrected by checking the continuity of adjacent arc segments to obtain the continuous cortical bone contour set. The continuous cortical bone contour set is preferably represented as a sequence of two-dimensional coordinate points arranged in arc length order to facilitate subsequent reference contour fitting, normal offset calculation, and curvature calculation.

[0077] Understandably, early marginal osteophytes, due to their small size, are usually attached to the edge of the bone cortex, and their grayscale is similar to that of the bone cortex. If only Canny edge detection, Sobel gradient detection, or fixed threshold segmentation is used, it is easy to mix the edges of the bone cortex, soft tissue, patellar projection, and noisy edges. In this case, even if the curvature or protrusion amplitude is calculated subsequently, misjudgment may occur because the input contour contains false edges. This step first defines the target bone body region through bone tissue segmentation, and then forms a continuous contour through boundary continuity optimization, so that the object processed in subsequent steps is closer to the real bone cortex boundary, reducing abnormal responses introduced by non-target structures.

[0078] It should be understood that, compared to traditional single-threshold segmentation or single-edge operator extraction, the improvements in this step for weakly visualized early osteophyte formation are mainly reflected in two aspects. First, this step does not directly search for all strong edges in the entire image. Instead, it segments the distal femur and proximal tibia regions under the local anatomical coordinates and scale constraints determined in step S10, thereby reducing the interference of soft tissue shadows, background noise, and image border regions on the target bone edge. Second, this step does not directly use the outer boundary of the segmentation result as the final bone edge. Instead, it further performs boundary probability prediction and active contour model optimization to ensure that the cortical bone edge maintains directional and morphological continuity within the local arc segment. Through the above processing, contour breaks caused by local gray-level unevenness can be avoided, as well as abnormal contour protrusions caused by local noise spurs being mistaken for early osteophytes.

[0079] For example, in an anteroposterior X-ray image of the knee joint, there is a slight soft tissue overlay at the edge of the medial tibial plateau. Conventional edge operators might extract an additional soft tissue edge on the outer side of the actual cortical bone. Directly calculating the protrusion based on this edge could lead to false positives. In this step, the proximal tibial body is first segmented based on the tibial plateau localization region obtained in step S10, yielding the bone segmentation result. Subsequently, cortical bone boundary probability prediction is performed only within the neighborhood of the outer cortical bone boundary, weakening the soft tissue edge due to the lack of bone body support. When a 5-8 pixel break occurs at the edge of the tibial plateau due to low contrast, the active contour model can compensate by incorporating the anterior and posterior arc directions, generating a continuous cortical bone contour of the medial tibial plateau. This continuous cortical bone contour set serves as the input to step S30, significantly reducing abnormal biases during subsequent reference contour fitting.

[0080] Step S30: Based on the continuous cortical bone contour set, a reference contour fitting and differential enhancement method is used to perform local protrusion preprocessing and output a local protrusion enhancement response set.

[0081] It should be noted that the reference contour fitting and differential enhancement method refers to constructing a reference contour that reflects the normal orientation of each target bone edge arc segment in a continuous bone cortex contour set, and then performing a differential comparison between the actual bone edge contour and the reference contour to obtain a local outward protrusion enhancement response. In specific implementation, an arc length parameter s can be established for the target bone edge arc segment, and adjacent bone edge sub-arc segments within a preset range on both sides of the target point can be used as the fitting basis. A cubic spline fitting is then used to obtain the reference contour B(s). Subsequently, the actual bone edge contour is represented as C(s), and the normal offset d(s) of C(s) relative to the outward normal direction of B(s) is calculated. Furthermore, the local curvature of the actual bone edge contour and the gray-level gradient difference between the inner and outer cortex are calculated. Based on this, the normal offset, the local curvature anomaly, and the gray-level gradient difference are fused to generate a set of local outward protrusion enhancement responses distributed along the bone edge arc length.

[0082] Understandably, the imaging manifestations of early-stage micro-marginal osteophytes are often not a single "obvious white block-like protrusion," but rather a combination of multiple weak signs, including slight deviation in the outward normal direction, increased curvature of the local bone margin, and abnormal grayscale transition between the inner and outer cortical bone. Relying on only one sign can lead to unstable identification results. For example, relying solely on grayscale difference is susceptible to exposure variations and noise; relying solely on curvature can easily misinterpret the natural curvature of the normal tibial plateau edge as abnormal; and relying solely on the outward protrusion distance may misinterpret the overall outward expansion caused by patient positioning deviation as osteophytes. This step provides a local normal baseline by referencing the contour, and then jointly calculates the outward protrusion direction, curvature abnormalities, and grayscale changes at the same location. This allows for the enhancement of micro-osteophytes that are truly attached to the edge of the cortical bone, while suppressing the normal, slowly changing bone margin morphology.

[0083] It should be understood that, compared to traditional methods that directly calculate the curvature of the bone edge contour or directly enhance the image grayscale, the improvement in this step lies in introducing "reference contour difference" to determine whether the abnormality occurs relative to the local normal bone edge. Traditional curvature calculation lacks local baseline constraints, and the physiological curvature inherent in the tibial plateau edge may produce a higher curvature; traditional grayscale enhancement lacks constraints on the outward protrusion direction, and overall brightening of the bone cortex or local noise enhancement may be mistaken for osteophytes. In this step, the reference contour is fitted by relatively stable adjacent sub-arcs on both sides of the target arc segment, which can represent the expected direction of the bone edge when no local protrusion occurs; only when the actual bone edge deviates from the reference contour in the outward normal direction, accompanied by local curvature abnormalities and changes in grayscale gradient difference, does a higher local protrusion enhancement response occur. Through this processing, it is possible to more effectively distinguish between "normal smooth bone edge," "smooth outward expansion caused by projection," and "early marginal osteophyte protrusion."

[0084] For example, such as Figure 2 As shown, Figure 2 This image is used to illustrate the distribution of the continuous cortical bone contour received in step S30 within the original local X-ray scene. In this image, the actual bone edge contour is located near the highlighted boundary of the cortical bone, but soft tissue grayscale fluctuations, local overlapping shadows, and imaging noise exist around it, making it difficult to clearly determine whether early micro-protrusions exist at the local bone edge when observed directly with the naked eye. Especially in routine outpatient knee X-ray images, early marginal osteophytes typically do not appear as obvious blocky high-density shadows, but rather form short-distance, low-amplitude local expansions only at the tibial plateau edge or femoral condyle edge. Figure 2By overlaying the actual bone edge contour with the initial position of the reference contour, it can be clearly seen that this step does not directly enhance all bright edges in the entire image. Instead, it establishes a local analysis range around the target bone edge arc segment based on the continuous cortical bone contour output in step S20. Therefore, this step can limit the subsequent processing object to the vicinity of the actual cortical bone edge, reducing the influence of soft tissue texture, background noise, and non-target bone projections on the preprocessing of local protrusions, providing a more stable input basis for subsequent reference contour fitting and differential enhancement. Figure 3 As shown, Figure 3 As shown, Figure 3 This is used to display the normal offset d(s) distributed along the arc length parameter s. The normal offset characterizes the degree of deviation of the actual bone edge contour from the reference contour in the outward normal direction. When d(s) is greater than zero and forms a peak in a local area, it indicates that there is a morphological tendency for the corresponding bone edge location to protrude outwards. Figure 3 It can be seen that the d(s) near the suspected early osteophyte location is significantly higher than that of the surrounding bone margin arc segment, while the d(s) in other areas is generally at a lower level, indicating that the abnormality is not a general outward expansion of the entire bone margin, but rather a localized outward displacement. This technique is significant in distinguishing between "true outward protrusions" and "outward displacement of the contour caused by overall posture deviation." If only the original image is observed, small osteophytes might be considered as natural unevenness of the bone margin; however, through the quantitative distribution of d(s), the local outward protrusion of the bone margin can be transformed into a numerical signal that continuously varies along the arc length, enabling subsequent steps to clearly locate the location of the outward protrusion and providing a quantitative basis for screening candidate abnormal arc segments. Figure 4 As shown, Figure 4 As shown, Figure 4 This is used to display the distribution of local curvature abnormalities and the gray-scale gradient difference between the inner and outer cortical bone. Local curvature abnormalities reflect the degree to which the target bone edge changes from a gentle slope to a pointed or curved shape within a local area, while the gray-scale gradient difference between the inner and outer cortical bone reflects whether abnormal changes have occurred in the gray-scale transition between the two sides of the bone edge. Figure 4 As can be seen, near the suspected early marginal osteophyte location, both the abnormal curvature and the gray-level gradient difference show an increasing trend, indicating that this location not only has morphological protrusion but also local changes in the gray-level transition between the inner and outer cortex. Compared to judging solely based on normal offset, the composite features reflected in this image can further eliminate some non-osteophyte interference. For example, if a region only shows increased gray-level but no abnormality in bone edge curvature, it may be imaging noise or overall enhancement of the bone cortex; if a region shows increased curvature but no significant gray-level gradient difference, it may be normal minor undulations in the bone edge or fluctuations in contour extraction. This step uses abnormal curvature and gray-level gradient difference as auxiliary evidence, which can improve the reliability of judging true early marginal osteophytes. Figure 5 As shown, Figure 5 As shown, Figure 5 This technique is used to demonstrate the overall technical effect of mapping the enhanced response of localized protrusions back to the actual bone margin contour. The figure uses color intensity to represent the magnitude of the enhanced response at different bone margin locations, and provides candidate intervals and key indicator summaries at the peak response locations. It can be seen that the enhanced response does not increase uniformly across the entire bone margin, but is concentrated in the short arc segment where suspected small marginal osteophytes are located. This indicates that this step can highlight the location of localized abnormal protrusions while preserving the overall structural continuity of the bone margin.

[0085] Step S40: Based on the local protrusion enhancement response set, perform marginal osteophyte risk parameter construction and output the marginal osteophyte risk parameter set.

[0086] It should be noted that the construction of marginal osteophyte risk parameters refers to screening candidate abnormal arc segments in the local outward protrusion enhancement response set output in step S30, and extracting multiple parameters that reflect osteophyte risk from the screened candidate abnormal arc segments. Since knee anteroposterior X-ray images may contain patellar projection, high-contrast soft tissue edges, fibular head overlap, local bone edge overlap, and random noise, these interferences can all form pseudo-responses in the local outward protrusion enhancement response set. Therefore, this step can construct artifact constraint regions such as the patellar adjacent overlap region, the high-contrast soft tissue edge region, and the bone edge overlap region based on the continuous cortical bone contour set output in step S20, and introduce artifact scoring and dual-threshold screening rules during the candidate abnormal arc segment screening process. For candidate abnormal arc segments that pass the screening, further extraction of outward protrusion amplitude parameters, outward protrusion length parameters, contour deviation parameters, grayscale change parameters, artifact scoring parameters, and image quality parameters forms the marginal osteophyte risk parameter set.

[0087] Understandably, step S40 is used to further convert the "local enhancement response" obtained in step S30 into "risk parameters that can be used for judgment." Local bulging enhancement response can indicate a possible abnormality at a certain location, but it cannot be directly equated to a positive osteophyte, because the enhancement response may originate from real osteophytes, or from artifacts, noise, or overlapping structures. This step, by adding an artifact constraint region and a dual-threshold screening rule, ensures that candidate abnormal arc segments not only have sufficient enhancement response but also meet certain requirements in terms of cortical continuity, anatomical location rationality, and image quality. This maintains the recall capability of early weak signs while preventing a large number of false abnormal areas from entering the final judgment step.

[0088] It should be understood that, compared to the traditional method of directly setting a fixed threshold on enhanced images to extract suspected osteophytes, the improvement in this step lies in changing the candidate screening from "single-condition judgment of response intensity" to "joint judgment of response intensity, artifact influence, and cortical continuity." Traditional fixed-threshold screening is easily affected by local image brightness and projection overlap, especially when the fibular head overlaps with the lateral tibial plateau, the patellar edge projection is close to the femoral condyle edge, or the soft tissue boundary is relatively clear, which may result in strong pseudo-responses. This step establishes an artifact constraint region for the common interferences in the above-mentioned outpatient anteroposterior X-ray images, and adopts a stricter dual-threshold retention rule when the artifact score is high, that is, the candidate abnormal arc segment must not only meet the local protrusion enhancement response threshold, but also meet the cortical context continuity threshold; when the artifact score is low, a relatively sensitive response threshold can be used to reduce the risk of over-removal of early micro-osteophytes.

[0089] For example, in an image, there is a projection overlap of the fibular head near the edge of the lateral tibial plateau. Step S30 obtains a high local outward protrusion enhancement response in this area. If only the magnitude of the enhancement response is considered, this area might be mistaken for an early osteophyte. This step constructs a bone edge overlap artifact constraint region based on the location of the fibular head and the position of the lateral tibial edge, and assigns a high artifact score to this candidate arc segment. Since this candidate arc segment is in a high artifact score region, it needs to simultaneously meet the local outward protrusion enhancement response threshold and the cortical continuity threshold; if the arc segment is inconsistent with the continuous direction of the true tibial plateau cortex, it is excluded. As another example, another candidate arc segment is located at the edge of the medial tibial plateau, with a low artifact score, moderate local outward protrusion response but stable outward offset and grayscale gradient change. This candidate arc segment can be retained, and the maximum outward protrusion amplitude, outward protrusion length, contour deviation degree, grayscale change degree, and image quality parameters are extracted to form a marginal osteophyte risk parameter set.

[0090] Step S50: Based on the marginal osteophyte risk parameter set, a cost-sensitive classification method is used to determine the enhancement of marginal osteophytes, and the determination result of the enhancement of marginal osteophytes is output.

[0091] It should be noted that the cost-sensitive classification method refers to assigning a higher cost weight to early, minor osteophyte omissions than to ordinary misjudgments during the determination of marginal osteophytes, making the determination process focus more on weak positive signs in the transition stage from KL level 0 to 1. Specifically, based on the marginal osteophyte risk parameter set output in step S40, a determination feature vector for candidate abnormal arc segments can be constructed. This determination feature vector includes at least the outward protrusion amplitude, outward protrusion length, contour deviation degree, grayscale change degree, artifact score, image quality parameters, and pose deviation parameters. Then, the cost-sensitive classification method is used to generate the positive probability of marginal osteophytes. Next, an adaptive determination threshold is generated based on image noise parameters, pose deviation parameters, and image quality parameters. The positive probability is compared with the adaptive determination threshold to obtain the positive determination result of marginal osteophytes. Finally, a structured determination result containing candidate location, outward protrusion degree, contour abnormality degree, and determination confidence is generated.

[0092] It is understandable that step S50 outputs not just an overall conclusion of "present" or "absent," but rather a verifiable, structured result for each candidate abnormal arc segment. For clinical image interpretation, doctors need not only to know whether the system has determined the presence of early marginal osteophytes, but also the specific location corresponding to that determination and its basis. For example, the determination result might include information such as "suspected early osteophyte on the medial tibial plateau edge," "maximum outward protrusion amplitude is several pixels," "local curvature abnormality is higher than the surrounding area," and "determination confidence level is several." By outputting structured determination results, doctors can return to the original X-ray image to verify the corresponding bone edge location, thereby improving the reliability of the auxiliary determination results.

[0093] It should be understood that the improvement of this step compared to traditional fixed threshold judgment or ordinary binary classification methods lies in the simultaneous introduction of cost-sensitive classification and adaptive judgment threshold. Traditional fixed thresholds often apply the same judgment standard to all images. While this may be able to detect obvious osteophytes when the image quality is good, it is prone to mismatch in images commonly found in primary care clinics, such as images with high noise, blurred bone edges, and slight pose deviations: a threshold that is too high will miss early micro-osteophytes, while a threshold that is too low will increase artifact misjudgments. Ordinary binary classification methods are easily affected by the distribution of training samples. Since the number of early micro-osteophyte samples is usually less than that of normal samples and obvious degenerative samples, the model may tend to judge them as negative. This step increases the weight of weak positive samples in training and judgment through cost-sensitive classification, and dynamically adjusts the judgment threshold through noise parameters, pose deviation parameters, and image quality parameters. This makes the judgment standard under different imaging conditions more closely match the actual quality of the current image, thus balancing early detection capability and misjudgment control capability.

[0094] For example, a candidate abnormal arc segment is located at the edge of the medial tibial plateau. The risk parameters extracted in step S40 show that the maximum outward protrusion amplitude is 2.6 pixels, the outward protrusion length is 18 pixels, the degree of local curvature abnormality is 1.5 times higher than the neighborhood average, the artifact score is low, and the image quality parameters are at a medium-to-high level. The cost-sensitive classification method generates a positive probability of 0.63 for marginal osteophytes; the adaptive judgment threshold generated based on the noise parameters, pose deviation parameters, and image quality parameters of the image is 0.58. Therefore, the candidate abnormal arc segment is judged as a positive region for marginal osteophytes and marked as "suspected early marginal osteophyte at the edge of the medial tibial plateau" in the structured judgment results. For example, another candidate abnormal arc segment is located in the area adjacent to the patellar projection, with a positive probability of 0.60. However, the image quality in this area is low and the artifact score is high. If the adaptive judgment threshold is increased to 0.66, this candidate abnormal arc segment will not be directly judged as positive, but can be output as a low-confidence suspicious area or an exclusion area, thereby reducing misjudgments caused by patellar projection or soft tissue overlap. Through the above processing, this embodiment can achieve more stable and interpretable enhancement judgment of early small marginal osteophytes under routine outpatient X-ray imaging conditions.

[0095] Example 2: Furthermore, the knee osteoarthritis local sign enhancement judgment system for early marginal osteophyte omission provided by the present invention employs the knee osteoarthritis local sign enhancement judgment method for early marginal osteophyte omission in the above embodiments, and can solve the technical problem of enhancing judgment of local signs of knee osteoarthritis for early marginal osteophyte omission. Compared with the prior art, the beneficial effects of the knee osteoarthritis local sign enhancement judgment system for early marginal osteophyte omission provided by the present invention are the same as the beneficial effects of the knee osteoarthritis local sign enhancement judgment method for early marginal osteophyte omission provided in the above embodiments, and other technical features of the knee osteoarthritis local sign enhancement judgment system for early marginal osteophyte omission are the same as those disclosed in the above embodiments, and will not be repeated here.

[0096] Example 3: This invention provides a device for enhancing the determination of local signs of knee osteoarthritis in cases of missed early marginal osteophyte diagnosis. The device includes: at least one processor; and a memory communicatively connected to the at least one processor. The memory stores instructions executable by the at least one processor, which are executed to enable the at least one processor to perform the method for enhancing the determination of local signs of knee osteoarthritis in cases of missed early marginal osteophyte diagnosis described in Example 1. The device for enhancing the determination of local signs of knee osteoarthritis in cases of missed early marginal osteophyte diagnosis in this invention can include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), and in-vehicle terminals (e.g., in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. This device for enhancing the determination of local signs of knee osteoarthritis in cases of missed early marginal osteophyte diagnosis is merely an example and should not be construed as limiting the functionality or scope of the invention. A device for enhancing the assessment of local signs of knee osteoarthritis to address the early omission of marginal osteophytes may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) that can perform various appropriate actions and processes based on a program stored in read-only memory or a program loaded from a storage device into random access memory. The random access memory also stores various programs and data required for the operation of the device. The processing unit, read-only memory, and random access memory are interconnected via a bus. An I / O interface is also connected to the bus. Typically, the following systems can be connected to the I / O interface: input devices including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices including, for example, magnetic tapes, hard disks, etc.; and communication devices. The communication device allows the device to exchange data wirelessly or via wired communication with other devices. While the figures illustrate enhanced assessment devices for local signs of knee osteoarthritis with various systems designed to detect early marginal osteophyte malformations, it should be understood that implementation of or possession of all the systems shown is not required. More or fewer systems may be implemented alternatively.

[0097] Example 4: This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method for determining the enhancement of local signs in knee osteoarthritis aimed at early detection of marginal osteophytes. The computer program product provided by this invention can solve the technical problem of determining the enhancement of local signs in knee osteoarthritis aimed at early detection of marginal osteophytes. Compared with the prior art, the beneficial effects of the computer program product provided by this invention are the same as those of the method for determining the enhancement of local signs in knee osteoarthritis aimed at early detection of marginal osteophytes provided in the above embodiments, and will not be repeated here.

[0098] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device, or installed from a read-only memory. When the computer program is executed by a processing device, it performs the functions defined in the methods of the embodiments disclosed in this invention.

[0099] It should be understood that the various parts disclosed in this invention can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0100] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for enhancing the assessment of local signs in knee osteoarthritis to address the early omission of marginal osteophytes, characterized by the following: include: Step S10: Obtain a routine outpatient knee joint anteroposterior X-ray image, perform local standardization processing based on the routine outpatient knee joint anteroposterior X-ray image, and output a local standardization parameter set; Step S20: Based on the local standardized parameter set, the target bone edge is extracted using the bone tissue segmentation and boundary continuity optimization method, and a continuous bone cortex contour set is output; Step S30: Based on the continuous cortical bone contour set, a reference contour fitting and differential enhancement method is used to perform local protrusion preprocessing, and output the local protrusion enhancement response set; Step S40: Based on the local protrusion enhancement response set, construct the marginal osteophyte risk parameters and output the marginal osteophyte risk parameter set; Step S50: Based on the marginal osteophyte risk parameter set, a cost-sensitive classification method is used to determine the enhancement of marginal osteophytes, and the determination result of the enhancement of marginal osteophytes is output.

2. The method for enhancing the assessment of local signs of knee osteoarthritis as described in claim 1, characterized in that, Step S10, which involves performing local standardization processing on the routine outpatient knee joint anteroposterior X-ray image and outputting a local standardization parameter set, specifically includes: Step S101: Perform percentile grayscale normalization on the routine outpatient knee joint anteroposterior X-ray image to obtain a standard grayscale image; Step S102: Based on the standard grayscale image, extract the femoral distal positioning region, tibial plateau positioning region, tibial spine positioning region and fibular head positioning region to construct a local anatomical coordinate system for the knee joint; Step S103: Generate local scale parameters, posture deviation parameters, and mode selection parameters based on the local anatomical coordinate system of the knee joint to form the local standardized parameter set; Wherein, the percentile grayscale normalization process satisfies: ； in, Represents the original grayscale value. This represents the grayscale value at the 100th percentile. This represents the 99th percentile gray value. This represents the normalized grayscale value. This indicates a bounded truncation function, used to restrict input values ​​to a preset range.

3. The method for enhancing the assessment of local signs in knee osteoarthritis as described in claim 1, characterized in that: Step S20, which involves extracting the target bone edge using a bone tissue segmentation and boundary continuity optimization method based on the local standardized parameter set, and outputting a continuous bone cortical contour set, specifically includes: Step S201: Based on the local standardized parameter set, a convolutional neural network segmentation method is used to segment the bone body of the distal femur region and the proximal tibia region to obtain the bone body segmentation result; Step S202: Based on the bone segmentation results, predict the probability of the bone cortex boundary and use an active contour model to optimize the boundary continuity to obtain the initial bone cortex contour. Step S203: Based on the initial cortical bone contour, perform continuity checks on adjacent arc segments and correct abnormal fractures to obtain the continuous cortical bone contour set.

4. The method for enhancing the assessment of local signs of knee osteoarthritis as described in claim 1, characterized in that, Step S30, which involves performing local protrusion preprocessing based on the continuous cortical bone contour set using a reference contour fitting and differential enhancement method, and outputting a local protrusion enhancement response set, specifically includes: Step S301: Establish arc length parameters for each bone edge arc segment in the continuous bone cortex contour set, and construct a reference contour using a cubic spline fitting method; Step S302: Obtain the actual bone edge contour, and generate local difference feature parameters based on the normal offset, local curvature and gray-level gradient difference between the reference contour and the actual bone edge contour; Step S303: Based on the local differential feature parameters, perform outward protrusion enhancement and physiological gradual contour suppression to obtain the local outward protrusion enhancement response set.

5. The method for enhancing the assessment of local signs of knee osteoarthritis as described in claim 4, characterized in that, The normal offset of the actual bone edge contour relative to the reference contour satisfies: ； in, Indicates the reference profile in the arc length parameter The coordinates of the point, This indicates the actual bone edge contour in the arc length parameter. The coordinates of the points are all represented using two-dimensional plane coordinates. Indicates the reference profile in the arc length parameter The outward normal unit vector at a given location satisfies: ； in, , These represent the reference profile in terms of arc length parameter. The first derivative at that point; This represents the transpose of the outward normal unit vector; This represents the normal offset of the actual bone edge contour relative to the reference contour, when... When it indicates an outward protrusion, This indicates that no outward protrusion has occurred; The local curvature satisfies: ； in, Represents arc length parameter Local curvature at that point; , These represent the actual bone edge contour in terms of arc length parameters. The second derivative at point .

6. The method for enhancing the assessment of local signs of knee osteoarthritis as described in claim 1, characterized in that, Step S40, which involves constructing marginal osteophyte risk parameters based on the local protrusion enhancement response set and outputting the marginal osteophyte risk parameter set, specifically includes: Step S401: Construct artifact constraint regions for the patellar adjacent overlapping region, soft tissue high-contrast edge region, and bone edge overlapping region based on the continuous bone cortex contour set; Step S402: Based on the artifact constraint region, perform artifact scoring and dual threshold screening on the candidate abnormal arc segments in the local outward enhancement response set, and output the candidate abnormal arc segments; Step S403: Extract the outward protrusion amplitude parameter, outward protrusion length parameter, contour deviation parameter, grayscale change parameter, artifact scoring parameter and image quality parameter from the candidate abnormal arc segment to form the marginal osteophyte risk parameter set.

7. The method for enhancing the assessment of local signs in knee osteoarthritis as described in claim 6, characterized in that, Step S50, which involves using a cost-sensitive classification method to determine marginal osteophyte enhancement based on the marginal osteophyte risk parameter set and outputting the determination result, specifically includes: Step S501: Construct a judgment feature vector for candidate abnormal arc segments based on the marginal osteophyte risk parameter set, and generate the positive probability of marginal osteophytes; Step S502: Introduce an adaptive judgment threshold and compare the positive probability of marginal osteophytes with the adaptive judgment threshold to obtain the positive judgment result of marginal osteophytes; Step S503: Generate a structured judgment result based on the positive determination result of marginal osteophytes, including candidate location, degree of protrusion, degree of contour abnormality, and judgment confidence. Wherein, the adaptive determination threshold satisfies: ； in, This represents the adaptive threshold for determining the probability of marginal osteocheilitis. This represents the base threshold, which is a fixed threshold obtained through statistical analysis of training samples. , and These represent weighting coefficients, which are used to adjust the degree of influence of image noise, pose deviation, and image quality on the decision threshold.

8. A system for enhancing the assessment of local signs in knee osteoarthritis for early detection of marginal osteophytes, applied to any one of the methods for enhancing the assessment of local signs in knee osteoarthritis for early detection of marginal osteophytes as described in claims 1 to 7, characterized in that, The enhanced local sign assessment system for knee osteoarthritis aimed at early detection of marginal osteophytes includes: The standardized parameter construction module is used to acquire routine outpatient knee joint anteroposterior X-ray images, perform local standardization processing based on the routine outpatient knee joint anteroposterior X-ray images, and output a local standardized parameter set. The bone edge extraction module is used to perform target bone edge extraction based on the local standardized parameter set using bone tissue segmentation and boundary continuity optimization methods, and output a continuous bone cortex contour set; The external protrusion enhancement module is used to perform local external protrusion preprocessing based on the continuous cortical bone contour set using a reference contour fitting and differential enhancement method, and output a local external protrusion enhancement response set; The risk construction module is used to construct risk parameters for marginal osteophytes based on the local protrusion enhancement response set and output the risk parameter set for marginal osteophytes. The enhancement determination module is used to perform edge osteophyte enhancement determination based on the edge osteophyte risk parameter set using a cost-sensitive classification method, and output the edge osteophyte enhancement determination result.

9. A device for enhancing the assessment of local signs in knee osteoarthritis to detect early-stage marginal osteophytes, characterized in that: The device for enhancing the local signs of knee osteoarthritis for early detection of marginal osteophytes includes: a memory, a processor, and a program for enhancing the local signs of knee osteoarthritis for early detection of marginal osteophytes stored in the memory and executable on the processor. When the program for enhancing the local signs of knee osteoarthritis for early detection of marginal osteophytes is executed by the processor, it implements the method for enhancing the local signs of knee osteoarthritis for early detection of marginal osteophytes as described in any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes a local sign enhancement determination program for knee osteoarthritis aimed at early omission of marginal osteophytes. When the local sign enhancement determination program for knee osteoarthritis aimed at early omission of marginal osteophytes is executed by the processor, it implements the local sign enhancement determination method for knee osteoarthritis aimed at early omission of marginal osteophytes as described in any one of claims 1 to 7.

Images