Intelligent articular cartilage MRI image enhancement and contrast optimization system

Abstract

The invention discloses an articular cartilage MRI (Magnetic Resonance Imaging) image intelligent enhancement and contrast optimization system, which belongs to the technical field of medical image processing and comprises a cartilage feature extraction module, a self-adaptive enhancement module, a contrast optimization module and a quality evaluation feedback module, the cartilage feature extraction module realizes cartilage region classification and structural analysis based on T2 relaxation time distribution features and multi-scale structure tensor features; the adaptive enhancement module performs differential enhancement according to the region type and the structural feature; the contrast optimization module realizes contrast dynamic optimization through nonlinear mapping and edge gradient adaptive adjustment; according to the articular cartilage MRI image display system and method, the display quality of the articular cartilage MRI image is remarkably improved, the cartilage display definition is improved by 73% or above compared with that of an original image, and a high-quality imaging basis is provided for accurate diagnosis of chondropathy.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of medical image processing and computer-aided diagnosis, and relates to a magnetic resonance imaging (MRI) image enhancement technique, in particular to an intelligent enhancement and contrast optimization system for joint cartilage MRI images. BACKGROUND

[0002] Joint cartilage is a special connective tissue covering the surface of joints, which has important physiological functions of bearing load, reducing friction and absorbing shock. Early lesions of joint cartilage are of great significance for the prevention and treatment of osteoarthritis. Magnetic resonance imaging (MRI) has become an important imaging examination method for diagnosing joint cartilage lesions due to its excellent soft tissue contrast and non-invasive characteristics. However, in conventional MRI imaging, the signal contrast of joint cartilage and surrounding tissues such as subchondral bone, joint effusion and ligaments is low, the boundary of cartilage is not clearly displayed, and the microstructure and early lesions in the cartilage are difficult to accurately identify, which brings significant challenges to clinical diagnosis and scientific research analysis.

[0003] Chinese patent CN103456004A discloses a joint cartilage segmentation method based on image sheet structure enhancement and a system thereof. The technology extracts the cartilage detection area of the magnetic resonance image, calculates the sheet structure feature function in the cartilage detection area, performs gray scale enhancement on the sheet structure in the cartilage detection area, and finally performs threshold connection to extract the cartilage. The method mainly has the following deficiencies: first, the technology uses Hessian matrix eigenvalue to detect sheet structure, only based on the second derivative information of the image to judge the structure, and does not fully utilize the physiological characteristic parameters such as T2 relaxation time of cartilage tissue, resulting in limited ability to distinguish different pathological state cartilage tissues; second, the gray scale enhancement method of the technology is a unified processing based on sheet structure feature function, lacking an adaptive adjustment mechanism for different regional characteristics of cartilage tissue, and being unable to perform differential enhancement according to the local features of cartilage tissue; third, the technology does not establish a closed-loop feedback optimization mechanism, and once the enhancement parameters are determined, they cannot be dynamically adjusted according to the enhancement effect, lacking adaptive ability when processing MRI images of different qualities, resulting in insufficient stability and robustness of the enhancement effect; fourth, the technology mainly focuses on cartilage segmentation rather than contrast optimization, and has limited ability to dynamically adjust the contrast between cartilage and surrounding tissues, making it difficult to meet the high-quality requirements of cartilage detail display in clinical practice.

[0004] Therefore, how to fully utilize the physiological characteristics such as T2 relaxation time of cartilage tissue, combine multi-scale structure information, establish an adaptive enhancement and contrast optimization mechanism, and realize dynamic adjustment of parameters through closed-loop feedback is a key technical problem to be solved in the current joint cartilage MRI image enhancement technique. SUMMARY

[0005] In view of the above-mentioned deficiencies existing in the prior art, the purpose of the present application is to provide a joint cartilage MRI image intelligent enhancement and contrast optimization system, through the deep coupling and cooperation of a cartilage feature extraction module, an adaptive enhancement module, a contrast optimization module and a quality evaluation feedback module, based on the T2 relaxation time characteristics and multi-scale structure tensor characteristics of cartilage tissue, the regional adaptive enhancement and contrast dynamic optimization of cartilage tissue are realized, and through the closed-loop feedback mechanism, the adaptive adjustment of parameters is realized, and the display quality and diagnostic value of joint cartilage MRI images are significantly improved.

[0006] In order to achieve the above-mentioned purpose, the technical scheme adopted by the present application is as follows:

[0007] The joint cartilage MRI image intelligent enhancement and contrast optimization system comprises a cartilage feature extraction module, an adaptive enhancement module, a contrast optimization module and a quality evaluation feedback module.

[0008] The cartilage feature extraction module calculates the T2 relaxation time distribution characteristics and multi-scale structure tensor characteristics of cartilage tissue based on the obtained joint cartilage MRI image data. The cartilage feature extraction module extracts the T2 relaxation time distribution information of cartilage tissue from the multi-echo spin echo sequence, utilizes the sensitivity of T2 relaxation time to the water content and collagen fiber arrangement state of cartilage tissue, and realizes the accurate differentiation of normal cartilage, early degenerative cartilage and pathological cartilage. The cartilage feature extraction module simultaneously calculates the multi-scale structure tensor characteristics, analyzes the local structure information of cartilage tissue at different spatial scales, obtains the edge, texture and anisotropy characteristics of cartilage tissue, and provides multi-dimensional feature support for subsequent adaptive enhancement.

[0009] The adaptive enhancement module is deeply coupled with the cartilage feature extraction module, generates differentiated enhancement weight parameters for different regions according to the cartilage tissue region type determined by the T2 relaxation time distribution characteristics. The adaptive enhancement module sets different enhancement weights for normal cartilage regions, early degenerative regions and pathological cartilage regions, so that the pathological cartilage region obtains higher enhancement intensity, and the cartilage lesion characteristics are highlighted. The adaptive enhancement module judges the local structure type based on the eigenvalue difference of the multi-scale structure tensor characteristics, determines the edge structure region when the eigenvalue difference is greater than a preset edge determination threshold and applies an edge preserving enhancement algorithm, and determines the homogeneous structure region when the eigenvalue difference is not greater than the preset edge determination threshold and applies a smoothing enhancement algorithm, so as to ensure that the contrast of the enhanced cartilage is maintained while the clarity of the cartilage edge and the integrity of the cartilage internal structure are maintained.

[0010] The contrast optimization module cooperates with the adaptive enhancement module, receives the enhanced image data output by the adaptive enhancement module, and dynamically adjusts the contrast mapping curve of the cartilage region according to the signal difference characteristics of the cartilage tissue and the surrounding tissue in the enhanced image data. The contrast optimization module determines the contrast stretching interval by counting the signal intensity distribution range of the cartilage tissue, subchondral bone and joint effusion in the enhanced image data, and increases the signal intensity difference between the cartilage tissue and the surrounding tissue through nonlinear mapping. The contrast optimization module adaptively adjusts the slope of the contrast mapping curve according to the gradient information at the edge of the cartilage, increases the contrast enhancement intensity in the area with weak edge gradient, and reduces the enhancement intensity in the area with strong edge gradient, so as to avoid edge artifacts caused by over-enhancement and realize accurate display of the cartilage boundary.

[0011] The quality evaluation feedback module and the contrast optimization module form a closed-loop feedback system, and the quality of the enhancement effect is evaluated by calculating the cartilage display clarity index and the edge retention index of the enhanced image. The quality evaluation feedback module calculates the contrast noise ratio of the cartilage tissue region to represent the cartilage display clarity, and extracts the gradient direction consistency of the cartilage edge before and after enhancement to represent the edge retention. When the cartilage display clarity index is lower than a first preset threshold or the edge retention index is lower than a second preset threshold, the quality evaluation feedback module generates a parameter adjustment instruction and feeds back to the adaptive enhancement module and the contrast optimization module, so as to dynamically adjust the enhancement weight parameter and the contrast mapping curve through the iterative optimization mechanism, and realize the best balance between the enhancement effect and the edge retention.

[0012] Through the deep coupling and closed-loop cooperation of the above four modules, the present application realizes intelligent enhancement and contrast dynamic optimization of cartilage tissue, and significantly improves the display quality of joint cartilage MRI images.

[0013] Compared with the prior art, the present application has the following beneficial effects:

[0014] Firstly, the present application calculates the T2 relaxation time distribution characteristics of the cartilage tissue through the cartilage feature extraction module, and uses the sensitivity of T2 relaxation time to the physiological state of the cartilage tissue to accurately distinguish normal cartilage, early degenerative cartilage and pathological cartilage, which provides a region classification basis based on physiological characteristics for adaptive enhancement. Compared with the existing technology which only detects sheet-like structures based on image second-order derivatives, the present application can more accurately identify cartilage tissue in different pathological states, and improve the pertinence and effectiveness of the enhancement processing.

[0015] Secondly, the application realizes the regional adaptive enhancement of cartilage tissue by the adaptive enhancement module according to the regional type and local structural features of cartilage tissue, sets different enhancement weight parameters for different regions, and applies different enhancement algorithms for different structural types, so that the application can highlight the pathological cartilage region while maintaining the natural contrast of the normal cartilage region and ensuring the clarity of the cartilage edge, and the fine degree and clinical application value of the enhancement processing are significantly improved.

[0016] Thirdly, the application realizes the significant improvement of the contrast of cartilage and subchondral bone and joint effusion by the contrast optimization module according to the signal difference features of cartilage tissue and surrounding tissue, dynamically adjusts the contrast mapping curve, and adaptively adjusts the slope of the mapping curve according to the edge gradient information, so that the application can better meet the high-quality requirement of the clinical cartilage detail display and provide a better image basis for the accurate diagnosis of cartilage lesions.

[0017] Fourthly, the application realizes the adaptive optimization of the enhancement processing by the quality evaluation feedback module to establish a closed-loop feedback optimization mechanism, dynamically adjusts the enhancement parameters and the contrast mapping curve according to the quality indicators of the enhancement effect, so that the application can automatically adjust the parameters according to different quality MRI images and different enhancement effects, significantly improves the robustness and adaptability of the system, and ensures that stable high-quality enhancement effect can be obtained under various imaging conditions.

[0018] In summary, the application realizes the intelligent enhancement and contrast optimization of joint cartilage MRI images through the deep integration of cartilage region classification based on T2 relaxation time characteristics, adaptive enhancement of multi-scale structural features, dynamic contrast optimization and closed-loop feedback mechanism, the cartilage display clarity of the processed MRI image is improved by more than 73% compared with the original image, the differentiation of subchondral bone and joint effusion is significantly improved, high-quality image support is provided for the accurate diagnosis and minimally invasive treatment planning of joint cartilage lesions, and the application has important clinical application value and broad application prospect. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 is the overall architecture schematic diagram of the joint cartilage MRI image intelligent enhancement and contrast optimization system of the application.

[0020] Figure 2 is the functional structure schematic diagram of the cartilage feature extraction module of the application.

[0021] Figure 3 is the processing flow schematic diagram of the adaptive enhancement module of the application.

[0022] Figure 4 This is a schematic diagram of the mapping curve adjustment of the contrast optimization module of the present invention.

[0023] Figure 5 This is a schematic diagram of the closed-loop feedback mechanism of the quality assessment feedback module of the present invention. Detailed Implementation

[0024] Please refer to the attached document. Figures 1-5 The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0025] Reference Figure 1 The intelligent enhancement and contrast optimization system for articular cartilage MRI images provided by the present invention includes a cartilage feature extraction module 1, an adaptive enhancement module 2, a contrast optimization module 3, and a quality assessment feedback module 4.

[0026] Reference Figure 2 The cartilage feature extraction module 1 receives MRI image data of articular cartilage as input. This MRI image data is preferably acquired using a multi-echo spin-echo (MESE) sequence to obtain signal attenuation information of the cartilage tissue at different echo times. In a preferred embodiment of the invention, the MRI imaging parameters are set as follows: field strength 3.0T, repetition time 2500ms, echo times of 10ms, 20ms, 30ms, 40ms, 50ms, 60ms, and 70ms, slice thickness 3mm, interslice spacing 0.5mm, field of view 160mm×160mm, and matrix 512×512.

[0027] The cartilage feature extraction module 1 first calculates the T2 relaxation time distribution characteristics of cartilage tissue. T2 relaxation time is an important parameter describing the decay rate of the transverse magnetization vector of an MRI signal, and it is highly sensitive to the water content, collagen fiber arrangement, and matrix integrity of cartilage tissue. The cartilage feature extraction module 1 extracts the signal intensity of each pixel at different echo times from the multi-echo signal sequence, denoted as... ,in Indicates the echo sequence number. The total number of echoes. Based on the exponential model of signal attenuation, the T2 relaxation time is determined as follows:

[0028] ,

[0029] in, Echo time The corresponding signal strength, The initial signal strength, The T2 relaxation time is the unknown value. The above equation is solved using a nonlinear least squares fitting method to obtain the T2 relaxation time value for each pixel in the cartilage tissue. In a preferred embodiment, the T2 relaxation time ranges from 30 to 45 ms for normal cartilage tissue, from 45 to 60 ms for early degenerative cartilage, and from greater than 60 ms for pathological cartilage.

[0030] Based on the statistical distribution characteristics of T2 relaxation time values, the cartilage feature extraction module 1 divides cartilage tissue into normal cartilage regions, early degeneration regions, and pathological cartilage regions. Specifically, the division method involves statistically analyzing the T2 relaxation time values ​​of the cartilage tissue regions and calculating the mean T2 relaxation time. and standard deviation When the T2 relaxation time of a pixel satisfy When, it is determined to be a normal cartilage area; when When, it is determined to be an early degradation region; when When the T2 relaxation time is reached, the area is identified as pathological cartilage. This region segmentation method based on T2 relaxation time characteristics can fully utilize the physiological characteristics of cartilage tissue to achieve accurate identification of cartilage tissue in different pathological states.

[0031] The cartilage feature extraction module 1 also calculates the multi-scale structural tensor features of cartilage tissue. Structural tensors are important tools for describing the local structural characteristics of images, characterizing the principal orientation, secondary orientation, and degree of anisotropy of local structures. The cartilage feature extraction module 1 performs multi-scale Gaussian smoothing on the MRI image data, at various scales... Calculate pixel points Image gradient at Construct the gradient covariance matrix, i.e., the structure tensor:

[0032] ,

[0033] in, variance Gaussian kernel, This represents the convolution operation. For scale The image gradient vector below, Indicates transpose. Structure tensor It is a 2×2 symmetric matrix, and its eigenvalues ​​are... and (agreement) The eigenvectors represent the principal and secondary directional intensities of the local structure, respectively, and the eigenvectors represent the directional information of the principal and secondary directions.

[0034] In a preferred embodiment of the present invention, three different scale parameters are used for multi-scale analysis: , For each scale, the eigenvalues ​​of the structure tensor are calculated. and eigenvalue differences Characterizes the degree of anisotropy of local structure. When eigenvalue differences Greater than the preset edge detection threshold Time (preferred) ), which is determined to be a marginal structure region; when the eigenvalue difference, Not greater than the preset edge detection threshold When the structure is homogeneous, it is identified as a homogeneous structural region. Multi-scale structural tensor features can capture local structural information of cartilage tissue at different spatial scales, providing rich structural feature support for adaptive enhancement.

[0035] Reference Figure 3 The adaptive enhancement module 2 is deeply coupled with the cartilage feature extraction module 1. It receives the T2 relaxation time distribution features and multi-scale structural tensor features output by the cartilage feature extraction module 1, and performs regional adaptive enhancement processing of cartilage tissue based on these features.

[0036] The adaptive enhancement module 2 first determines the cartilage tissue region type based on the T2 relaxation time distribution characteristics and generates differentiated enhancement weight parameters. In the innovative design of this invention, different enhancement weights are set for cartilage tissue in different pathological states, specifically:

[0037] ,

[0038] in, The enhancement weight parameters for normal cartilage regions, early degeneration regions, and pathological cartilage regions are respectively, satisfying the following relationship. In a preferred embodiment, , , This differentiated enhancement weight setting allows pathological cartilage areas to receive higher enhancement intensity, highlighting the characteristics of cartilage lesions, while maintaining the natural contrast of normal cartilage areas and avoiding image distortion caused by over-enhancement.

[0039] Adaptive enhancement module 2 further determines the local structure type and selects the corresponding enhancement algorithm based on multi-scale structural tensor features. For edge structure regions, adaptive enhancement module 2 applies an edge-preserving enhancement algorithm, which is implemented through the following innovative formula:

[0040] ,

[0041] in, The pixel intensity of the original image. To maintain the enhanced pixel intensity at the edges, Edge enhancement coefficient (preferred) ), and These are the eigenvalues ​​of the structure tensor. The gradient magnitude at the pixel. Gradient control parameters (preferred) This formula is obtained through... The term characterizes edge strength, through an exponential decay term. By controlling the enhancement amplitude, the enhancement intensity is reduced when the edge gradient is large to avoid over-enhancing the edges, and the enhancement intensity is increased when the edge gradient is small to improve the edge contrast, thereby achieving accurate edge preservation and appropriate enhancement.

[0042] For homogeneous regions, adaptive enhancement module 2 applies a smoothing enhancement algorithm, achieving contrast enhancement through a weighted average of local neighborhoods while suppressing noise. The smoothing enhancement algorithm is implemented by calculating pixel values... In the neighborhood The weighted average value within the range, with weights determined based on spatial distance and gray-level similarity, employs a bilateral filtering weight design. The enhanced pixel intensity is:

[0043] ,

[0044] in, To smooth the enhanced pixel intensity, For smoothing enhancement coefficient (preferred) ), It is a neighborhood-weighted average. This algorithm uses... The local contrast enhancement is calculated by... and By controlling the enhancement intensity, the contrast of homogeneous areas is improved while maintaining the smoothness and consistency within the areas.

[0045] By combining the edge-preserving enhancement algorithm and the smoothing enhancement algorithm, the adaptive enhancement module 2 can perform differentiated enhancement based on the local structural features of cartilage tissue, which not only ensures the clarity of the cartilage edge but also improves the contrast of the internal structure of the cartilage, thus achieving fine control of the enhancement effect.

[0046] Reference Figure 4 The contrast optimization module 3 works in conjunction with the adaptive enhancement module 2. It receives the enhanced image data output by the adaptive enhancement module 2 and dynamically adjusts the contrast mapping curve of the cartilage region according to the signal difference characteristics between the cartilage tissue and the surrounding tissue in the enhanced image data, thereby further improving the display contrast of the cartilage tissue.

[0047] The contrast optimization module 3 first statistically analyzes the signal intensity distribution ranges of cartilage tissue, subchondral bone, and joint effusion in the enhanced image data. By labeling and statistically analyzing the signal intensity of the cartilage tissue, subchondral bone, and joint effusion regions in the enhanced image data, the mean and standard deviation of the signal intensity for each tissue are obtained: the mean signal intensity of cartilage tissue is denoted as... The standard deviation is denoted as The mean signal intensity of subchondral bone is denoted as The standard deviation is denoted as The mean signal intensity of joint effusion is denoted as . The standard deviation is denoted as .

[0048] Contrast optimization module 3 determines the contrast stretching range based on the aforementioned signal intensity distribution range. ,in:

[0049] ,

[0050] ,

[0051] Within the contrast stretching range, contrast optimization module 3 applies a nonlinear mapping function to transform the signal intensity of cartilage tissue, increasing the signal intensity difference between cartilage tissue and surrounding tissues. The innovative contrast mapping curve proposed in this invention adopts the form of a Sigmoid function:

[0052] ,

[0053] in, For adaptively enhanced pixel intensity, Pixel intensity optimized for contrast. The slope parameter of the contrast mapping curve. Mapping center point (preferred) The S-shaped curve characteristic of the Sigmoid function maps the signal intensity of cartilage tissue to the steep region of the curve, while the signal intensity of surrounding tissue is mapped to the flat region of the curve, thereby achieving a significant improvement in the contrast of cartilage tissue.

[0054] Contrast optimization module 3 further adaptively adjusts the slope parameter of the contrast mapping curve based on gradient information at the cartilage edge. For each pixel at the cartilage edge, calculate its gradient magnitude. The slope parameter is increased when the gradient magnitude is below the third preset threshold, and decreased when the gradient magnitude is above the third preset threshold. The slope parameter is dynamically adjusted according to the gradient magnitude.

[0055] ,

[0056] in, Basic slope parameter (preferred) ), Gradient adjustment coefficient (preferred) ), Edge gradient control parameters (preferred) When the magnitude of the marginal gradient is small, the exponential term is close to 1, and the slope parameter... Increasing the value of the gradient makes the contrast mapping curve steeper, thus enhancing contrast; when the edge gradient magnitude is large, the exponential term approaches 0, and the slope parameter... By reducing the contrast ratio, the contrast mapping curve becomes gentler, weakening the enhancement intensity and avoiding artifacts and distortion caused by excessive edge enhancement. This slope adaptive adjustment mechanism based on edge gradients ensures that cartilage boundaries are clearly displayed while avoiding the negative effects of over-enhancement.

[0057] Through the nonlinear transformation and slope adaptive adjustment of the contrast mapping curve, the contrast optimization module 3 significantly improves the contrast between cartilage tissue and subchondral bone and joint effusion, greatly improving the clarity of the cartilage boundary and providing a high-quality imaging basis for clinical diagnosis.

[0058] Reference Figure 5 The quality assessment feedback module 4 and the contrast optimization module 3 form a closed-loop feedback system. By calculating the quality index of the enhanced image, the enhancement effect is evaluated, and parameter adjustment instructions are generated based on the evaluation results. These instructions are then fed back to the adaptive enhancement module 2 and the contrast optimization module 3 to achieve dynamic optimization of the enhancement parameters.

[0059] The quality assessment feedback module 4 calculates two types of key quality indicators: cartilage display clarity and edge preservation.

[0060] Cartilage visualization sharpness is characterized by the contrast-to-noise ratio (CNR). The quality assessment feedback module 4 calculates the mean signal intensity of the cartilage tissue region in the enhanced image. and standard deviation and the mean signal intensity of the background region (joint effusion region). and standard deviation The contrast-to-noise ratio is defined as:

[0061] ,

[0062] A higher contrast-to-noise ratio indicates higher clarity of cartilage tissue relative to the background. In a preferred embodiment, a first preset threshold, i.e., a cartilage display clarity threshold, is set. When the calculated contrast-to-noise ratio If the cartilage is not clearly visible, the enhancement parameters need to be adjusted.

[0063] Edge preservation index is characterized by the consistency of gradient direction of cartilage edges before and after enhancement. Quality assessment feedback module 4 first extracts the gradient vectors of cartilage tissue edges from the original and enhanced images. and Calculate the cosine of the angle between the gradient vectors at the edge pixels:

[0064] ,

[0065] Edge gradient direction consistency is defined as the average of the cosine values ​​of the included angles of all edge pixels:

[0066] ,

[0067] in, This represents the set of pixels at the edge of the cartilage. This represents the total number of edge pixels. The closer the edge gradient direction consistency is to 1, the smaller the change in edge direction caused by the enhancement process, and the better the edge preservation. In a preferred embodiment, a second preset threshold, i.e., the edge preservation threshold, is set. When the calculated edge gradient directions are consistent If the edge preservation is insufficient, the enhancement parameters need to be adjusted.

[0068] The quality assessment feedback module 4 generates parameter adjustment instructions based on the assessment results of the quality indicators. The parameter adjustment strategy is as follows:

[0069] when Below the first preset threshold and Not lower than the second preset threshold When the cartilage display is deemed to have insufficient clarity but good edge retention, an adjustment instruction to increase the enhancement weight parameter is generated, specifically:

[0070] ,

[0071] Simultaneously increase the base slope parameter of the contrast mapping curve:

[0072] ,

[0073] in, (Preferred value) (Preferred value).

[0074] when Not lower than the first preset threshold and Below the second preset threshold When the cartilage is deemed to have sufficient clarity but insufficient edge preservation, an adjustment instruction is generated to reduce the enhancement weight parameter, specifically:

[0075] ,

[0076] Simultaneously reduce the base slope parameter of the contrast mapping curve:

[0077] ,

[0078] when Below the first preset threshold and Below the second preset threshold If the cartilage visualization clarity and edge preservation are deemed insufficient, a comprehensive adjustment instruction is generated, adjusting the edge enhancement coefficient accordingly. and gradient adjustment coefficient Achieving a balance between enhancement and edge preservation:

[0079] ,

[0080] The above parameter adjustment instructions are fed back to the adaptive enhancement module 2 and the contrast optimization module 3 through the iterative optimization mechanism, and the enhancement processing and contrast optimization are carried out again until the cartilage display clarity index is not lower than the first preset threshold and the edge preservation index is not lower than the second preset threshold, or the maximum number of iterations (preferably 5 times) is reached.

[0081] Through the closed-loop feedback optimization mechanism of the quality assessment feedback module 4, this invention achieves adaptive adjustment of enhancement parameters, which can automatically optimize parameters according to MRI images of different qualities and different enhancement effects, ensuring stable high-quality enhancement effects under various imaging conditions, and significantly improving the robustness and adaptability of the system.

[0082] The overall workflow of the intelligent enhancement and contrast optimization system for articular cartilage MRI images of this invention is as follows:

[0083] Step 1: The cartilage feature extraction module 1 receives MRI image data of articular cartilage, calculates the T2 relaxation time distribution characteristics and multi-scale structural tensor characteristics of cartilage tissue, divides the cartilage tissue into normal cartilage region, early degeneration region and pathological cartilage region according to the T2 relaxation time value, and determines the local structure type based on the comparison result of the structural tensor feature value difference with the preset edge determination threshold. When the feature value difference is greater than the preset edge determination threshold, it is determined to be an edge structure region, and when the feature value difference is not greater than the preset edge determination threshold, it is determined to be a homogeneous structure region.

[0084] Step 2: The adaptive enhancement module 2 receives the feature data output by the cartilage feature extraction module 1, generates differentiated enhancement weight parameters according to the type of cartilage tissue region, sets different enhancement weights for different regions, selects the corresponding enhancement algorithm according to the local structure type, applies the edge-preserving enhancement algorithm to the edge structure region, applies the smoothing enhancement algorithm to the homogeneous structure region, performs regional adaptive enhancement processing on the cartilage tissue, and outputs the preliminary enhanced image.

[0085] Step 3: The contrast optimization module 3 receives the preliminary enhanced image output by the adaptive enhancement module 2, statistically analyzes the signal intensity distribution range of cartilage tissue, subchondral bone, and joint effusion, determines the contrast stretching interval, applies the nonlinear Sigmoid mapping function to transform the signal intensity of cartilage tissue, adaptively adjusts the slope parameter of the contrast mapping curve according to the gradient information at the cartilage edge, and outputs the final enhanced image after contrast optimization.

[0086] Step 4: The quality assessment feedback module 4 receives the final enhanced image output by the contrast optimization module 3, calculates the cartilage display sharpness index (contrast-to-noise ratio) and edge preservation index (edge ​​gradient direction consistency), and determines whether the cartilage display sharpness index is not lower than the first preset threshold and whether the edge preservation index is not lower than the second preset threshold. If both threshold requirements are met, the final enhanced image is output, and the enhancement process is complete. If any index does not meet the corresponding threshold requirement, a parameter adjustment instruction is generated and fed back to the adaptive enhancement module 2 and the contrast optimization module 3 to adjust the enhancement weight parameters and contrast mapping curve parameters, and returns to step 2 to re-enhance the image until the quality index meets the requirements or the maximum number of iterations is reached.

[0087] Through the above workflow, the present invention achieves deep coupling and closed-loop collaboration of cartilage feature extraction module 1, adaptive enhancement module 2, contrast optimization module 3 and quality assessment feedback module 4, forming a complete intelligent enhancement and contrast optimization system.

[0088] To verify the technical effectiveness of this invention, enhancement processing experiments were conducted on MRI images of 50 cases of articular cartilage. The experiment used a Siemens Skyra 3.0T MRI scanner, employing a multi-echo spin-echo sequence to acquire MRI images of the knee joint. Imaging parameters were: repetition time 2500 ms, echo time 10-70 ms (10 ms interval), slice thickness 3 mm, slice spacing 0.5 mm, field of view 160 mm × 160 mm, and matrix 512 × 512. The 50 images included 15 normal knee joints, 20 cases of early-stage osteoarthritis, and 15 cases of mid-to-late-stage osteoarthritis.

[0089] Experimental results show that, after processing with the system of this invention, the MRI images exhibit an average improvement of 73.2% in the clarity of cartilage tissue compared to the original images, a 58.7% improvement in contrast between subchondral bone and cartilage tissue, and a 61.3% improvement in contrast between joint effusion and cartilage tissue. The clarity of cartilage edges is significantly improved, with an average edge gradient direction consistency of 0.94, indicating that the enhancement process effectively maintains the integrity of the cartilage edges while improving contrast. For early-stage osteoarthritis, the system of this invention can clearly display early pathological features such as slight unevenness of cartilage, local signal elevation, and prolonged T2 relaxation time, providing important evidence for early diagnosis. For mid-to-late-stage osteoarthritis, the system of this invention can clearly display significant cartilage defects, full-thickness destruction, and subchondral bone changes, providing accurate imaging information for treatment planning.

[0090] Compared with the existing technology CN103456004A, this invention demonstrates significant advantages in cartilage display clarity, contrast enhancement, and edge preservation. Existing technologies primarily detect sheet-like structures and perform grayscale enhancement using Hessian matrix eigenvalues, which has limited ability to distinguish cartilage in different pathological states, lacks specificity in enhancement effects, and cannot adjust parameters based on the enhancement results. This invention, by introducing T2 relaxation time characteristics and multi-scale structural tensor features, achieves adaptive enhancement based on physiological characteristics for region classification and structure perception. Through dynamic contrast optimization and a closed-loop feedback mechanism, it achieves adaptive adjustment of enhancement parameters, significantly improving the quality and stability of the enhancement effect.

[0091] In summary, the intelligent enhancement and contrast optimization system for articular cartilage MRI images provided by this invention achieves intelligent enhancement and dynamic contrast optimization of cartilage tissue through the deep coupling and synergy of the cartilage feature extraction module, adaptive enhancement module, contrast optimization module, and quality assessment feedback module. This provides high-quality imaging support for the accurate diagnosis and treatment planning of articular cartilage lesions, and has significant clinical application value and broad application prospects.

Claims

1. A system for intelligent enhancement and contrast optimization of MRI images of articular cartilage, characterized in that, include: The cartilage feature extraction module calculates the T2 relaxation time distribution features and multi-scale structural tensor features of cartilage tissue based on the acquired articular cartilage MRI image data. The T2 relaxation time distribution features characterize the water content and collagen fiber arrangement of cartilage tissue, and the multi-scale structural tensor features characterize the local structural information of cartilage tissue at different spatial scales. An adaptive enhancement module, connected to the cartilage feature extraction module, determines the region type of cartilage tissue based on the T2 relaxation time distribution characteristics, generates corresponding enhancement weight parameters for different region types, and performs region-specific adaptive enhancement processing on the cartilage tissue based on the multi-scale structural tensor features and the enhancement weight parameters. A contrast optimization module, connected to the adaptive enhancement module, receives the enhanced image data output by the adaptive enhancement module. Based on the signal difference characteristics between cartilage tissue and surrounding tissue in the enhanced image data, it dynamically adjusts the contrast mapping curve of the cartilage region. The contrast mapping curve is determined in real time based on the signal intensity distribution and edge gradient information of the cartilage tissue, thereby significantly improving the contrast between the cartilage tissue and the subchondral bone and joint effusion. The quality assessment feedback module, connected to the contrast optimization module, calculates the cartilage display sharpness index and edge preservation index of the enhanced image. When the cartilage display sharpness index is lower than a first preset threshold or the edge preservation index is lower than a second preset threshold, it generates parameter adjustment instructions and feeds them back to the adaptive enhancement module and the contrast optimization module to adjust the enhancement weight parameters and the contrast mapping curve, forming a closed-loop feedback optimization mechanism.

2. The system according to claim 1, characterized in that, The process of calculating the T2 relaxation time distribution characteristics by the cartilage feature extraction module includes: extracting multi-echo signal sequences from the articular cartilage MRI image data, fitting the T2 relaxation time values ​​of each pixel in the cartilage tissue based on the multi-echo signal sequences, and dividing the cartilage tissue into normal cartilage regions, early degeneration regions, and pathological cartilage regions according to the statistical distribution characteristics of the T2 relaxation time values.

3. The system according to claim 1, characterized in that, The process of calculating multi-scale structural tensor features by the cartilage feature extraction module includes: performing multi-scale decomposition on the articular cartilage MRI image data, calculating the gradient covariance matrix of the neighborhood of the pixel at different scales, wherein the eigenvalues ​​of the gradient covariance matrix represent the principal and secondary directional intensities of the local structure, and the eigenvectors of the gradient covariance matrix represent the directional information of the local structure.

4. The system according to claim 1, characterized in that, The process of generating enhancement weight parameters by the adaptive enhancement module includes: setting a first enhancement weight for normal cartilage areas, setting a second enhancement weight for early degeneration areas, and setting a third enhancement weight for pathological cartilage areas. The first enhancement weight is less than the second enhancement weight, and the second enhancement weight is less than the third enhancement weight, so that the pathological cartilage areas obtain higher enhancement intensity.

5. The system according to claim 1, characterized in that, When the adaptive enhancement module performs regional adaptive enhancement processing, it determines the local structure type based on the feature value difference of the multi-scale structural tensor features. When the feature value difference is greater than the preset edge determination threshold, it is determined to be an edge structure region, and an edge preservation enhancement algorithm is applied to the edge structure region. When the feature value difference is not greater than the preset edge determination threshold, it is determined to be a homogeneous structure region, and a smoothing enhancement algorithm is applied to the homogeneous structure region.

6. The system according to claim 1, characterized in that, The process of dynamically adjusting the contrast mapping curve by the contrast optimization module includes: statistically analyzing the signal intensity distribution ranges of cartilage tissue, subchondral bone, and joint effusion; determining the contrast stretching interval based on the signal intensity distribution range; and performing nonlinear mapping on the signal intensity of cartilage tissue within the contrast stretching interval to increase the signal intensity difference between cartilage tissue and surrounding tissues.

7. The system according to claim 1, characterized in that, When the contrast optimization module adjusts the contrast mapping curve based on edge gradient information, it calculates the gradient magnitude at the edge of the cartilage tissue. When the gradient magnitude is lower than the third preset threshold, it increases the slope of the contrast mapping curve. When the gradient magnitude is higher than the third preset threshold, it decreases the slope of the contrast mapping curve to avoid artifacts caused by excessive edge enhancement.

8. The system according to claim 1, characterized in that, The process of calculating the cartilage display clarity index by the quality assessment feedback module includes: calculating the mean and variance of the signal intensity of the cartilage tissue region in the enhanced image; determining the contrast-to-noise ratio of the cartilage tissue based on the mean signal intensity and the variance; and the contrast-to-noise ratio characterizing the display clarity of the cartilage tissue relative to the background noise.

9. The system according to claim 1, characterized in that, The process of calculating the edge retention index by the quality assessment feedback module includes: extracting the gradient direction consistency of the cartilage tissue edge before and after enhancement treatment. The gradient direction consistency is determined by the cosine value of the angle between the gradient vectors of the edge before and after enhancement. When the gradient direction consistency is higher than the second preset threshold, the edge is judged to be well maintained.

10. The system according to claim 1, characterized in that, The process of generating parameter adjustment instructions by the quality assessment feedback module includes: when the cartilage display clarity index is lower than the first preset threshold, generating an adjustment instruction to increase the enhancement weight parameter; when the edge preservation index is lower than the second preset threshold, generating an adjustment instruction to decrease the enhancement weight parameter or adjust the slope of the contrast mapping curve. The parameter adjustment instructions achieve a balance between enhancement effect and edge preservation through an iterative optimization mechanism.

Images