Hip replacement preoperative reconstruction system and device based on intelligent learning

Abstract

The invention provides a hip joint replacement preoperative reconstruction system and device based on intelligent learning, and relates to the technical field of auxiliary medical instruments. The hip joint replacement preoperative reconstruction system based on intelligent learning comprises an image quality evaluation module, an image fusion evaluation module and a prosthesis planning module. According to the invention, whether to carry out image quality optimization is judged through the image quality evaluation value obtained through the image data, and then whether to carry out image fusion quality optimization is judged based on the image fusion quality evaluation value obtained through the image fusion quality data. Finally, whether feedback is conducted or not is judged based on the prosthesis position data, the image fusion precision, the image quality evaluation value and a prosthesis planning evaluation value obtained through the image fusion quality evaluation value, so that hip joint prosthesis position selection is optimized, and then improvement of the relevance between the hip joint replacement preoperative reconstruction process and the image quality is achieved; the problem that the relevance between the hip joint replacement preoperative reconstruction process and the image quality is low in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of auxiliary medical devices, and particularly to a hip replacement preoperative reconstruction system and device based on intelligent learning. Background Art

[0002] Total hip arthroplasty is one of the common methods for treating hip joint abnormalities. In order to restore the function of the affected hip joint, surgeons often need to select an appropriate placement position of the acetabular prosthesis according to the pelvic anatomy. Hip joint function abnormalities in common patients are often caused by congenital skeletal dysplasia and acquired hip osteoarthritis for various reasons. In such patients, the pelvis develops normally or is slightly smaller on the affected side, and there is often a clear true acetabulum position. Surgeons usually choose to place the acetabular prosthesis at the true acetabulum position. The hip replacement preoperative reconstruction system and device based on intelligent learning realize the whole process of intelligence and precision from data acquisition to preoperative planning, which helps to improve the safety and success rate of hip replacement surgery and bring better treatment effects and quality of life to patients.

[0003] In the prior art, by acquiring the imaging data of patients, automatically analyzing the hip joint state of patients, and providing a more accurate preoperative reconstruction plan, the improvement of the surgical effect and the recovery speed of patients is realized.

[0004] However, in the process of implementing the technical solutions of the embodiments of the present invention, it is found that the above technologies have at least the following technical problems:

[0005] In the prior art, hip replacement preoperative reconstruction often requires fusing multiple types of imaging data. Each data type has its specific imaging characteristics. Due to various reasons, problems such as poor image quality and difficult data fusion processing may occur, and there is a problem of low correlation between the hip replacement preoperative reconstruction process and image quality.

[0006] However, there are four types of patients who cannot provide the current preoperative placement plan of the acetabular prosthesis: (1) The bilateral hip joints of the patient are fused, and the acetabulum position can no longer be identified; (2) The bilateral pelvises are severely asymmetric in development. If the acetabular prosthesis is placed at the true acetabulum position, it will cause pelvic tilt and spinal scoliosis; (3) The femoral head is highly dislocated, and the original acetabulum position has not developed; (4) The patient has undergone osteotomy to preserve the hip joint, but with the progression of the disease, the affected side of the patient has a high dislocation again or the original structure of the pelvic bone has been completely damaged due to the osteotomy surgery. Summary of the Invention

[0007] The present invention provides a hip joint replacement preoperative reconstruction system and device based on intelligent learning, which can not only improve the quality of images, reduce errors caused by image noise, blurring or unclear images, but also optimize the fusion of multi-modal images, make the overlap and docking between different images more accurate, and ensure that the preoperative image data of hip joint replacement reaches the best state, ultimately improving the correlation between the preoperative reconstruction process of hip joint replacement and image quality.

[0008] To solve the above technical problems, the technical solutions provided by the present invention are as follows:

[0009] A hip joint replacement preoperative reconstruction system based on intelligent learning, comprising: an image quality evaluation module, an image fusion evaluation module and a prosthesis planning module; wherein, the image quality evaluation module is used to evaluate the obtained image data to obtain an image quality evaluation value, and judge whether to optimize the image quality based on the image quality evaluation value, and the image quality evaluation value is used to evaluate the image quality; the image fusion evaluation module is used to evaluate the obtained image fusion quality data to obtain an image fusion quality evaluation value, and judge whether to optimize the image fusion quality based on the image fusion quality evaluation value, and the image fusion quality evaluation value is used to evaluate the image fusion quality; the prosthesis planning module is used to, if image optimization is performed, evaluate the obtained prosthesis planning evaluation value based on the prosthesis position data, the image fusion accuracy, the image quality evaluation value obtained after image quality optimization and the image fusion quality evaluation value obtained after image fusion quality optimization, and judge whether to give feedback based on the prosthesis planning evaluation value, the image optimization includes image quality optimization and image fusion quality optimization, and the image fusion quality evaluation value is used to evaluate the rationality of the hip joint prosthesis position.

[0010] Optionally, the image data includes image signal-to-noise ratio, structural similarity index, information entropy, and contrast; the image fusion quality data includes registration error, relative information gain, mutual information, and fused image signal-to-noise ratio; the image fusion accuracy is obtained by processing the ratio of the pixel values of the fused image to the pixel values of the standard image; the prosthesis position data includes prosthesis anteversion angle, prosthesis insertion depth, and prosthesis pitch angle; the image signal-to-noise ratio includes fused image signal-to-noise ratio, CT image signal-to-noise ratio, MRI image signal-to-noise ratio, and X-ray image signal-to-noise ratio; the structural similarity index is obtained by processing the ratio operations of the average pixel value, pixel standard deviation, and pixel squared difference of the image corresponding to the average pixel value, pixel standard deviation, and pixel squared difference of the standard image, and the structural similarity index includes CT image structural similarity index, MRI image structural similarity index, and X-ray image structural similarity index; the information entropy is obtained by processing the occurrence frequency of each image pixel value obtained from the constructed image histogram, and the information entropy includes fused image information entropy, CT image information entropy, MRI image information entropy, and X-ray image information entropy; the contrast is obtained by performing a ratio operation on the difference between the maximum image pixel value and the minimum image pixel value and the average image pixel value, and the contrast includes CT image contrast, MRI image contrast, and X-ray image contrast; the relative information gain is obtained by processing the ratio operations of the fused image information entropy, CT image information entropy, MRI image information entropy, and X-ray image information entropy to the theoretical maximum entropy value in the database; the mutual information is obtained by processing the joint probability distribution and the marginal probability distribution, and the marginal probability distribution is obtained by performing a ratio operation on the occurrence frequency of the pixel values of the corresponding image in the image to the total number of pixels of the corresponding image; the mutual information includes first mutual information, second mutual information, third mutual information, fourth mutual information, fifth mutual information, and sixth mutual information, the first mutual information represents the mutual information between the CT image and the MRI image, the second mutual information represents the mutual information between the CT image and the X-ray image, the third mutual information represents the mutual information between the CT image and the fused image, the fourth mutual information represents the mutual information between the MRI image and the X-ray image, the fifth mutual information represents the mutual information between the MRI image and the fused image, and the sixth mutual information represents the mutual information between the X-ray image and the fused image.

[0011] Optionally, the specific method for obtaining the image quality evaluation value by evaluating according to the acquired image data is as follows: obtaining the signal-to-noise ratio deviation according to the relative relationship between the signal-to-noise ratio and the corresponding preset signal-to-noise ratio in the database; obtaining the information entropy deviation according to the relative relationship between the information entropy and the corresponding preset information entropy in the database; obtaining the contrast deviation according to the relative relationship between the contrast and the corresponding preset contrast in the database; and processing the signal-to-noise ratio deviation, information entropy deviation, contrast deviation, and structural similarity index to obtain the image quality evaluation value.

[0012] Optionally, the specific limiting expression of the image quality evaluation value is:

[0013]

[0014] In the formula, i represents the number of the image type, i = 1 represents CT image, i = 2 represents MRI image, i = 3 represents X-ray image, YXZ i represents the image quality evaluation value of the i-th image type, SNR i represents the signal-to-noise ratio of the i-th image type, SSIM i represents the structural similarity index of the i-th image type, XXS i represents the information entropy of the i-th image type, DBD i represents the contrast of the i-th image type, SNR i.0 represents the preset signal-to-noise ratio of the i-th image type, XXS i.0 represents the preset information entropy of the i-th image type, DBD i.0 represents the preset contrast of the i-th image type.

[0015] Optionally, the specific method for evaluating the image fusion quality evaluation value based on the obtained image fusion quality data is as follows: obtaining the registration error deviation according to the relative relationship between the registration error and the corresponding preset registration error in the database; obtaining the mutual information deviation according to the relative relationship between the total mutual information and the corresponding preset mutual information in the database; obtaining the fused image signal-to-noise ratio deviation according to the relative relationship between the fused image signal-to-noise ratio and the corresponding preset fused image signal-to-noise ratio in the database; processing the signal-to-noise ratio deviation, relative information gain, mutual information deviation and fused image signal-to-noise ratio deviation to obtain the image fusion quality evaluation value.

[0016] Optionally, the specific method for evaluating the prosthesis planning evaluation value based on the prosthesis position data, the image fusion accuracy, the image quality evaluation value obtained after image quality optimization, and the image fusion quality evaluation value obtained after image fusion quality optimization is as follows: Obtain the image fusion accuracy deviation according to the relative relationship between the image fusion accuracy and the corresponding preset image fusion accuracy in the database. The image fusion accuracy deviation is obtained by processing the ratio of the image fusion accuracy to the corresponding preset image fusion accuracy in the database; Obtain the image quality optimization coefficient according to the relative relationship between the image quality evaluation value obtained after image quality optimization, the image quality evaluation value obtained before image quality optimization, and the preset image quality threshold; Obtain the image fusion quality optimization coefficient according to the relative relationship between the image fusion quality evaluation value obtained after image fusion quality optimization, the image fusion quality evaluation value obtained before image fusion quality optimization, and the preset image fusion quality threshold; If the prosthesis anteversion angle is not within the preset prosthesis anteversion angle range, obtain the anteversion angle coefficient according to the relative relationship between the prosthesis anteversion angle and the preset prosthesis anteversion angle, otherwise record the anteversion angle coefficient as 1; If the prosthesis insertion depth is not within the preset prosthesis insertion depth range, obtain the depth coefficient according to the relative relationship between the prosthesis insertion depth and the preset prosthesis insertion depth, otherwise record the depth coefficient as 1; If the prosthesis pitch angle is not within the preset prosthesis pitch angle range, obtain the pitch angle coefficient according to the relative relationship between the prosthesis pitch angle and the preset prosthesis pitch angle, otherwise record the pitch angle coefficient as 1; Process the anteversion angle coefficient, the depth coefficient, the pitch angle coefficient, the image fusion accuracy deviation, the image quality optimization coefficient, and the image fusion quality optimization coefficient to obtain the prosthesis planning evaluation value.

[0017] Optionally, the specific process for determining whether to give feedback based on the prosthesis planning evaluation value is as follows: Determine whether the prosthesis planning evaluation value is less than the preset prosthesis planning threshold. If the prosthesis planning evaluation value is less than the preset prosthesis planning threshold, adjust the number of wavelet transform layers, otherwise do not give feedback; Determine whether the prosthesis planning evaluation value obtained after adjusting the number of wavelet transform layers is less than the preset prosthesis planning threshold. If the prosthesis planning evaluation value is less than the preset prosthesis planning threshold, adjust the prosthesis position, otherwise end the feedback; Determine whether the prosthesis planning evaluation value obtained after adjusting the prosthesis position is less than the preset prosthesis planning threshold. If the prosthesis planning evaluation value is less than the preset prosthesis planning threshold, give feedback to the preset personnel, otherwise end the feedback.

[0018] The embodiments of the present invention provide a device applied to the hip joint replacement preoperative reconstruction system based on intelligent learning, including: a data acquisition device, a storage device, and a processing device; wherein, the data acquisition device is used to collect the image fusion accuracy, image data, prosthesis position data, and image fusion quality data of the hip joint, and the data acquisition device includes an X-ray machine, a computed tomography scanner, and a magnetic resonance imaging machine; the storage device is used to store the image fusion accuracy, image data, prosthesis position data, and image fusion quality data of the hip joint; the processing device is used to perform image evaluation according to the data in the storage device, and the image evaluation includes image quality evaluation, image fusion quality evaluation, and prosthesis planning evaluation.

[0019] The above technical solution has at least the following beneficial effects compared with the prior art:

[0020] 1. Determine whether to optimize the image quality through the image quality evaluation value obtained from the image data, then determine whether to optimize the image fusion quality through the image fusion quality evaluation value obtained from the image fusion quality data, and finally determine whether to give feedback through the prosthesis planning evaluation value obtained from the prosthesis position data, image fusion accuracy, image quality evaluation value, and image fusion quality evaluation value, thereby optimizing the selection of the prosthesis position, and then improving the relevance between the hip joint replacement preoperative reconstruction process and the image quality, effectively solving the problem of low relevance between the hip joint replacement preoperative reconstruction process and the image quality in the prior art.

[0021] 2. Obtain the signal-to-noise ratio deviation through the relative relationship between the signal-to-noise ratio and the corresponding preset signal-to-noise ratio in the database, then obtain the information entropy deviation according to the relative relationship between the information entropy and the corresponding preset information entropy in the database, then obtain the contrast deviation according to the relative relationship between the contrast and the corresponding preset contrast in the database, and finally process the signal-to-noise ratio deviation, information entropy deviation, contrast deviation, and structural similarity index to obtain the image quality evaluation value, thereby quantitatively evaluating the image quality, and then providing optimization support for image fusion in the subsequent hip joint replacement preoperative reconstruction process.

[0022] 3. Obtain the image fusion accuracy deviation through the image fusion accuracy and the preset image fusion accuracy, then obtain the anterior inclination angle coefficient according to the prosthesis anterior inclination angle and the preset prosthesis anterior inclination angle, then obtain the depth coefficient according to the prosthesis insertion depth and the preset prosthesis insertion depth, then obtain the pitch angle coefficient according to the prosthesis pitch angle and the preset prosthesis pitch angle, and finally obtain the prosthesis planning evaluation value according to the anterior inclination angle coefficient, depth coefficient, pitch angle coefficient, image fusion accuracy deviation, image quality optimization coefficient, and image fusion quality optimization coefficient, thereby quantitatively evaluating the rationality of the hip joint prosthesis position, and then improving the accuracy of prosthesis planning. Brief Description of the Drawings

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0024] Figure 1 It is a schematic structural diagram of a hip joint replacement preoperative reconstruction system based on intelligent learning provided by an embodiment of the present invention. Specific embodiments

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the described embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0026] Unless otherwise defined, the technical terms or scientific terms used in the present invention should have the ordinary meaning understood by those of ordinary skill in the art to which the present invention belongs. The "first", "second", and similar terms used in the present invention do not indicate any order, quantity, or importance, but are only used to distinguish different components. Similarly, terms such as "a", "an", or "the" do not indicate a quantity limitation, but indicate the existence of at least one. The terms "including" or "comprising" and the like mean that the elements or items appearing before the term cover the elements or items listed after the term and their equivalents, without excluding other elements or items. The terms "connected" or "coupled" and the like are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect.

[0027] It should be noted that the "upper", "lower", "left", "right", "front", "rear", etc. used in the present invention are only used to represent relative positional relationships. When the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0028] In view of the problem of low correlation between the existing hip joint replacement preoperative reconstruction process and image quality, the present invention provides a hip joint replacement preoperative reconstruction system and device based on intelligent learning that can improve the quality of images, reduce errors caused by image noise, blurring, or unclear images, and can optimize multi-modal image fusion.

[0029] As Figure 1As shown in the figure, an embodiment of the present invention provides a hip replacement preoperative reconstruction system based on intelligent learning. The system includes: an image quality evaluation module, an image fusion evaluation module, and a prosthesis planning module. Among them, the image quality evaluation module is used to evaluate the acquired image data to obtain an image quality evaluation value, and based on the image quality evaluation value, it determines whether to optimize the image quality. The image quality evaluation value is used to evaluate the image quality. The image fusion evaluation module is used to evaluate the acquired image fusion quality data to obtain an image fusion quality evaluation value, and based on the image fusion quality evaluation value, it determines whether to optimize the image fusion quality. The image fusion quality evaluation value is used to evaluate the image fusion quality. The prosthesis planning module is used to, if image optimization is performed, evaluate the prosthesis position data, the image fusion accuracy, the image quality evaluation value obtained after image quality optimization, and the image fusion quality evaluation value obtained after image fusion quality optimization to obtain a prosthesis planning evaluation value, and based on the prosthesis planning evaluation value, it determines whether to give feedback. Image optimization includes image quality optimization and image fusion quality optimization. The image fusion quality evaluation value is used to evaluate the rationality of the hip prosthesis position.

[0030] It should be added that the images include CT images (Computed Tomography), MRI images (Magnetic Resonance Imaging), and X-ray images; the image data includes the signal-to-noise ratio of the image, the structural similarity index, the information entropy, and the contrast; the image fusion quality data includes the registration error, the relative information gain, the mutual information, and the signal-to-noise ratio of the fused image; the image fusion accuracy is obtained by processing the ratio of the pixel values of the fused image to the pixel values of the standard image; the prosthesis position data includes the anteversion angle of the prosthesis, the insertion depth of the prosthesis, and the pitch angle of the prosthesis.

[0031] The signal-to-noise ratio of the image is obtained by processing the ratio operation of the average power of the signal region and the average power of the noise region in the image obtained by an image processing software (such as Photoshop). The signal-to-noise ratio of the image includes the signal-to-noise ratio of the fused image, the signal-to-noise ratio of the CT image, the signal-to-noise ratio of the MRI image, and the signal-to-noise ratio of the X-ray image.

[0032] The structural similarity index is obtained by processing the ratio operation of the average pixel value, the pixel standard deviation, and the pixel squared difference of the image respectively with the average pixel value, the pixel standard deviation, and the pixel squared difference of the standard image. The structural similarity index includes the structural similarity index of the CT image, the structural similarity index of the MRI image, and the structural similarity index of the X-ray image.

[0033] The information entropy is obtained by processing the occurrence frequency of each image pixel value obtained from the constructed image histogram. The specific formula of the information entropy is: where x represents the number of the gray level, x = 1, 2,..., 255, px It represents the occurrence probability of the x-th gray level; if the occurrence probability of a certain gray level is 0, the information entropy is 0; the information entropy includes the information entropy of the fused image, the information entropy of the CT image, the information entropy of the MRI image, and the information entropy of the X-ray image.

[0034] The contrast is obtained by calculating the ratio of the difference between the maximum value and the minimum value of the image pixel values to the average value of the image pixel values. The contrast includes the contrast of the CT image, the contrast of the MRI image, and the contrast of the X-ray image.

[0035] The registration error is obtained by the Euclidean distance between the corresponding points between the image obtained by the image registration algorithm and the fused image; the relative information gain is obtained by processing the ratio of the information entropy of the fused image, the information entropy of the CT image, the information entropy of the MRI image, and the information entropy of the X-ray image to the theoretical maximum entropy value in the database.

[0036] The mutual information is obtained by processing the joint probability distribution and the marginal probability distribution. The joint probability distribution is obtained by calculating the ratio of the occurrence frequency of each pair of pixel values in the two images obtained by statistics to the total number of pixels in the two images, and the marginal probability distribution is obtained by calculating the ratio of the occurrence frequency of the pixel values in the corresponding image to the total number of pixels in the corresponding image; the mutual information includes the first mutual information, the second mutual information, the third mutual information, the fourth mutual information, the fifth mutual information, and the sixth mutual information. The first mutual information represents the mutual information between the CT image and the MRI image, the second mutual information represents the mutual information between the CT image and the X-ray image, the third mutual information represents the mutual information between the CT image and the fused image, the fourth mutual information represents the mutual information between the MRI image and the X-ray image, the fifth mutual information represents the mutual information between the MRI image and the fused image, and the sixth mutual information represents the mutual information between the X-ray image and the fused image.

[0037] The anteversion angle of the prosthesis, the insertion depth of the prosthesis, and the pitch angle of the prosthesis are measured by analyzing the preoperative three-dimensional image data.

[0038] Through the above steps, the image quality assessment module conducts image quality assessment based on the acquired CT images, MRI images, and X-ray images. These images are crucial in the pre-operative reconstruction of hip replacement as they provide detailed information on joint structure, bone condition, and potential complications. By means of the image quality assessment module, the accuracy of pre-operative planning is ensured, and the surgical risks caused by image quality problems are reduced. The image fusion assessment module evaluates the quality of multi-modal image fusion. In the pre-operative reconstruction of hip replacement, multi-source image fusion can provide more comprehensive and accurate joint structure information, helping doctors formulate surgical plans more accurately and ensuring the precision and reliability of pre-operative planning. The prosthesis planning module can further improve the accuracy and practicality of pre-operative planning for hip replacement by continuously optimizing image quality and fusion quality. It realizes the comprehensive assessment and optimization of image quality and image fusion quality, and further improves the correlation between the pre-operative reconstruction process of hip replacement and image quality.

[0039] The specific method for obtaining the image quality assessment value by evaluating according to the acquired image data is as follows: The signal-to-noise ratio deviation is obtained based on the relative relationship between the signal-to-noise ratio and the corresponding preset signal-to-noise ratio in the database. The signal-to-noise ratio deviation is obtained through the ratio operation of the signal-to-noise ratio and the corresponding preset signal-to-noise ratio in the database; the information entropy deviation is obtained based on the relative relationship between the information entropy and the corresponding preset information entropy in the database. The information entropy deviation is obtained through the ratio operation of the information entropy and the corresponding preset information entropy in the database; the contrast deviation is obtained based on the relative relationship between the contrast and the corresponding preset contrast in the database. The contrast deviation is obtained through the ratio operation of the contrast and the corresponding preset contrast in the database; the signal-to-noise ratio deviation, information entropy deviation, contrast deviation, and structural similarity index are processed to obtain the image quality assessment value.

[0040] Among them, the specific limit expression of the image quality assessment value is:

[0041]

[0042] In the formula, i represents the number of the image type. i = 1 represents CT images, i = 2 represents MRI images, and i = 3 represents X-ray images, YXZ i represents the image quality assessment value of the i-th image type, SNR i represents the signal-to-noise ratio of the i-th image type, SSIM i represents the structural similarity index of the i-th image type, XXS i represents the information entropy of the i-th image type, DBD i represents the contrast of the i-th image type, SNR i.0Denote the preset signal-to-noise ratio of the i-th type of image, XXS i.0 Denote the preset information entropy of the i-th type of image, DBD i.0 Denote the preset contrast of the i-th type of image.

[0043] In this embodiment, the preset signal-to-noise ratio, preset information entropy, and preset contrast are all set according to industry standards. For example, the preset signal-to-noise ratio is set to 35 dB, the preset information entropy is set to 7 bits, and the preset contrast is set to 0.7.

[0044] In this algorithm, the image quality evaluation value is obtained by processing multiple independent variables (image signal-to-noise ratio, structural similarity index, information entropy, and contrast). There are mutual influence relationships among these independent variables; images with higher image signal-to-noise ratios usually have lower noise, so it is easier to obtain higher structural similarity indices in terms of structure and detail retention; because low-noise images contain more effective information, the higher the image signal-to-noise ratio, the greater the information entropy; the higher the image signal-to-noise ratio, the more obvious the brightness difference in the image, and the stronger the contrast; the higher the structural similarity index, the increase in the similarity of the image structure may lead to an increase in the information entropy because high-quality images usually contain more details and rich information; the greater the contrast helps to enhance the details and structure in the image, making the structural similarity index of the image higher; the higher the contrast of the image, usually contains more details and variations, so the information entropy may also be higher; the image signal-to-noise ratio, structural similarity index, information entropy, and contrast are positively correlated with the image quality evaluation value.

[0045] Through the above steps, the quality of the image is quantitatively evaluated, effectively reflecting the overall quality of the image, and further providing optimization support for image fusion in the subsequent pre-operative reconstruction process of hip replacement.

[0046] The specific process of determining whether to optimize the image quality based on the image quality evaluation value is as follows: A1. Determine whether the image quality evaluation value is less than the preset image quality threshold. If the image quality evaluation value is less than the preset image quality threshold, execute A2; otherwise, do not perform image quality optimization. A2. Perform denoising processing, and determine whether the image quality evaluation value obtained after denoising is less than the preset image quality threshold. If the image quality evaluation value is less than the preset image quality threshold, execute A3; otherwise, end the image quality optimization. A3. Perform contrast enhancement, and determine whether the image quality evaluation value obtained after contrast enhancement is less than the preset image quality threshold. If the image quality evaluation value is less than the preset image quality threshold, execute A4; otherwise, end the image quality optimization. A4. Perform image reconstruction and repair, and determine whether the image quality evaluation value obtained after image reconstruction and repair is less than the preset image quality threshold. If the image quality evaluation value is less than the preset image quality threshold, execute A5; otherwise, end the image quality optimization. A5. Perform structure optimization, and determine whether the image quality evaluation value obtained after structure optimization is less than the preset image quality threshold. If the image quality evaluation value is less than the preset image quality threshold, give feedback; otherwise, end the image quality optimization.

[0047] In this embodiment, the preset image quality threshold is represented by the maximum value of the image quality evaluation values of qualified images in the historical time period; the denoising processing is performed on the local noise characteristics of the image by the Feed-forward Denoising Convolutional Neural Networks (DnCNN); the contrast enhancement adjusts the image contrast through a convolutional neural network; the image reconstruction and repair reconstructs the missing area through a generative adversarial network; the structure optimization optimizes the image structure through super-resolution technology and edge enhancement technology; through the above steps, the improvement of the image quality in the pre-operative reconstruction process of hip replacement is realized.

[0048] The specific method for obtaining the image fusion quality evaluation value by evaluating the obtained image fusion quality data is as follows: Obtain the registration error deviation according to the relative relationship between the registration error and the corresponding preset registration error in the database (i.e., in the specific method for obtaining the image fusion quality evaluation value), and the registration error deviation is obtained by processing the ratio operation of the registration error and the corresponding preset registration error in the database; obtain the mutual information deviation according to the relative relationship between the total mutual information and the corresponding preset mutual information in the database (i.e., ) The mutual information deviation is obtained by processing the ratio of the total mutual information to the corresponding preset mutual information in the database. The total mutual information represents the average value of the mutual information; the fusion image signal-to-noise ratio deviation is obtained according to the relative relationship between the fusion image signal-to-noise ratio and the corresponding preset fusion image signal-to-noise ratio in the database (i.e., in the specific method for obtaining the image fusion quality evaluation value ) The fusion image signal-to-noise ratio deviation is obtained by processing the ratio of the fusion image signal-to-noise ratio to the corresponding preset fusion image signal-to-noise ratio in the database; the signal-to-noise ratio deviation, relative information gain, mutual information deviation, and fusion image signal-to-noise ratio deviation are processed to obtain the image fusion quality evaluation value.

[0049] Among them, the specific method for obtaining the image fusion quality evaluation value is:

[0050]

[0051] In the formula, i represents the number of the image type. i = 1 represents CT image, i = 2 represents MRI image, i = 3 represents X-ray image, RHZ represents the image fusion quality evaluation value, YXZ i ′ represents the image quality evaluation value obtained after optimizing the image quality of the i-th image type, PZWC represents the registration error, ZEY represents the relative information gain, HXX represents the mutual information, SNR represents the fusion image signal-to-noise ratio, SNR 0 represents the preset fusion image signal-to-noise ratio, MI 1,2 represents the first mutual information, MI 1,3 represents the second mutual information, MI 1,4 represents the third mutual information, MI 2,3 represents the fourth mutual information, MI 2,4 represents the fifth mutual information, MI 3,4 represents the sixth mutual information, PZWC 0 represents the preset registration error, HXX 0 represents the preset mutual information.

[0052] In this embodiment, the preset registration error is represented by the average value of the registration errors in the historical time period; the preset mutual information is represented by the average value of the mutual information in the historical time period; the preset fusion image signal-to-noise ratio is set according to industry standards. For example, the preset fusion image signal-to-noise ratio is set to 35 dB.

[0053] In this algorithm, the image fusion quality evaluation value is obtained by processing multiple independent variables (registration error, relative information gain, mutual information, and signal-to-noise ratio of the fused image). There are mutual influence relationships among these independent variables. When the registration error is larger, obvious misalignment or ghosting phenomena may occur in the fused image, resulting in a decrease in the signal-to-noise ratio of the fused image. Both relative information gain and mutual information are indicators for measuring the amount of information, and relative information gain can be regarded as a supplement to mutual information. If the signal-to-noise ratio of the fused image is low, useful information may not be effectively extracted, leading to a decrease in relative information gain and mutual information. The registration error is negatively correlated with the image fusion quality evaluation value, while relative information gain, mutual information, and the signal-to-noise ratio of the fused image are positively correlated with the image fusion quality evaluation value.

[0054] Through the above steps, the fusion effect between multi-modal images is quantitatively evaluated, and thus a reference basis is provided for optimizing subsequent image fusion.

[0055] The specific process for determining whether to optimize the image fusion quality based on the image fusion quality evaluation value is as follows: Determine whether the image fusion quality evaluation value is less than the preset image fusion quality threshold. If the image fusion quality evaluation value is less than the preset image fusion quality threshold, perform image quality optimization on the fused image; otherwise, do not perform image fusion quality optimization. Then determine whether the image fusion quality evaluation value obtained after image quality optimization is less than the preset image fusion quality threshold. If the image fusion quality evaluation value is less than the preset image fusion quality threshold, perform wavelet transform optimization; otherwise, end the image fusion quality optimization.

[0056] In this embodiment, the preset image fusion quality threshold is represented by the maximum value of the image fusion quality evaluation values of the images with qualified fusion quality within the historical time period. Wavelet transform optimization performs multi-resolution decomposition of the image through wavelet transform and performs fusion on different frequency bands. Optimization is performed on the low-frequency part because the low-frequency part usually contains the general structure and background information of the image, and optimizing the low-frequency part improves the overall effect of the image. Fusion is performed on the high-frequency part because the high-frequency part contains the detailed information of the image, and optimizing the details enhances the clarity and detail retention of the image. For example, if the sampling frequency of the original signal is 1000Hz, the highest frequency is 500Hz, the low-frequency part is 0 - 250Hz (the part below half of the highest frequency), and the high-frequency part is 250 - 500Hz (the part not lower than half of the highest frequency).

[0057] Through the cyclic adjustment of quantitative evaluation and optimization measures, it is ensured that the quality of the finally fused image meets the expected requirements, which helps to improve the effect of image fusion during the pre-operative reconstruction of hip replacement.

[0058] The specific method for obtaining the prosthesis planning evaluation value by evaluating according to the prosthesis position data, the image fusion accuracy, the image quality evaluation value obtained after optimizing the image quality, and the image fusion quality evaluation value obtained after optimizing the image fusion quality is as follows: The image fusion accuracy deviation is obtained according to the relative relationship between the image fusion accuracy and the corresponding preset image fusion accuracy in the database (i.e., ) in the specific method for obtaining the prosthesis planning evaluation value. The image fusion accuracy deviation is obtained by processing the ratio of the image fusion accuracy to the corresponding preset image fusion accuracy in the database; The image quality optimization coefficient is obtained according to the relative relationship between the image quality evaluation value obtained after optimizing the image quality, the image quality evaluation value obtained before optimizing the image quality, and the preset image quality threshold (i.e., YX in the specific method for obtaining the prosthesis planning evaluation value); The image fusion quality optimization coefficient is obtained according to the relative relationship between the image fusion quality evaluation value obtained after optimizing the image fusion quality, the image fusion quality evaluation value obtained before optimizing the image fusion quality, and the preset image fusion quality threshold (i.e., RH in the specific method for obtaining the prosthesis planning evaluation value); If the anterior inclination angle of the prosthesis is not within the preset anterior inclination angle range of the prosthesis, the anterior inclination angle coefficient is obtained according to the relative relationship between the anterior inclination angle of the prosthesis and the preset anterior inclination angle of the prosthesis (i.e., JD in the specific method for obtaining the prosthesis planning evaluation value), otherwise the anterior inclination angle coefficient is recorded as 1; If the insertion depth of the prosthesis is not within the preset insertion depth range of the prosthesis, the depth coefficient is obtained according to the relative relationship between the insertion depth of the prosthesis and the preset insertion depth of the prosthesis (i.e., JS in the specific method for obtaining the prosthesis planning evaluation value), otherwise the depth coefficient is recorded as 1; If the pitch angle of the prosthesis is not within the preset pitch angle range of the prosthesis, the pitch angle coefficient is obtained according to the relative relationship between the pitch angle of the prosthesis and the preset pitch angle of the prosthesis (i.e., JF in the specific method for obtaining the prosthesis planning evaluation value), otherwise the pitch angle coefficient is recorded as 1; The anterior inclination angle coefficient, the depth coefficient, the pitch angle coefficient, the image fusion accuracy deviation, the image quality optimization coefficient, and the image fusion quality optimization coefficient are processed to obtain the prosthesis planning evaluation value.

[0059] Among them, the specific method for obtaining the prosthesis planning evaluation value is:

[0060]

[0061] In the formula, YXH represents the prosthesis planning evaluation value, RHJ represents the image fusion accuracy, RHJ 0 represents the preset image fusion accuracy, YX represents the image quality optimization coefficient, RH represents the image fusion quality optimization coefficient, YXZ i ′ represents the image quality evaluation value obtained after optimizing the image quality of the i-th type of image, YXZ i represents the image quality evaluation value of the i-th type of image, YXZ 0 represents the preset image quality threshold, RHZ′ represents the image fusion quality evaluation value obtained after optimizing the image fusion quality. RHZ represents the image fusion quality evaluation value, and RHZ 0 represents the preset image fusion quality threshold. JD represents the anteversion coefficient, JS represents the depth coefficient, JF represents the pitch angle coefficient, JTD represents the prosthesis anteversion angle, JTS represents the prosthesis insertion depth, JTF represents the prosthesis pitch angle, and JTD 0 represents the preset prosthesis anteversion angle, and JTS 0 represents the preset prosthesis insertion depth, and JTF 0 represents the preset prosthesis pitch angle, and JTD min represents the preset minimum prosthesis anteversion angle, and JTD max represents the preset maximum prosthesis anteversion angle, and JTS min represents the preset minimum prosthesis insertion depth, and JTS max represents the preset maximum prosthesis insertion depth, and JTF min represents the preset minimum prosthesis pitch angle, and JTF max represents the preset maximum prosthesis pitch angle.

[0062] In this embodiment, the preset image fusion accuracy is represented by the average value of the image fusion accuracies of qualified fusion quality within a historical time period; the anteversion angle refers to the angle between the prosthesis and the front of the acetabulum, which affects the stability between the prosthesis and the femoral head; the preset prosthesis anteversion angle range is set according to industry standards. For example, the preset anteversion angle range is usually 20° ± 5°, the preset minimum prosthesis anteversion angle is set to 15°, the preset maximum prosthesis anteversion angle is set to 25°, and the preset prosthesis anteversion angle is set to the median 20° of the preset anteversion angle range; preoperative three-dimensional image data analysis (such as CT, MRI) can measure the anteversion angle of the acetabulum to obtain a personalized anteversion angle.

[0063] The pitch angle is the angle between the prosthesis and the pelvic plane during insertion, which affects the contact surface and stability of the prosthesis; the preset prosthesis pitch angle range is set according to industry standards. For example, the preset prosthesis pitch angle range is usually set from 40° to 50°, the preset minimum prosthesis pitch angle is set to 40°, the preset maximum prosthesis pitch angle is set to 50°, and the preset prosthesis pitch angle is set to the median 45° of the preset prosthesis pitch angle range; through preoperative image reconstruction, the anatomical shape of the patient's pelvis can be measured, and a personalized pitch angle can be set based on this data.

[0064] The depth of the prosthesis is usually determined based on the anatomical depth of the acetabulum. The adjustment of the depth should balance the prosthesis stability and the compression of soft tissues. The preset prosthesis insertion depth range is set according to industry standards and the individual conditions of patients. For example, the preset prosthesis insertion depth range is usually set to 50% to 70% of the acetabular depth. If the acetabular depth of the patient is 13 mm, then the preset prosthesis insertion depth range is 6.5 - 9.1, the preset minimum prosthesis insertion depth is 6.5 mm, the preset maximum prosthesis insertion depth is 9.1 mm, and the preset prosthesis insertion depth is set to the median value of the preset prosthesis insertion depth range, which is 7.8 mm.

[0065] In this algorithm, the prosthesis planning evaluation value is obtained by processing multiple independent variables (anteversion coefficient, depth coefficient, pitch angle coefficient, image fusion accuracy deviation, image quality optimization coefficient, and image fusion quality optimization coefficient). There are mutual influence relationships among these independent variables. The prosthesis anteversion angle and the prosthesis pitch angle jointly determine the spatial position and orientation of the prosthesis, and thus affect the stability and function of the prosthesis. Adjusting the prosthesis anteversion angle may affect the prosthesis pitch angle. If the prosthesis anteversion angle is not within the preset prosthesis anteversion angle range, it may be necessary to compensate by adjusting the insertion depth to ensure the overall stability and function of the prosthesis. If the prosthesis insertion depth is not within the preset prosthesis insertion depth range, the pitch angle can be adjusted to ensure the matching of the prosthesis and the acetabulum. The higher the image fusion accuracy, the closer the pixel values of the fused image are to the pixel values of the standard image, and the higher the image quality optimization coefficient. The improvement of image fusion accuracy helps to improve the image fusion quality optimization coefficient because a high-precision fusion result can retain more original information and thus show better performance in terms of fusion quality. The improvement of image quality helps to improve the fusion quality because better quality provides more useful information for the fusion process. The anteversion coefficient, depth coefficient, pitch angle coefficient, image fusion accuracy deviation, image quality optimization coefficient, and image fusion quality optimization coefficient are positively correlated with the prosthesis planning evaluation value.

[0066] Taking the preset image fusion accuracy as 1 as an example, the change statistical table of the prosthesis planning evaluation value is shown in Table 1:

[0067] Table 1 Change Statistical Table of Prosthesis Planning Evaluation Value

[0068]

[0069] It can be seen from the data in the first and second groups in the table that the prosthesis planning evaluation value increases as the image fusion accuracy increases; it can be seen from the data in the second and third groups that the prosthesis planning evaluation value increases as the image quality optimization coefficient increases; it can be seen from the data in the third and fourth groups that the prosthesis planning evaluation value increases as the image fusion quality optimization coefficient increases; it can be seen from the data in the fourth and fifth groups that the anteversion angle coefficient increases as the image fusion quality optimization coefficient increases; it can be seen from the data in the fifth and sixth groups that the depth coefficient increases as the image fusion quality optimization coefficient increases; it can be seen from the data in the sixth and seventh groups that the pitch angle coefficient increases as the image fusion quality optimization coefficient increases.

[0070] Through the above steps, the rationality of the hip prosthesis position is quantified, and thus the accuracy of the hip prosthesis position planning is improved.

[0071] The specific process of judging whether to give feedback based on the prosthesis planning evaluation value is as follows: judge whether the prosthesis planning evaluation value is less than the preset prosthesis planning threshold. If the prosthesis planning evaluation value is less than the preset prosthesis planning threshold, adjust the number of wavelet transform layers; otherwise, do not give feedback. Judge whether the prosthesis planning evaluation value obtained after adjusting the number of wavelet transform layers is less than the preset prosthesis planning threshold. If the prosthesis planning evaluation value is less than the preset prosthesis planning threshold, adjust the prosthesis position; otherwise, end the feedback. Judge whether the prosthesis planning evaluation value obtained after adjusting the prosthesis position is less than the preset prosthesis planning threshold. If the prosthesis planning evaluation value is less than the preset prosthesis planning threshold, give feedback to the preset personnel; otherwise, end the feedback.

[0072] In this embodiment, the preset prosthesis planning threshold is represented by the maximum value of the prosthesis planning evaluation value within the historical time period;

[0073] Adjusting the number of wavelet transform layers is achieved by increasing the number of wavelet transform layers. For example, if the initial state is to process the image using 3 layers of wavelet transform, then it is adjusted to 4 layers of wavelet transform.

[0074] Adjusting the prosthesis position is carried out by a random search method; the random search method means randomly generating a series of candidate positions around the current prosthesis position, calculating the prosthesis planning evaluation value of each candidate position, and selecting the position with the highest prosthesis planning evaluation value as the new prosthesis position.

[0075] Through the above steps, the accuracy of hip replacement during the hip replacement preoperative reconstruction process is improved.

[0076] An embodiment of the present invention provides a device for a hip replacement preoperative reconstruction system based on intelligent learning. The device includes: a data acquisition device, a storage device, and a processing device; wherein, the data acquisition device is used to collect the image fusion accuracy, image data, prosthesis position data, and image fusion quality data of the hip joint. The data acquisition device includes an X-ray machine, a computed tomography scanner, and a magnetic resonance imaging machine; the storage device is used to store the image fusion accuracy, image data, prosthesis position data, and image fusion quality data of the hip joint, and also includes the storage of corresponding historical data; the processing device is used to perform image quality assessment on the image data, perform image fusion quality assessment on the image fusion quality data, and perform prosthesis planning assessment on the prosthesis position data, the image fusion accuracy, the image quality assessment value obtained after image quality optimization, and the image fusion quality assessment value obtained after image fusion quality optimization.

[0077] In this embodiment, hip replacement surgery represents a surgical procedure used to treat hip joint diseases or injuries, and it restores joint function by replacing damaged parts of the hip joint (such as the femoral head and acetabulum); the image quality assessment module optimizes the image quality through intelligent learning, improving the image quality; the image fusion assessment module and the prosthesis planning module optimize the image fusion quality through intelligent learning, improving the image fusion quality; through the above steps, the effective optimization of the image quality and the optimization of multi-modal image fusion are achieved, and further the improvement of the relevance between the hip replacement preoperative reconstruction process and the image quality is realized.

[0078] In summary, in the embodiment of the present invention, it is judged whether to perform image quality optimization based on the image quality assessment value obtained from the image data, then it is judged whether to perform image fusion quality optimization based on the image fusion quality assessment value obtained from the image fusion quality data, and finally it is judged whether to give feedback based on the prosthesis planning assessment value obtained from the optimized image quality assessment value and the optimized image fusion quality assessment value, thereby optimizing the fusion of multi-modal images, and further realizing the improvement of the relevance between the hip replacement preoperative reconstruction process and the image quality, effectively solving the problem of low relevance between the hip replacement preoperative reconstruction process and the image quality in the prior art.

[0079] The following points need to be explained:

[0080] (1) The drawings in the embodiment of the present invention only relate to the structures involved in the embodiment of the present invention, and other structures can refer to the general design.

[0081] (2) For clarity, in the drawings used to describe the embodiments of the present invention, the thickness of layers or regions is enlarged or reduced, that is, these drawings are not drawn to actual scale. It can be understood that when an element such as a layer, film, region, or substrate is referred to as being "on" or "under" another element, the element can be "directly" on or under the other element or there can be intervening elements.

[0082] (3) Without conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other to obtain new embodiments.

[0083] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. The protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A hip replacement preoperative reconstruction system based on intelligent learning, characterized in that: include: Image quality assessment module, image fusion assessment module and prosthesis planning module; The image quality assessment module is used to evaluate the acquired image data to obtain an image quality assessment value, and determine whether to perform image quality optimization based on the image quality assessment value, and the image quality assessment value is used to evaluate the image quality; The image fusion evaluation module is used to evaluate the acquired image fusion quality data to obtain an image fusion quality evaluation value, and determine whether to perform image fusion quality optimization based on the image fusion quality evaluation value, and the image fusion quality evaluation value is used to evaluate the image fusion quality; The prosthesis planning module is used to obtain a prosthesis planning evaluation value based on the prosthesis position data, image fusion accuracy, image quality evaluation value obtained after image quality optimization, and image fusion quality evaluation value obtained after image fusion quality optimization if image optimization is performed, and to determine whether to provide feedback based on the prosthesis planning evaluation value. The image optimization includes image quality optimization and image fusion quality optimization, and the image fusion quality evaluation value is used to evaluate the rationality of the hip prosthesis position.

2. The hip replacement preoperative reconstruction system based on intelligent learning according to claim 1, characterized in that: The image data includes image signal-to-noise ratio, structural similarity index, information entropy and contrast; The image fusion quality data includes registration error, relative information gain, mutual information and fused image signal-to-noise ratio; The image fusion accuracy is obtained by processing the ratio of the pixel value of the fused image to the pixel value of the standard image; The prosthesis position data includes the prosthesis anteversion angle, the prosthesis insertion depth and the prosthesis pitch angle; The image signal-to-noise ratio includes fusion image signal-to-noise ratio, CT image signal-to-noise ratio, MRI image signal-to-noise ratio and X-ray image signal-to-noise ratio; The structural similarity index is obtained by performing a ratio operation on the pixel average value, pixel standard deviation and pixel square difference corresponding to the image and the pixel average value, pixel standard deviation and pixel square difference of the standard image, respectively. The structural similarity index includes a CT image structural similarity index, an MRI image structural similarity index and an X-ray image structural similarity index; The information entropy is obtained by processing the occurrence frequency of each image pixel value obtained by constructing the image histogram, and the information entropy includes fusion image information entropy, CT image information entropy, MRI image information entropy and X-ray image information entropy; The contrast is obtained by performing a ratio operation on the difference between the maximum image pixel value and the minimum image pixel value and the average image pixel value, and the contrast includes CT image contrast, MRI image contrast and X-ray image contrast; The relative information gain is obtained by calculating the ratio of the fusion image information entropy, the CT image information entropy, the MRI image information entropy and the X-ray image information entropy to the theoretical maximum entropy value in the database; The mutual information is obtained by processing the joint probability distribution and the marginal probability distribution, and the marginal probability distribution is obtained by performing a ratio operation between the frequency of occurrence of the pixel value of the corresponding image in the image and the total number of pixels of the corresponding image; The mutual information includes first mutual information, second mutual information, third mutual information, fourth mutual information, fifth mutual information and sixth mutual information, the first mutual information represents the mutual information between the CT image and the MRI image, the second mutual information represents the mutual information between the CT image and the X-ray image, the third mutual information represents the mutual information between the CT image and the fused image, the fourth mutual information represents the mutual information between the MRI image and the X-ray image, the fifth mutual information represents the mutual information between the MRI image and the fused image, and the sixth mutual information represents the mutual information between the X-ray image and the fused image.

3. The hip replacement preoperative reconstruction system based on intelligent learning according to claim 2, characterized in that: The specific method of evaluating the image quality evaluation value according to the acquired image data is as follows: The signal-to-noise ratio deviation is obtained according to the relative relationship between the signal-to-noise ratio and the corresponding preset signal-to-noise ratio in the database; The information entropy deviation is obtained according to the relative relationship between the information entropy and the corresponding preset information entropy in the database; Obtaining a contrast deviation according to a relative relationship between the contrast and a corresponding preset contrast in a database; The signal-to-noise ratio deviation, information entropy deviation, contrast deviation and structural similarity index are processed to obtain the image quality assessment value.

4. The hip replacement preoperative reconstruction system based on intelligent learning according to claim 3, characterized in that: The specific limiting expression of the image quality evaluation value is: Where i represents the image type number, i=1 represents CT image, i=2 represents MRI image, i=3 represents X-ray image, YXZ i Represents the image quality assessment value of the i-th image type, SNR i Represents the signal-to-noise ratio of the i-th image type, SSIM i Represents the structural similarity index of the i-th image type, XXS i represents the information entropy of the i-th image type, DBD i Represents the contrast of the i-th image type, SNR i.0 Represents the preset signal-to-noise ratio of the i-th image type, XXS i.0 represents the preset information entropy of the i-th image type, DBD i.0 Indicates the preset contrast of the i-th image type.

5. The intelligent learning-based preoperative reconstruction system for hip replacement according to claim 4, characterized in that: The specific process of determining whether to optimize the image quality based on the image quality evaluation value is as follows: A1, determining whether the image quality evaluation value is less than a preset image quality threshold. If the image quality evaluation value is less than the preset image quality threshold, executing A2, otherwise, not performing image quality optimization; A2, perform denoising, and determine whether the image quality evaluation value obtained after denoising is less than a preset image quality threshold. If the image quality evaluation value is less than the preset image quality threshold, execute A3, otherwise, end the image quality optimization; A3, performing contrast enhancement, determining whether the image quality evaluation value obtained after the contrast enhancement is less than a preset image quality threshold, if the image quality evaluation value is less than the preset image quality threshold, executing A4, otherwise, ending the image quality optimization; A4, performing image reconstruction and restoration, determining whether the image quality evaluation value obtained after the image reconstruction and restoration is less than a preset image quality threshold, if the image quality evaluation value is less than the preset image quality threshold, executing A5, otherwise, ending the image quality optimization; A5, perform structural optimization, and determine whether the image quality evaluation value obtained after the structural optimization is less than the preset image quality threshold. If the image quality evaluation value is less than the preset image quality threshold, feedback is given; otherwise, the image quality optimization is terminated.

6. The hip replacement preoperative reconstruction system based on intelligent learning according to claim 2, characterized in that: The specific method for evaluating the image fusion quality evaluation value according to the acquired image fusion quality data is as follows: Obtaining a registration error deviation according to a relative relationship between the registration error and a corresponding preset registration error in a database; The mutual information deviation is obtained according to the relative relationship between the total mutual information and the corresponding preset mutual information in the database; The fused image signal-noise ratio deviation is obtained according to the relative relationship between the fused image signal-noise ratio and the corresponding preset fused image signal-noise ratio in the database; The signal-to-noise ratio deviation, relative information gain, mutual information deviation and fusion image signal-to-noise ratio deviation are processed to obtain the image fusion quality evaluation value.

7. The intelligent learning-based preoperative reconstruction system for hip replacement according to claim 6, characterized in that: The specific process of judging whether to optimize the image fusion quality based on the image fusion quality evaluation value is as follows: Determine whether the image fusion quality evaluation value is less than the preset image fusion quality threshold. If the image fusion quality evaluation value is less than the preset image fusion quality threshold, the image quality of the fused image is optimized; otherwise, the image fusion quality is not optimized. It is determined whether the image fusion quality evaluation value obtained after image quality optimization is less than the preset image fusion quality threshold. If the image fusion quality evaluation value is less than the preset image fusion quality threshold, wavelet transform optimization is performed, otherwise the image fusion quality optimization is terminated.

8. The hip replacement preoperative reconstruction system based on intelligent learning according to claim 2, characterized in that: The specific method for evaluating the prosthesis planning evaluation value according to the prosthesis position data, the image fusion accuracy, the image quality evaluation value obtained after the image quality is optimized, and the image fusion quality evaluation value obtained after the image fusion quality is optimized is as follows: The image fusion accuracy deviation is obtained according to the relative relationship between the image fusion accuracy and the corresponding preset image fusion accuracy in the database, and the image fusion accuracy deviation is obtained by processing the ratio operation of the image fusion accuracy and the corresponding preset image fusion accuracy in the database; Obtaining an image quality optimization coefficient according to a relative relationship between an image quality evaluation value obtained after image quality optimization, an image quality evaluation value obtained before image quality optimization, and a preset image quality threshold; The image fusion quality optimization coefficient is obtained according to the relative relationship between the image fusion quality evaluation value obtained after the image fusion quality optimization, the image fusion quality evaluation value obtained before the image fusion quality optimization, and the preset image fusion quality threshold; If the prosthesis anteversion angle is not within the preset prosthesis anteversion angle range, the anteversion angle coefficient is obtained according to the relative relationship between the prosthesis anteversion angle and the preset prosthesis anteversion angle, otherwise the anteversion angle coefficient is recorded as 1; If the prosthesis insertion depth is not within the preset prosthesis insertion depth range, the depth coefficient is obtained according to the relative relationship between the prosthesis insertion depth and the preset prosthesis insertion depth, otherwise the depth coefficient is recorded as 1; If the prosthesis pitch angle is not within the preset prosthesis pitch angle range, the pitch angle coefficient is obtained according to the relative relationship between the prosthesis pitch angle and the preset prosthesis pitch angle, otherwise the pitch angle coefficient is recorded as 1; The anteversion angle coefficient, depth coefficient, pitch angle coefficient, image fusion accuracy deviation, image quality optimization coefficient and image fusion quality optimization coefficient are processed to obtain the prosthesis planning evaluation value.

9. The hip replacement preoperative reconstruction system based on intelligent learning according to claim 1, characterized in that: The specific process of judging whether to provide feedback based on the prosthesis planning evaluation value is as follows: Determine whether the prosthesis planning evaluation value is less than a preset prosthesis planning threshold, if the prosthesis planning evaluation value is less than the preset prosthesis planning threshold, adjust the number of wavelet transform layers, otherwise no feedback is performed; Determine whether the prosthesis planning evaluation value obtained after adjusting the number of wavelet transform layers is less than a preset prosthesis planning threshold. If the prosthesis planning evaluation value is less than the preset prosthesis planning threshold, adjust the prosthesis position, otherwise end the feedback; Determine whether the prosthesis planning evaluation value obtained after adjusting the prosthesis position is less than the preset prosthesis planning threshold. If the prosthesis planning evaluation value is less than the preset prosthesis planning threshold, feedback is given to the preset personnel, otherwise the feedback is terminated.

10. A device for a hip replacement preoperative reconstruction system based on intelligent learning as claimed in any one of claims 1 to 9, characterized in that: include: Data acquisition equipment, storage equipment and processing equipment; The data acquisition equipment is used to collect image fusion accuracy, image data, prosthesis position data and image fusion quality data of the hip joint, and the data acquisition equipment includes an X-ray machine, a computer tomography machine, and a magnetic resonance imaging machine; The storage device is used to store image fusion accuracy, image data, prosthesis position data and image fusion quality data of the hip joint; The processing device is used to perform image evaluation based on the data in the storage device, and the impact evaluation includes image quality evaluation, image fusion quality evaluation and prosthesis planning evaluation.

Images