Method, apparatus, device and storage medium for evaluating quality of pelvic anteroposterior X-ray film

Abstract

The present invention discloses a method for evaluating the quality of pelvic anteroposterior X-ray films, including: constructing a data set based on the annotation data of pelvic anteroposterior X-ray film images to train the Mask R-CNN network model; sending the pelvic anteroposterior X-ray film image to be evaluated into the trained Mask R-CNN network for processing; predicting the positions of bony key points according to the regions near the bony key points and their bounding boxes obtained, and calculating the slopes of the H line and the P line; calculating the transverse diameter and longitudinal diameter of the obturator foramen based on the slopes of the H line and the P line, and calculating the pelvic rotation index and pelvic tilt index; evaluating the quality of the pelvic anteroposterior X-ray film to be evaluated according to the pelvic rotation index and pelvic tilt index. The present invention can perform individualized and accurate measurement of the pelvic rotation index and pelvic tilt index for different obturator foramen morphologies of children, can quickly and accurately obtain the quality control indexes of pelvic anteroposterior X-ray films, and realize the automatic judgment of whether the pelvic anteroposterior X-ray films are standard or not.

Description

Technical Field

[0001] The present invention belongs to the technical field of computer image processing, and particularly relates to a method, device, equipment and storage medium for evaluating the quality of pelvic anteroposterior X-ray films. Background Technique

[0002] The pelvic anteroposterior X-ray film is the preferred imaging examination for developmental dysplasia of the hip (DDH) in children aged 4 to 6 months and above. Many studies have confirmed that non-standard imaging positions (i.e., pelvic tilt or rotation during pelvic X-ray examination) have a great impact on the measurement of hip joint parameters such as the acetabular index (AI) and the center-edge angle (CEA), thus affecting the correct diagnosis of subsequent pediatric DDH. Due to the high requirements for the standard imaging position of the pelvic anteroposterior X-ray film and the poor cooperation of children, "non-standard" pelvic anteroposterior X-ray films are not uncommon in clinical practice. Therefore, before diagnosis, it is first necessary to verify the quality of the pelvic anteroposterior X-ray film, that is, to judge whether the pelvic position is standard.

[0003] Currently, the evaluation of the quality of pelvic anteroposterior X-ray films only has the method of manual determination (i.e., clinical doctor's manual measurement). It has relatively high professional requirements in clinical use. It not only requires marking key points, drawing several reference lines, estimating the longitudinal and transverse diameters of the obturator foramen, but also needs to calculate the pelvic tilt index (pelvic tilt index of Ball and Kommenda, PTI) and the pelvic rotation index (pelvic rotation index of and Brunken, PRI) based on the actual measurement values, and make a final judgment by comparing with the normal reference value range.

[0004] However, the above method is not only time-consuming and laborious, but also has large measurement errors. In addition, due to the particularity of children's skeletal development, the obturator morphology of children of different ages varies greatly, so there is a lack of a unified obturator morphology, which may affect the measurement of the transverse / longitudinal diameter of the obturator foramen, thus affecting the quality determination of pelvic anteroposterior X-ray films. Summary of the Invention

[0005] In order to solve the above problems existing in the prior art, the present invention provides a method, device, equipment and storage medium for evaluating the quality of pelvic anteroposterior X-ray films. The technical problems to be solved by the present invention are realized through the following technical solutions:

[0006] In the first aspect, the present invention provides a method for evaluating the quality of pelvic anteroposterior X-ray films, including:

[0007] Construct a dataset based on the annotation data of pelvic anteroposterior X-ray images, and use the dataset to train the Mask R-CNN network model to obtain the trained Mask R-CNN network;

[0008] Send the pelvic anteroposterior X-ray image to be evaluated into the trained Mask R-CNN network for processing to obtain the regions near the bony key points, the obturator region and its bounding box; wherein, the bony key points include the Y points at the vertices of the Y-shaped cartilages of the two acetabula, the lateral points of the symphysis pubis, and the end points at the openings of the two obturators;

[0009] Predict the positions of the bony key points according to the regions near the bony key points and their bounding boxes, and calculate the slopes of the H line and the P line according to the predicted positions of the Y points at the vertices of the Y-shaped cartilages of the two acetabula;

[0010] Convert the obturator region into a binary image, and calculate the transverse diameter and longitudinal diameter of the obturator based on the slopes of the H line and the P line and in combination with the predicted positions of the end points at the openings of the two obturators;

[0011] Calculate the pelvic rotation index according to the transverse diameter of the obturator, and calculate the pelvic tilt index according to the longitudinal diameter of the obturator;

[0012] Evaluate the quality of the pelvic anteroposterior X-ray image to be evaluated according to the pelvic rotation index and the pelvic tilt index.

[0013] In an embodiment of the present invention, constructing a dataset based on the annotation data of pelvic anteroposterior X-ray images, and using the dataset to train the Mask R-CNN network model to obtain the trained Mask R-CNN network includes:

[0014] Perform bony key point annotation on the pelvic anteroposterior X-ray image;

[0015] Perform the following preprocessing on the data of the pelvic anteroposterior X-ray image after annotation: taking the bony key point as the center, generate a region of a certain size as a category for instance segmentation;

[0016] Divide the dataset obtained after preprocessing into a training set and a validation set;

[0017] Send the training set and the validation set into the Mask R-CNN network model for training to determine the weights of the network model and obtain the trained Mask R-CNN network.

[0018] In an embodiment of the present invention, predicting the positions of the bony key points according to the regions near the bony key points and their bounding boxes, and calculating the slopes of the H line and the P line according to the predicted positions of the Y points at the vertices of the Y-shaped cartilages of the two acetabula includes:

[0019] Take the center point of the bounding box of the area near the bone key point as the predicted position of the bone key point;

[0020] Calculate the slope of the H line according to the predicted positions of the Y points at the vertices of the Y-shaped cartilage of the acetabula on both sides;

[0021] Take the negative reciprocal of the slope of the H line as the slope of the P line.

[0022] In one embodiment of the present invention, converting the obturator region into a binary image, and calculating the transverse diameter and longitudinal diameter of the obturator based on the slopes of the H line and the P line and in combination with the predicted positions of the end points at the openings of the obturators on both sides includes:

[0023] Convert the obturator region into a binary image, and obtain the contour information of the obturators on both sides therefrom to obtain two first obturator contours;

[0024] Delete the contour information at the openings of the obturators on the two first obturator contours respectively according to the predicted positions of the end points at the openings of the obturators on both sides to correspondingly obtain two second obturator contours;

[0025] Take one side of the obturator as a reference, and obtain the minimum circumscribed rectangle of the second obturator contour corresponding to this side of the obturator;

[0026] Traverse the points on the left side of the minimum circumscribed rectangle in turn, draw a straight line H' from the points on the left side of the minimum circumscribed rectangle to the opposite side, the slope of which is equal to the slope of the H line, and record the length of the diameter line intercepted by the second obturator contour by the current H' line;

[0027] Among the diameter lines obtained by traversing, take the maximum value as the transverse diameter of the selected reference obturator; at this time, the length of the diameter line intercepted by the second obturator contour corresponding to the other side of the obturator by the H' line is used as the transverse diameter of the other side of the obturator.

[0028] In one embodiment of the present invention, converting the obturator region into a binary image, and calculating the transverse diameter and longitudinal diameter of the obturator based on the slopes of the H line and the P line and in combination with the predicted positions of the end points at the openings of the obturators on both sides further includes:

[0029] Convert the obturator region into a binary image, and obtain the contour information of the obturators on both sides therefrom to obtain two first obturator contours;

[0030] Delete the contour information at the openings of the obturators on the two first obturator contours respectively according to the predicted positions of the end points at the openings of the obturators on both sides to correspondingly obtain two second obturator contours;

[0031] Take one side of the obturator as a reference, and obtain the minimum circumscribed rectangle of the second obturator contour corresponding to this side of the obturator;

[0032] Traverse the points on the upper side of the minimum circumscribed rectangle of the second obturator contour corresponding to this side of the obturator in sequence. Draw a straight line P' from the points on the upper side of the minimum circumscribed rectangle to the opposite side, and its slope is equal to the slope of the P line. Record the vertical line length intercepted by the current P' line by the second obturator contour.

[0033] Among the vertical lines obtained by traversing, take the maximum value as the vertical diameter of the selected reference obturator, denoted as Dv. In an embodiment of the present invention, the calculation formula of the pelvic rotation index is:

[0034] PRI = Rt / Lt

[0035] where PRI represents the pelvic rotation index, Rt represents the transverse diameter of the right obturator, and Lt represents the transverse diameter of the left obturator;

[0036] The calculation formula of the pelvic tilt index is:

[0037] PTI = Dv / h

[0038] where PTI represents the pelvic tilt index, Dv represents the vertical diameter of the selected reference obturator, and h represents the distance from the outer point of the pubic symphysis on one side of the reference obturator to the H line.

[0039] In an embodiment of the present invention, the quality assessment of the pelvic anteroposterior X-ray film to be evaluated according to the pelvic rotation index and the pelvic tilt index includes:

[0040] Compare the pelvic rotation index and the pelvic tilt index with the corresponding normal reference value ranges. Only when both of them meet the normal reference value ranges, it is determined that the pelvic anteroposterior X-ray is a standard pelvic anteroposterior X-ray film, otherwise it is a "non-standard" pelvic anteroposterior X-ray film.

[0041] In a second aspect, the present invention provides a device for quality assessment of pelvic anteroposterior X-ray films, including:

[0042] A network training module, used to construct a data set based on the annotation data of pelvic anteroposterior X-ray film images, and use the data set to train the Mask R-CNN network model to obtain a trained Mask R-CNN network;

[0043] A data inference module, used to send the pelvic anteroposterior X-ray film image to be evaluated into the trained Mask R-CNN network for processing to obtain the area near the bony key points, the obturator area and their bounding boxes; wherein, the bony key points include the Y points at the vertices of the Y-shaped cartilages of the two acetabula, the vertex of the pubic symphysis, and the end points at the openings of the two obturators;

[0044] A first calculation module, configured to predict the positions of bone key points based on the regions near the bone key points and their bounding boxes, and calculate the slopes of the H line and the P line according to the predicted positions of the Y points at the vertices of the Y-shaped cartilages on both sides of the acetabulum;

[0045] A second calculation module, configured to convert the obturator region into a binary image, and calculate the transverse diameter and longitudinal diameter of the obturator based on the slopes of the H line and the P line, in combination with the predicted positions of the end points at the openings of the obturators on both sides;

[0046] A third calculation module, configured to calculate a pelvic rotation index and a pelvic tilt index based on the transverse diameter and longitudinal diameter of the obturator;

[0047] An evaluation module, configured to perform quality evaluation on the pelvic anteroposterior X-ray film to be evaluated according to the pelvic rotation index and the pelvic tilt index.

[0048] In a third aspect, the present invention provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory complete communication with each other through the communication bus;

[0049] The memory is used for storing a computer program;

[0050] The processor is configured to implement the method steps described in the above embodiments when executing the program stored on the memory.

[0051] In a fourth aspect, the present invention provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, and the computer program realizes the method steps described in the above embodiments when executed by a processor.

[0052] Advantages of the present invention:

[0053] 1. By training an instance segmentation network model Mask R-CNN to extract bone key points and segment the obturator region, the present invention can quickly and accurately obtain relevant indicators for determining whether a pelvic anteroposterior X-ray film is standard, and can perform individualized and accurate measurement of the pelvic rotation index and pelvic tilt index for different obturator morphologies of children, efficiently and accurately realizing automatic analysis and evaluation of the quality of pelvic anteroposterior X-ray films;

[0054] 2. The present invention applies an automated method to the quality evaluation of pelvic anteroposterior X-ray films for the first time. Compared with the manual analysis method, the present invention can output analysis results efficiently and accurately;

[0055] 3. For the special obturator morphology where the obturators of some young children are not yet closed, the present invention designs a calculation method for the transverse diameter and longitudinal diameter of the obturator, which can be widely applicable to the obturator morphologies of children of different ages.

[0056] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Description of the Drawings

[0057] Figure 1 is a schematic flowchart of a method for evaluating the quality of a pelvic anteroposterior X-ray film provided by an embodiment of the present invention;

[0058] Figure 2 is a schematic diagram of key point annotation of a pelvic anteroposterior X-ray film provided by an embodiment of the present invention;

[0059] Figure 3 is a schematic diagram of calculating the transverse / longitudinal diameter of the obturator foramen based on the right obturator foramen provided by an embodiment of the present invention;

[0060] Figure 4 is an analysis result diagram of a pelvic anteroposterior X-ray film provided by an embodiment of the present invention;

[0061] Figure 5 is a schematic structural diagram of a device for evaluating the quality of a pelvic anteroposterior X-ray film provided by an embodiment of the present invention;

[0062] Figure 6 is a schematic structural diagram of an electronic device for evaluating the quality of a pelvic anteroposterior X-ray film provided by an embodiment of the present invention;

[0063] Figure 7 is a schematic structural diagram of a computer-readable storage medium provided by an embodiment of the present invention. Detailed Embodiments

[0064] The present invention will be further described in detail below with reference to specific embodiments, but the embodiments of the present invention are not limited thereto.

[0065] Embodiment 1

[0066] Please refer to Figure 1 , Figure 1 which is a schematic flowchart of a method for evaluating the quality of a pelvic anteroposterior X-ray film provided by an embodiment of the present invention, and includes:

[0067] Step 1: Construct a data set based on the annotation data of the pelvic anteroposterior X-ray film image, and use the data set to train the Mask R-CNN network model to obtain a trained Mask R-CNN network.

[0068] 11) Perform bone key point annotation on the pelvic anteroposterior X-ray film image.

[0069] Specifically, in this embodiment, the coordinates of the bone key points and the left / right obturator foramen regions can be marked by manual annotation method, such as in the attached Figure 2As shown in the figure. Among them, the bony key points include the Y points at the vertices of the Y-shaped cartilages of the two acetabula and the lateral points of the symphysis pubis; for the case where the obturator foramen of some young children is not closed, the endpoints at the openings of the two obturator foramina also need to be marked.

[0070] 12) Perform the following preprocessing on the labeled pelvic anteroposterior X-ray image data: Take the bony key points as the center and generate a region of a certain size as a category for instance segmentation.

[0071] Specifically, in this embodiment, a square region with a size of 60*60 pixels centered on the bony key points can be generated through code as a category for instance segmentation.

[0072] 13) Divide the dataset obtained after preprocessing into a training set and a validation set;

[0073] 14) Feed the training set and the validation set into the Mask R-CNN network model for training to determine the weights of the network model and obtain the trained Mask R-CNN network.

[0074] Optionally, as an implementation method, this embodiment selects a standard Mask R-CNN network structure for training, and its specific training process can refer to the existing relevant technologies and will not be introduced in detail here.

[0075] Step 2: Feed the pelvic anteroposterior X-ray image to be evaluated into the trained Mask R-CNN network for processing to obtain the region near the bony key points, the obturator foramen region and their bounding boxes.

[0076] The present invention first uses the method of deep learning to extract the obturator foramen region, which not only avoids the time-consuming and laborious problem of the traditional manual measurement method, but also reduces the measurement error.

[0077] Step 3: Predict the positions of the bony key points according to the region near the bony key points and their bounding boxes, and calculate the slopes of the H line and the P line according to the predicted positions of the Y points at the vertices of the Y-shaped cartilages of the two acetabula.

[0078] 31) Take the center point of the bounding box of the region near the bony key points as the predicted position of the bony key points;

[0079] 32) Calculate the slope of the H line according to the predicted positions of the Y points at the vertices of the Y-shaped cartilages of the two acetabula;

[0080] 33) Take the negative reciprocal of the slope of the H line as the slope of the P line.

[0081] Specifically, the H line, also known as the Hilgenreiner line, refers to the line connecting the Y points at the vertices of the Y-shaped cartilages of the two acetabula on the pelvic anteroposterior X-ray image, so its slope can be calculated through the positions of the Y points at the vertices of the Y-shaped cartilages of the two acetabula.

[0082] The P line is the perpendicular line of the H line drawn through point P on the outer edge of the acetabulum. Then, the slope of the P line is the negative reciprocal of the slope of the H line.

[0083] Step 4: Convert the obturator region into a binary image. Based on the slopes of the H line and the P line, and combined with the predicted endpoint positions at the openings of both sides of the obturator, calculate the transverse diameter and longitudinal diameter of the obturator.

[0084] Step 5: Calculate the pelvic rotation index according to the transverse diameter of the obturator, and calculate the pelvic tilt index according to the longitudinal diameter of the obturator.

[0085] Specifically, for the calculation process of the transverse diameter of the obturator, it can be realized according to the following steps:

[0086] 41) Convert the obturator region into a binary image, and obtain the contour information of both sides of the obturator from it to get two first obturator contours.

[0087] 42) According to the predicted endpoint positions at the openings of both sides of the obturator, delete the contour information at the openings of the obturator on the two first obturator contours respectively, and correspondingly obtain two second obturator contours.

[0088] 43) Take one side of the obturator as a reference, and obtain the minimum circumscribed rectangle of the second obturator contour corresponding to this side of the obturator.

[0089] 44) Traverse the points on the left side of the minimum circumscribed rectangle in turn. Draw a straight line H' from the points on the left side of the minimum circumscribed rectangle to the opposite side, and its slope is equal to the slope of the H line. Record the length of the diameter intercepted by the second obturator contour by the current H' line.

[0090] It should be noted that when the slope of the H line is greater than 0, start traversing from the lower left corner vertex to the upper left corner vertex of the minimum circumscribed rectangle; when the slope of the H line is less than 0, start traversing from the upper left corner vertex to the lower left corner vertex of the minimum circumscribed rectangle.

[0091] 45) Among the diameters obtained by traversing, take the maximum value as the transverse diameter of the selected reference obturator; at this time, the length of the diameter intercepted by the second obturator contour corresponding to the other side of the obturator by the H' line is used as the transverse diameter of the other side of the obturator.

[0092] Then, in step 5, the pelvic rotation index can be calculated according to the transverse diameter of the obturator according to the following formula:

[0093] PRI = Rt / Lt

[0094] Where, PRI represents the pelvic rotation index, Rt represents the transverse diameter of the right obturator, and Lt represents the transverse diameter of the left obturator.

[0095] Furthermore, for the calculation of the longitudinal diameter of the obturator, it can be carried out according to the following steps:

[0096] 4a) Convert the closed-hole region into a binary image, and obtain the contour information of the two closed holes on both sides from it to get two first closed-hole contours;

[0097] 4b) Delete the contour information at the openings of the closed holes on the two first closed-hole contours respectively according to the predicted endpoint positions at the openings of the closed holes on both sides, and correspondingly obtain two second closed-hole contours;

[0098] 4c) Take one side of the closed hole as a reference, and obtain the minimum circumscribed rectangle of the second closed-hole contour corresponding to this side of the closed hole;

[0099] 4d) Traverse the points on the upper side of the minimum circumscribed rectangle of the second closed-hole contour corresponding to this side of the closed hole in sequence. Make a straight line P' from the points on the upper side of the minimum circumscribed rectangle to the opposite side, and its slope is equal to the slope of the P line. Record the vertical line length intercepted by the current P' line by this second closed-hole contour;

[0100] 4e) Among the vertical lines obtained by traversing, take the maximum value as the longitudinal diameter of the selected reference closed hole, denoted as Dv. Then in step 5, the pelvic tilt index can be calculated according to the longitudinal diameter of the closed hole according to the following formula:

[0101] PTI = Dv / h

[0102] Wherein, PTI represents the pelvic tilt index, Dv represents the longitudinal diameter of the reference closed hole, and h represents the distance from the outer point of the pubic symphysis on one side of the reference closed hole to the H line.

[0103] It should be noted that since the pelvis is divided into left and right sides, and the change in the pelvis position has different effects on both sides, when calculating the pelvic rotation index and the pelvic tilt index, it is necessary to be based on the longitudinal diameter of the closed hole / the distance from the outer point of the pubic symphysis to the H line on the same side.

[0104] Furthermore, for the convenience of calculation, the closed hole on the same side can be selected as a reference to calculate the pelvic rotation index and the pelvic tilt index.

[0105] For example, please refer to Figure 3 , Figure 3 is a schematic diagram of calculating the transverse / longitudinal diameter of the closed hole based on the right closed hole provided by the embodiment of the present invention. It should be noted that on the X-ray film, the left position in the figure corresponds to the right side in reality, and the right position in the figure corresponds to the left side in reality. Then specifically, obtain Figure 3 the minimum circumscribed rectangle of the closed-hole contour at the left position (the right side in reality) in, and traverse the points on the left side of the minimum circumscribed rectangle in sequence. Make a straight line H' from the points on the left side of the minimum circumscribed rectangle to the opposite side, and its slope is equal to the slope of the H line. Record the radial line length intercepted by the current H' line by this side of the closed hole. Among the radial lines obtained by traversing, take the maximum value as the transverse diameter of this side of the closed hole, that is, the right closed hole in reality, denoted as Rt. At this time, the H' line is intercepted by Figure 3The length of the diameter intercepted on the right side of the obturator foramen is used as the transverse diameter of the left obturator foramen in practice, denoted as Lt. Further combining with the formula PRI = Rt / Lt, the pelvic rotation index PRI can be calculated.

[0106] The calculation method of the longitudinal diameter of the obturator foramen is similar to that of the transverse diameter, and it is obtained by traversing Figure 3 On the left side in the middle, that is, the maximum diameter of the right obturator foramen in practice is used as the longitudinal diameter Dv of this side of the obturator foramen. The ratio of this maximum longitudinal diameter to the distance h from the lateral point of the symphysis pubis of this side to the H line is the pelvic tilt index PTI.

[0107] In view of the special obturator foramen morphology where the obturator foramen of some young children has not closed, the present invention designs the above-mentioned specific calculation methods for the transverse and longitudinal diameters of the obturator foramen, which can be widely applicable to the obturator foramen morphologies of children of different ages.

[0108] Step 6: Evaluate the quality of the pelvic anteroposterior X-ray film to be evaluated according to the pelvic rotation index and the pelvic tilt index.

[0109] Specifically, compare the pelvic rotation index and the pelvic tilt index with the corresponding normal reference value ranges. Only when both of them meet the normal reference value ranges, it is determined that the pelvic anteroposterior X-ray film is a standard pelvic anteroposterior X-ray film, otherwise it is a "non-standard" pelvic anteroposterior X-ray film.

[0110] It should be noted that for the normal reference value ranges, the standard of PRI = 0.56 - 1.80; PTI = 0.75 - 1.20 can be followed.

[0111] The present invention trains an instance segmentation network model Mask R-CNN to extract bony key points and segment the obturator foramen area, so as to quickly and accurately obtain the relevant indicators for determining whether a pelvic anteroposterior X-ray film is standard, and can perform individual precise measurement of the pelvic rotation index and the pelvic tilt index for different obturator foramen morphologies of children, and efficiently and accurately realizes the automatic analysis and evaluation of the quality of pelvic anteroposterior X-ray films.

[0112] To verify the beneficial effects of the present invention, in this embodiment, a pelvic anteroposterior X-ray film to be evaluated is also selected, and it is processed by the above-mentioned automated method provided by the present invention, and the output result is as Figure 4 shown. It can be seen that compared with the manual analysis method, the present invention can output the analysis result efficiently and accurately.

[0113] Embodiment 2

[0114] Based on the above Embodiment 1, this embodiment provides a device for evaluating the quality of pelvic anteroposterior X-ray films. Please refer to Figure 5 , Figure 5 which is a schematic structural diagram of a device for evaluating the quality of pelvic anteroposterior X-ray films provided by an embodiment of the present invention, and it includes:

[0115] A network training module 1, configured to construct a data set based on the labeled data of pelvic anteroposterior X-ray images, and use the data set to train a Mask R-CNN network model to obtain a trained Mask R-CNN network;

[0116] A data inference module 2, configured to send the pelvic anteroposterior X-ray image to be evaluated into the trained Mask R-CNN network for processing to obtain the area near the bony key points, the obturator area and their bounding boxes; wherein, the bony key points include the Y points at the vertices of the Y-shaped cartilages of the two acetabula, the vertex of the symphysis pubis, and the end points at the openings of the two obturator foramina;

[0117] A first calculation module 3, configured to predict the positions of the bony key points according to the area near the bony key points and their bounding boxes, and calculate the slopes of the H line and the P line according to the predicted positions of the Y points at the vertices of the Y-shaped cartilages of the two acetabula;

[0118] A second calculation module 4, configured to convert the obturator area into a binary image, and calculate the transverse diameter and longitudinal diameter of the obturator based on the slopes of the H line and the P line and in combination with the predicted positions of the end points at the openings of the two obturator foramina;

[0119] A third calculation module 5, configured to calculate a pelvic rotation index and a pelvic tilt index according to the transverse diameter and longitudinal diameter of the obturator;

[0120] An evaluation module 6, configured to perform quality evaluation on the pelvic anteroposterior X-ray image to be evaluated according to the pelvic rotation index and the pelvic tilt index.

[0121] The pelvic anteroposterior X-ray image quality evaluation device provided in this embodiment can be used to implement the method provided in the above-mentioned Embodiment 1, and the detailed process will not be elaborated here.

[0122] Therefore, the device provided in this embodiment also has the advantage of being able to efficiently and accurately realize the automatic analysis and evaluation of the quality of pelvic anteroposterior X-ray images.

[0123] Embodiment 3

[0124] Based on the above-mentioned Embodiment 1, this embodiment provides an electronic device for pelvic anteroposterior X-ray image quality evaluation. Please refer to Figure 6 , Figure 6 which is a schematic structural diagram of the electronic device for pelvic anteroposterior X-ray image quality evaluation provided by the embodiment of the present invention. It includes a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0125] The memory is used to store a computer program;

[0126] A processor, when executing a program stored in a memory, implements the method steps provided in the first embodiment above. The implementation principle and technical effects are similar and will not be elaborated here.

[0127] Embodiment 4

[0128] Based on the above-mentioned first embodiment, please refer to Figure 7 , Figure 7 which is a schematic structural diagram of a computer-readable storage medium provided by an embodiment of the present invention. A computer-readable storage medium provided in this embodiment stores a computer program. When the above computer program is executed by a processor, the method steps provided in the first embodiment above are implemented. The implementation principle and technical effects are similar and will not be elaborated here.

[0129] The above content is a further detailed description of the present invention in combination with specific preferred implementation manners. It cannot be determined that the specific implementation of the present invention is only limited to these descriptions. For those of ordinary skill in the technical field to which the present invention belongs, without departing from the concept of the present invention, several simple deductions or substitutions can still be made, and all should be regarded as belonging to the protection scope of the present invention.

Claims

1. A method for evaluating the quality of pelvic anteroposterior X-ray films, characterized in that, Including: Construct a dataset based on the annotation data of pelvic anteroposterior X-ray images, and use the dataset to train the Mask R-CNN network model to obtain the trained Mask R-CNN network; Send the pelvic anteroposterior X-ray image to be evaluated into the trained Mask R-CNN network for processing to obtain the regions near the osseous key points, the obturator region and its bounding boxes; wherein, the osseous key points include the Y points at the vertices of the Y-shaped cartilages of the acetabula on both sides, the lateral points of the symphysis pubis, and the end points at the openings of the obturators on both sides; Predict the positions of the osseous key points according to the regions near the osseous key points and their bounding boxes, and calculate the slopes of the H line and the P line according to the predicted positions of the Y points at the vertices of the Y-shaped cartilages of the acetabula on both sides; Convert the obturator region into a binary image, and calculate the transverse diameter and longitudinal diameter of the obturator based on the slopes of the H line and the P line and in combination with the predicted positions of the end points at the openings of the obturators on both sides; Calculate the pelvic rotation index according to the transverse diameter of the obturator, and calculate the pelvic tilt index according to the longitudinal diameter of the obturator; Perform quality assessment on the pelvic anteroposterior X-ray image to be evaluated according to the pelvic rotation index and the pelvic tilt index.

2. The quality assessment method of a pelvic anteroposterior X-ray film according to claim 1, wherein Construct a dataset based on the annotation data of pelvic anteroposterior X-ray images, and use the dataset to train the Mask R-CNN network model to obtain the trained Mask R-CNN network, including: Perform osseous key point annotation on the pelvic anteroposterior X-ray images; Perform the following preprocessing on the annotated pelvic anteroposterior X-ray image data: taking the osseous key points as the centers, generating regions of a certain size as a category for instance segmentation; Divide the dataset obtained after preprocessing into a training set and a validation set; Send the training set and the validation set into the Mask R-CNN network model for training to determine the weights of the network model and obtain the trained Mask R-CNN network.

3. The quality assessment method of a pelvic anteroposterior X-ray film according to claim 1, wherein, Predict the positions of the osseous key points according to the regions near the osseous key points and their bounding boxes, and calculate the slopes of the H line and the P line according to the predicted positions of the Y points at the vertices of the Y-shaped cartilages of the acetabula on both sides, including: Taking the center points of the bounding boxes of the regions near the osseous key points as the predicted positions of the osseous key points; Calculate the slope of the H line according to the predicted positions of the Y points at the vertices of the Y-shaped cartilages of the acetabula on both sides; Taking the negative reciprocal of the slope of the H line as the slope of the P line.

4. The method for evaluating the quality of a pelvic anteroposterior X-ray film according to claim 1, wherein, Convert the obturator region into a binary image, and calculate the transverse diameter and longitudinal diameter of the obturator based on the slopes of the H line and the P line and in combination with the predicted positions of the end points at the openings of the obturators on both sides, including: Convert the obturator region into a binary image, and obtain the contour information of the obturators on both sides from it to get two first obturator contours; Delete the contour information at the openings of the obturators on the two first obturator contours respectively according to the predicted positions of the end points at the openings of the obturators on both sides to correspondingly obtain two second obturator contours; Taking one side of the obturator as a reference, obtain the minimum bounding rectangle of the second obturator contour corresponding to this side of the obturator; Traverse the points on the left side of the minimum bounding rectangle in sequence, draw a straight line H' from the points on the left side of the minimum bounding rectangle to the opposite side, whose slope is equal to the slope of the H line, and record the radial line length intercepted by the current H' line by the second closed hole contour; Among the radial lines obtained by traversal, take the maximum value as the transverse diameter of the selected reference closed hole; at this time, the radial line length intercepted by the H' line by the second closed hole contour corresponding to the closed hole on the other side is used as the transverse diameter of the closed hole on the other side.

5. The method for evaluating the quality of a pelvic anteroposterior X-ray film according to claim 4, characterized in that, Convert the closed hole area into a binary image, and based on the slopes of the H line and the P line, and in combination with the predicted endpoint positions at the openings of the closed holes on both sides, calculate the transverse diameter and longitudinal diameter of the closed hole, further including: Convert the closed hole area into a binary image, and obtain the contour information of the closed holes on both sides from it, to get two first closed hole contours; According to the predicted endpoint positions at the openings of the closed holes on both sides, delete the contour information at the openings of the closed holes on the two first closed hole contours respectively, and correspondingly obtain two second closed hole contours; Take one side of the closed hole as a reference, and obtain the minimum bounding rectangle of the second closed hole contour corresponding to this side of the closed hole; Traverse the points on the upper side of the minimum bounding rectangle of the second closed hole contour corresponding to one side of the closed hole in sequence, draw a straight line P' from the points on the upper side of the minimum bounding rectangle to the opposite side, whose slope is equal to the slope of the P line, and record the longitudinal line length intercepted by the current P' line by the second closed hole contour; Among the longitudinal lines obtained by traversal, take the maximum value as the longitudinal diameter of the selected reference closed hole, denoted as Dv.

6. The method for evaluating the quality of a pelvic anteroposterior X-ray film according to claim 5, wherein, The calculation formula of the pelvic rotation index is: PRI = Rt / Lt Wherein, PRI represents the pelvic rotation index, Rt represents the transverse diameter of the right closed hole, and Lt represents the transverse diameter of the left closed hole; The calculation formula of the pelvic tilt index is: PTI = Dv / h Wherein, PTI represents the pelvic tilt index, Dv represents the longitudinal diameter of the selected reference closed hole, and h represents the distance from the outer point of the pubic symphysis on one side of the reference closed hole to the H line.

7. A method for evaluating the quality of a pelvic anteroposterior X-ray film according to claim 1, characterized in that, Perform quality assessment on the pelvic anteroposterior X-ray film to be evaluated according to the pelvic rotation index and the pelvic tilt index, including: Compare the pelvic rotation index and the pelvic tilt index with the corresponding normal reference value ranges. Only when both of them meet the normal reference value ranges at the same time, determine that the pelvic anteroposterior X-ray is a standard pelvic anteroposterior X-ray film, otherwise it is a "non-standard" pelvic anteroposterior X-ray film.

8. A quality assessment device for pelvic anteroposterior X-ray films, characterized in that, Including: A network training module (1), which is used to construct a data set based on the annotation data of pelvic anteroposterior X-ray film images, and use the data set to train the Mask R-CNN network model to obtain a trained Mask R-CNN network; A data inference module (2), which is used to send the pelvic anteroposterior X-ray film image to be evaluated into the trained Mask R-CNN network for processing to obtain the areas near the bony key points, the closed hole area and their bounding boxes; wherein, the bony key points include the Y points at the vertices of the Y-shaped cartilages of the two acetabula, the vertex of the pubic symphysis, and the endpoint positions at the openings of the two closed holes; A first calculation module (3), which is used to predict the positions of the bony key points according to the areas near the bony key points and their bounding boxes, and calculate the slopes of the H line and the P line according to the predicted positions of the Y points at the vertices of the Y-shaped cartilages of the two acetabula; A second calculation module (4) for converting the closed pore region into a binary image, calculating the transverse diameter and longitudinal diameter of the closed pore based on the slopes of the H line and the P line, and in combination with the predicted endpoint positions at the openings of the closed pores on both sides; A third calculation module (5) for calculating a pelvic rotation index and a pelvic tilt index according to the transverse diameter and the longitudinal diameter of the closed pore; An evaluation module (6) for performing quality evaluation on the pelvic anteroposterior X-ray film to be evaluated according to the pelvic rotation index and the pelvic tilt index.

9. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete mutual communication through the communication bus; The memory is used for storing a computer program; The processor is configured to implement the method steps described in any one of claims 1-7 when executing the program stored on the memory.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the method steps described in any one of claims 1-7 are implemented.

Images