Magnetic resonance imaging image-based shoulder cyst positioning method and system
Through deep learning algorithm detection model and spatial positioning algorithm, the problems of MRI image quality decline and doctor dependence are solved, efficient and automated positioning of shoulder cysts are achieved, and diagnostic accuracy is improved.
Patent Information
- Application Number
- CN202411079276.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-07
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2044-08-07
AI Technical Summary
MRI images are susceptible to spot noise and echo perturbations in shoulder cyst detection, resulting in a decrease in image quality. It is difficult for doctors to distinguish between effusion caused by humeral head cyst and rotator cuff injury, and it is highly dependent on doctors' cognitive ability and clinical experience. .
The humeral head area object detection model and cyst object detection model based on deep learning algorithm are adopted to reduce noise interference, improve image quality through the input and output of multiple MRI images, and the spatial positioning algorithm is used to automatically determine the location of the cyst.
The MRI image quality is improved, the cyst is significantly differentiated from normal tissues, the dependence on doctors' cognitive ability is reduced, and the automated positioning is achieved, avoiding the confusion between humeral head cyst and rotator cuff injury effusion.
Smart Images

Figure CN119048593A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of medical image processing, and in particular to a shoulder cyst positioning method and system based on magnetic resonance imaging images. Background Art
[0002] With the advancement of modern medical imaging technology, the detection and diagnosis of shoulder diseases have gradually improved. Humeral head cysts are a common and frequent shoulder disease in clinical practice, which is closely related to aging and rotator cuff lesions. According to statistics from relevant associations, the incidence of shoulder diseases has risen sharply in the past decade, among which humeral head cysts account for 15%-45% of the entire shoulder diseases, ranking in the top three among all shoulder diseases. Humeral head cysts can cause pain in the shoulder joint area and in severe cases lead to limited shoulder joint movement, making it difficult for patients to perform normal exercises. Therefore, the detection and treatment of humeral head cysts are particularly important.
[0003] Among the current technologies, magnetic resonance imaging (MRI) has greater advantages than CT (computed tomography). MRI can generate high-quality images in any plane (such as cross-section, coronal plane, sagittal plane), providing more comprehensive anatomical information. In addition, MRI is more sensitive in the detection of tumors and soft tissues, and displays shoulder structures more clearly. The advantages of MRI are particularly obvious in the detection of humeral head cysts.
[0004] However, clinical evaluation of MRI-based shoulder cyst images remains challenging. On the one hand, MRI images are affected by speckle noise and echo disturbance, which affects the image quality and makes the difference between cysts and normal tissues less obvious. On the other hand, doctors are prone to confuse humeral head cysts with effusions caused by rotator cuff injuries during diagnosis, which is highly dependent on doctors' cognitive ability and clinical experience. Summary of the invention
[0005] In order to solve the technical problem that traditional MRI images are affected by speckle noise and echo disturbance, thereby affecting the image quality, making the difference between cysts and normal tissues unclear, and doctors are prone to confuse humeral head cysts with effusions caused by rotator cuff injuries during the diagnosis process, which is highly dependent on the doctor's cognitive ability and clinical experience, the present invention provides a shoulder cyst positioning method and system based on magnetic resonance imaging images.
[0006] The technical solution provided by the embodiment of the present invention is as follows:
[0007] First aspect:
[0008] An embodiment of the present invention provides a method for locating a shoulder cyst based on magnetic resonance imaging images, comprising:
[0009] S1: Acquire multiple magnetic resonance imaging images;
[0010] S2: Build a humeral head region target detection model and a cyst target detection model based on deep learning algorithm respectively;
[0011] S3: inputting the magnetic resonance imaging image into the humeral head region target detection model for detection, and outputting a humeral head region image;
[0012] S4: inputting the humeral head region image into the cyst target detection model for detection, and outputting the cyst region image;
[0013] S5: Determine the position of the cyst according to the coordinates of the center point of the cyst area image by using a spatial positioning algorithm.
[0014] Second aspect:
[0015] An embodiment of the present invention provides a shoulder cyst localization system based on magnetic resonance imaging images, comprising: a memory and one or more processors;
[0016] One or more application programs are stored in the memory, and the one or more application programs are suitable for being executed by the one or more processors to implement the above-mentioned shoulder cyst positioning method based on magnetic resonance imaging images.
[0017] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:
[0018] In the present invention, by inputting the magnetic resonance imaging image into the humeral head area target detection model for detection, outputting the humeral head area image, and then inputting the humeral head area image into the cyst target detection model for detection, and outputting the cyst area image, the influence of speckle noise and echo disturbance on the MRI image is reduced, the image quality is improved, and the difference between the cyst and normal tissue becomes more obvious. The position of the cyst is determined according to the center point coordinates of the cyst area image through a spatial positioning algorithm, reducing the dependence on the doctor's cognitive ability and clinical experience, realizing automatic positioning, and avoiding confusion between the humeral head cyst and the effusion caused by rotator cuff injury. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0020] Figure 1A schematic flow chart of a method for locating a shoulder cyst based on magnetic resonance imaging images provided by an embodiment of the present invention;
[0021] Figure 2 A schematic diagram of humeral head region detection provided by an embodiment of the present invention;
[0022] Figure 3 A schematic structural diagram of a shoulder cyst localization system based on magnetic resonance imaging images provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0023] The technical solution of the present invention is described below in conjunction with the accompanying drawings.
[0024] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.
[0025] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.
[0026] Reference Manual Attached Figure 1 , showing a flow chart of a method for locating a shoulder cyst based on magnetic resonance imaging images provided by an embodiment of the present invention.
[0027] The embodiment of the present invention provides a method for locating a shoulder cyst based on magnetic resonance imaging images. The method can be implemented by a shoulder cyst locating device based on magnetic resonance imaging images. The shoulder cyst locating device based on magnetic resonance imaging images can be a terminal or a server. The processing flow of the method for locating a shoulder cyst based on magnetic resonance imaging images can include the following steps:
[0028] S1: Acquire multiple magnetic resonance imaging images.
[0029] For example, 15 MRI images are acquired for each patient.
[0030] Among them, Magnetic Resonance Imaging (MRI) is a medical imaging technology that generates detailed images of structures in the body through powerful magnetic fields and radio waves. It does not use ionizing radiation, but instead uses the arrangement and energy release of hydrogen atoms to obtain high-resolution soft tissue images. MRI can clearly show the details of the brain, spinal cord, joints and other soft tissues, and is widely used in the diagnosis of the nervous system, musculoskeletal system, abdomen and cardiovascular system. This technology has important diagnostic value in the medical field due to its superior image contrast and radiation-free characteristics.
[0031] S2: Construct a humeral head area target detection model and a cyst target detection model based on deep learning algorithm respectively.
[0032] Optionally, the deep learning algorithm includes a long short-term memory neural network and a recurrent neural network.
[0033] In the present invention, the deep learning model can automatically extract useful features from the image without manually defining the features, so more complex and subtle features can be learned, thereby improving the accuracy of target detection. Magnetic resonance imaging (MRI) images usually have high dimensions and complex structures, and deep learning algorithms are particularly suitable for processing these high-dimensional data because they can learn high-level representations and patterns in the images. In addition, long short-term memory networks (LSTM) or recurrent neural networks (RNN) can handle time dependencies in image sequences. If the image sequence has temporal or spatial correlations, these algorithms can capture these correlations, thereby further improving the accuracy and stability of detection.
[0034] Reference Manual Attached Figure 2 , showing a schematic diagram of humeral head area detection provided by an embodiment of the present invention.
[0035] S3: Input the magnetic resonance imaging image into the humeral head region target detection model for detection, and output the humeral head region image.
[0036] In a possible implementation, S3 specifically includes sub-steps S301 to S304:
[0037] S301: Inputting the magnetic resonance imaging image into the humeral head area target detection model for detection, and outputting the humeral head position information.
[0038] The specific information of the humeral head position is:
[0039] R1={cx1,cy1,w1,h1}
[0040] Among them, R1 represents the position information of the humeral head, cx1 and cy1 respectively represent the ratios of the horizontal coordinate and vertical coordinate of the center point of the humeral head area image in the magnetic resonance imaging image to the image height and image width of the magnetic resonance imaging image, and w1 and h1 respectively represent the ratios of the width and height of the humeral head detection box to the image width and image height of the magnetic resonance imaging image.
[0041] S302: Calculate the center point coordinates of the humeral head region image in the magnetic resonance imaging image and the width and height of the humeral head detection frame according to the magnetic resonance imaging image and the humeral head position information.
[0042] In a possible implementation manner, S302 specifically includes:
[0043] According to the following formula, the center point coordinates of the humeral head region image in the magnetic resonance imaging image and the width and height of the humeral head detection box are calculated:
[0044] X A =cx1×W
[0045] Y A =cy1×H
[0046] W A =w1×W
[0047] H A =h1×H
[0048] Among them, X A and Y A W and H represent the width and height of the MRI image, respectively. A and H A Respectively represent the width and height of the humeral head detection box.
[0049] S303: Calculate the vertex coordinates of the humeral head detection frame according to the center point coordinates of the humeral head region image in the magnetic resonance imaging image and the width and height of the humeral head detection frame.
[0050] In a possible implementation manner, S303 specifically includes:
[0051] Calculate the vertex coordinates of the humeral head detection box according to the following formula:
[0052]
[0053] Among them, X1 and Y1 respectively represent the horizontal coordinate and vertical coordinate of the vertex A of the humeral head detection frame, X2 and Y2 respectively represent the horizontal coordinate and vertical coordinate of the vertex B of the humeral head detection frame, X3 and Y3 respectively represent the horizontal coordinate and vertical coordinate of the vertex C of the humeral head detection frame, and X4 and Y4 respectively represent the horizontal coordinate and vertical coordinate of the vertex D of the humeral head detection frame.
[0054] S304: Outputting a humeral head region image according to the vertex coordinates of the humeral head detection frame.
[0055] It should be noted that the width of the humeral head area image output at this time is W A , high is H A .
[0056] In the present invention, through the deep learning model, the humeral head area can be automatically identified and located from the MRI image, reducing the subjectivity and error of manual annotation, thereby improving the accuracy of positioning. When detecting the humeral head area, the vertex coordinates of the detection frame can be accurately calculated, thereby extracting a more accurate humeral head area. This is very important for subsequent cyst detection and positioning, ensuring that subsequent steps can be performed in the correct area. During training and use, standardized humeral head area images can improve the robustness and stability of the model, so that the model can still maintain a high detection performance when facing MRI images of different patients.
[0057] S4: Input the humeral head area image into the cyst target detection model for detection, and output the cyst area image.
[0058] It should be noted that the width of the output cyst area image is W A , high is H A .
[0059] In a possible implementation, S4 specifically includes sub-steps S401 to S402:
[0060] S401: Inputting the humeral head area image into the cyst target detection model for detection, and outputting the cyst location information.
[0061] The specific information of cyst location is as follows:
[0062] R2={cx2,cy2,w2,h2}
[0063] Among them, R2 represents the cyst location information, cx2 and cy2 respectively represent the ratios of the horizontal coordinate and vertical coordinate of the center point of the cyst area image in the humeral head area image to the image height and image width of the humeral head area image, and w2 and h2 respectively represent the ratios of the width and height of the cyst detection frame to the image width and image height of the humeral head area image.
[0064] S402: Outputting the center point coordinates of the cyst region image in the humeral head region image according to the humeral head region image and the cyst location information.
[0065] In a possible implementation manner, S402 specifically includes:
[0066] According to the following formula, the coordinates of the center point of the cyst area image in the humeral head area image are output:
[0067] X C =cx2×W A
[0068] Y C =cy2×H A
[0069] Among them, X C and Y C They respectively represent the abscissa and ordinate of the center point of the cyst area image in the humeral head area image.
[0070] In the present invention, by specifically detecting the cyst in the humeral head area image, the influence of background interference is reduced. The cyst detection model only focuses on the specific area within the humeral head area, so that the cyst can be detected and located more accurately, reducing the interference of noise and irrelevant parts in the image on the detection results. The center point coordinates of the cyst area image are calculated using cyst location information (such as center point coordinates, ratio of detection frame), so that the cyst can be located more accurately. This method enables the model to focus on the detected cyst area and improves the accuracy of detection. By outputting the center point coordinates of the cyst area image, the accurate positioning of the cyst area is ensured. This provides consistent and reliable data support for subsequent analysis, evaluation and possible treatment plan formulation.
[0071] S5: According to the coordinates of the center point of the cyst area image, the position of the cyst is determined by a spatial positioning algorithm.
[0072] In a possible implementation, S5 specifically includes sub-steps S501 to S507:
[0073] S501: Taking the center point of the humeral head region image as the coordinate origin, a rectangular coordinate system is established:
[0074]
[0075] Among them, X B and Y B They respectively represent the abscissa and ordinate of the center point of the humeral head area image.
[0076] Specifically, in each magnetic resonance imaging image, a rectangular coordinate system is established with the center point of the humeral head region image as the coordinate origin.
[0077] In the present invention, by establishing a rectangular coordinate system with the center point of the humeral head region image as the coordinate origin, the position of the cyst can be accurately located in each magnetic resonance imaging image. Such a coordinate system helps to uniformly convert the position data of the cyst from different images into the same coordinate system, thereby improving the accuracy of positioning.
[0078] S502: Save the cyst detection result of each magnetic resonance imaging image as a sequence vector:
[0079] L i =(l x ,l y ,l z )
[0080]
[0081] Among them, L i The sequence vector representing the cyst detection result of the i-th MRI image, l x and l y They represent the relative positions of the cyst area image on the X-axis and Y-axis in the rectangular coordinate system, respectively. z Indicates the serial number of the magnetic resonance imaging image, x i B and i B represents the horizontal coordinate and vertical coordinate of the center point of the i-th humeral head area image, respectively, i C and i C They respectively represent the abscissa and ordinate of the center point of the cyst area image in the i-th humeral head area image.
[0082] In the present invention, the cyst detection results of each magnetic resonance imaging image are saved as a sequence vector, and the correlation in the image sequence is considered, which helps to capture the changes of cysts in different images. This method can process and analyze the temporal changes in the image sequence, and enhance the stability and reliability of detection and positioning.
[0083] S503: Counting the serial number of each magnetic resonance imaging image in which a cyst detection result appears.
[0084] S504: Determine whether the cyst is located at the front or back of the humeral head region according to the number of the sequence numbers of the magnetic resonance imaging images counted:
[0085] C front >C back , located at the front
[0086] C front <C back, located at the rear
[0087] Among them, C front Indicates the number of cysts located in the anterior part of the humeral head area, C back Indicates the number of cysts located posteriorly in the humeral head region.
[0088] For example, the serial number of each magnetic resonance imaging image showing a cyst detection result is counted, and the number of magnetic resonance imaging images with serial numbers 1-8 showing a cyst detection result is counted as C front The number of cyst detection results in the magnetic resonance imaging images with sequence numbers 8-15 is C back , C front With C back A comparison is made to determine whether the cyst is located in the front or back of the humeral head area, and the rectangular coordinate system is divided into eight areas.
[0089] Specifically, when the cyst is located at the front side of the humeral head region, i.e., the magnetic resonance imaging images with sequence numbers 1-8, the second quadrant and the third quadrant in the rectangular coordinate system are the side close to the body. When the cyst is located at the back side of the humeral head region, i.e., the magnetic resonance imaging images with sequence numbers 8-15, the first quadrant and the fourth quadrant in the rectangular coordinate system are the side close to the body. In this way, the rectangular coordinate system is divided into eight regions.
[0090] In the present invention, by counting the number of cyst area images in each quadrant in a rectangular coordinate system, the main distribution area of the cyst can be determined. This method helps to understand the distribution of cysts in the humeral head area, which is helpful for further analysis and processing. By comparing image data with different sequence numbers, it is possible to effectively determine whether the cyst is located on the front or back side of the humeral head area. This partitioning method makes the positioning of the cyst more accurate, providing strong data support for the formulation of diagnosis and treatment plans.
[0091] S505: Determine the quadrant where the cyst region image is located in the rectangular coordinate system according to the coordinates of the center point of the cyst region image.
[0092] S506: Count the number of cyst region images in each quadrant of the rectangular coordinate system, and take the quadrant with the largest number of cyst region images as the quadrant where the cyst is located:
[0093] max{C first ,C second ,C third ,C fourth}
[0094] Among them, C first represents the number of cyst area images in the first quadrant, C second represents the number of cyst area images in the second quadrant, Cthird represents the number of cyst area images in the third quadrant, C fourth Indicates the number of images of the cystic area in the fourth quadrant.
[0095] S507: Determine the location of the cyst according to the quadrant where the cyst is located.
[0096] Furthermore, after the cyst positioning task is judged to be completed, the location of the cyst is output to a folder with the positioning name.
[0097] In the present invention, the position of the cyst can be accurately located by determining the quadrant in which the cyst region image is located based on the coordinates of the center point of the cyst region image. Different quadrants represent different positions of the cyst in the humeral head region, and this precise spatial positioning can provide useful position information. Counting the number of cyst region images in each quadrant and selecting the quadrant with the largest number as the quadrant where the cyst is located helps to identify the area where the cyst most often occurs. This statistical analysis can reveal the distribution pattern of the cyst, which is helpful for diagnosis and treatment. Outputting the location of the cyst to a specified folder helps to organize and manage the positioning results. Structured file management can facilitate subsequent data query, processing and analysis, and improve work efficiency.
[0098] In a possible implementation, the model training method includes:
[0099] A total of 512 magnetic resonance imaging images of humeral head cysts in the shoulder were collected as the initial dataset.
[0100] Magnetic resonance imaging images of humeral head cysts in the shoulder were annotated by two surgeons using LabelMe image annotation software and reviewed by a sports injury expert.
[0101] After the data expansion, 1239 datasets were used as training datasets and 99 were used as validation datasets.
[0102] According to the training data set, the batch size is set to 2, the number of training times is 150, the model is trained, and the model parameters are updated through the AMD optimization algorithm, where the initial learning rate is 0.001.
[0103] In the present invention, 512 magnetic resonance imaging images of humeral head cysts in the shoulder were collected and annotated and checked by two surgeons and a kinematic injury expert to ensure the quality and accuracy of data annotation. This multi-level annotation and review mechanism effectively reduces annotation errors and omissions and improves the reliability of training data. Setting the batch size to 2 and the number of training times to 150 times can balance the computational efficiency and model performance during training. A smaller batch size helps the model adjust the weights more finely each time it is updated, while a larger number of training times ensures that the model fully learns the data features.
[0104] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:
[0105] In the present invention, by inputting the magnetic resonance imaging image into the humeral head area target detection model for detection, outputting the humeral head area image, and then inputting the humeral head area image into the cyst target detection model for detection, and outputting the cyst area image, the influence of speckle noise and echo disturbance on the MRI image is reduced, the image quality is improved, and the difference between the cyst and normal tissue becomes more obvious. The position of the cyst is determined according to the center point coordinates of the cyst area image through a spatial positioning algorithm, reducing the dependence on the doctor's cognitive ability and clinical experience, realizing automatic positioning, and avoiding confusion between the humeral head cyst and the effusion caused by rotator cuff injury.
[0106] Reference Manual Attached Figure 2 , showing a structural schematic diagram of a shoulder cyst positioning system based on magnetic resonance imaging images provided by the present invention.
[0107] The present invention also provides a shoulder cyst localization system 30 based on magnetic resonance imaging images, comprising: a memory 303 and one or more processors 301 .
[0108] One or more application programs are stored in the memory 303 , and the one or more application programs are suitable for being executed by the one or more processors 301 to implement the shoulder cyst localization method based on magnetic resonance imaging images described in the method embodiment.
[0109] The shoulder cyst localization system 30 based on magnetic resonance imaging images includes: a processor 301 and a memory 303. The processor 301 and the memory 303 are connected, for example, via a bus 302.
[0110] The structure of the shoulder cyst localization system 30 based on magnetic resonance imaging images does not constitute a limitation on the embodiment of the present invention.
[0111] Processor 301 may be a CPU, a general purpose processor, a DSP, an ASIC, an FPGA or other programmable logic device, a transistor logic device, a hardware component or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of the present invention. Processor 301 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.
[0112] The bus 302 may include a path to transmit information between the above components. The bus 302 may be a PCI bus or an EISA bus, etc. The bus 302 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.
[0113] The memory 303 can be a ROM or other type of static storage device that can store static information and instructions, a RAM or other type of dynamic storage device that can store information and instructions, or an EEPROM, a CD-ROM or other optical disk storage, an optical disk storage (including a compressed optical disk, a laser disk, an optical disk, a digital versatile disk, a Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited to these.
[0114] It should be noted that the shoulder cyst localization system 30 based on magnetic resonance imaging images can implement the above-mentioned shoulder cyst localization method based on magnetic resonance imaging images, and can achieve the same or similar technical effects. To avoid repetition, the present invention will not go into details.
[0115] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:
[0116] In the present invention, by inputting the magnetic resonance imaging image into the humeral head area target detection model for detection, outputting the humeral head area image, and then inputting the humeral head area image into the cyst target detection model for detection, and outputting the cyst area image, the influence of speckle noise and echo disturbance on the MRI image is reduced, the image quality is improved, and the difference between the cyst and normal tissue becomes more obvious. The position of the cyst is determined according to the center point coordinates of the cyst area image through a spatial positioning algorithm, reducing the dependence on the doctor's cognitive ability and clinical experience, realizing automatic positioning, and avoiding confusion between the humeral head cyst and the effusion caused by rotator cuff injury.
[0117] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program can be loaded and executed by a processor to implement the method for locating a shoulder cyst based on magnetic resonance imaging images as described in the first aspect.
[0118] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.
[0119] There are a few points to note:
[0120] (1) The drawings of the embodiments of the present invention only relate to the structures related to the embodiments of the present invention, and other structures may refer to the general design.
[0121] (2) For the sake of clarity, in the drawings used to describe the embodiments of the present invention, the thickness of the layers or regions is exaggerated or reduced, that is, these drawings are not drawn according to the actual scale. It is 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 may be "directly" "on" or "under" the other element or there may be intermediate elements.
[0122] (3) In the absence of conflict, the embodiments of the present invention and the features therein may be combined with each other to obtain new embodiments.
[0123] The above are only specific embodiments 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 based on the protection scope of the claims.
Claims
1. A method for locating a shoulder cyst based on magnetic resonance imaging images, characterized in that: include: S1: Acquire multiple magnetic resonance imaging images; S2: Build a humeral head region target detection model and a cyst target detection model based on deep learning algorithm respectively; S3: inputting the magnetic resonance imaging image into the humeral head region target detection model for detection, and outputting a humeral head region image; S4: inputting the humeral head region image into the cyst target detection model for detection, and outputting the cyst region image; S5: Determine the position of the cyst according to the coordinates of the center point of the cyst area image by using a spatial positioning algorithm.
2. The method for locating a shoulder cyst based on magnetic resonance imaging according to claim 1, characterized in that: The S3 specifically includes: S301: inputting the magnetic resonance imaging image into the humeral head area target detection model for detection, and outputting humeral head position information; S302: Calculating the center point coordinates of the humeral head region image in the magnetic resonance imaging image and the width and height of the humeral head detection frame according to the magnetic resonance imaging image and the humeral head position information; S303: Calculating the vertex coordinates of the humeral head detection frame according to the center point coordinates of the humeral head region image in the magnetic resonance imaging image and the width and height of the humeral head detection frame; S304: Outputting the humeral head region image according to the vertex coordinates of the humeral head detection frame.
3. The method for locating shoulder cysts based on magnetic resonance imaging images according to claim 2, characterized in that: The humeral head position information is specifically: R1={cx1,cy1,w1,h1} Among them, R1 represents the position information of the humeral head, cx1 and cy1 respectively represent the ratios of the horizontal coordinate and vertical coordinate of the center point of the humeral head area image in the magnetic resonance imaging image to the image height and image width of the magnetic resonance imaging image, and w1 and h1 respectively represent the ratios of the width and height of the humeral head detection box to the image width and image height of the magnetic resonance imaging image.
4. The method for locating shoulder cysts based on magnetic resonance imaging images according to claim 3, characterized in that: The S302 is specifically as follows: The center point coordinates of the humeral head region image in the magnetic resonance imaging image and the width and height of the humeral head detection frame are calculated according to the following formula: X A =cx1×W AND A =cy1×H IN A =w1×W H A =h1×H Among them, X A and Y A W and H represent the width and height of the MRI image, respectively. A and H A Respectively represent the width and height of the humeral head detection box.
5. The method for locating shoulder cysts based on magnetic resonance imaging images according to claim 4, characterized in that: The S303 is specifically: The vertex coordinates of the humeral head detection frame are calculated according to the following formula: Among them, X1 and Y1 respectively represent the horizontal coordinate and vertical coordinate of the vertex A of the humeral head detection frame, X2 and Y2 respectively represent the horizontal coordinate and vertical coordinate of the vertex B of the humeral head detection frame, X3 and Y3 respectively represent the horizontal coordinate and vertical coordinate of the vertex C of the humeral head detection frame, and X4 and Y4 respectively represent the horizontal coordinate and vertical coordinate of the vertex D of the humeral head detection frame.
6. The method for locating shoulder cysts based on magnetic resonance imaging images according to claim 1, characterized in that: The S4 specifically includes: S401: inputting the humeral head region image into the cyst target detection model for detection, and outputting cyst location information; S402: Outputting the center point coordinates of the cyst region image in the humeral head region image according to the humeral head region image and the cyst location information.
7. The method for locating shoulder cysts based on magnetic resonance imaging images according to claim 6, characterized in that: The cyst location information is specifically: R2={cx2,cy2,w2,h2} Among them, R2 represents the cyst location information, cx2 and cy2 respectively represent the ratios of the horizontal coordinate and vertical coordinate of the center point of the cyst area image in the humeral head area image to the image height and image width of the humeral head area image, and w2 and h2 respectively represent the ratios of the width and height of the cyst detection frame to the image width and image height of the humeral head area image.
8. The method for locating shoulder cysts based on magnetic resonance imaging images according to claim 7, characterized in that: The S402 is specifically as follows: According to the following formula, the center point coordinates of the cyst area image in the humeral head area image are output: X C =cx2×W A AND C =cy2×H A Among them, X C and Y C They respectively represent the abscissa and ordinate of the center point of the cyst area image in the humeral head area image.
9. The method for locating shoulder cysts based on magnetic resonance imaging images according to claim 8, characterized in that: The S5 specifically includes: S501: Taking the center point of the humeral head region image as the coordinate origin, a rectangular coordinate system is established: Among them, X B and Y B Respectively represent the abscissa and ordinate of the center point of the humeral head region image; S502: Save the cyst detection result of each magnetic resonance imaging image as a sequence vector: L i =(l x ,l y ,l z ) Among them, L i The sequence vector representing the cyst detection result of the i-th MRI image, l x and l y They represent the relative positions of the cyst area image on the X-axis and Y-axis in the rectangular coordinate system, respectively. z Indicates the serial number of the magnetic resonance imaging image, x i B and i B represents the horizontal coordinate and vertical coordinate of the center point of the i-th humeral head area image, respectively, i C and i C represent the abscissa and ordinate of the center point of the cyst region image in the i-th humeral head region image respectively; S503: Counting the sequence number of each magnetic resonance imaging image showing a cyst detection result; S504: Determine whether the cyst is located at the front or back of the humeral head region according to the number of the sequence numbers of the magnetic resonance imaging images counted: C front >C back , located at the front C front <C back , located at the rear Among them, C front Indicates the number of cysts located in the anterior part of the humeral head area, C back Indicates the number of cysts located posterior to the humeral head region; S505: determining the quadrant where the cyst region image is located in the rectangular coordinate system according to the coordinates of the center point of the cyst region image; S506: Count the number of cyst region images in each quadrant of the rectangular coordinate system, and take the quadrant with the largest number of cyst region images as the quadrant where the cyst is located: max{C first ,C second ,C third ,C fourth } Among them, C first represents the number of cyst area images in the first quadrant, C second represents the number of cyst area images in the second quadrant, C third represents the number of cyst area images in the third quadrant, C fourth represents the number of images of the cystic area in the fourth quadrant; S507: Determine the location of the cyst according to the quadrant where the cyst is located.
10. A shoulder cyst localization system based on magnetic resonance imaging, characterized in that: include: memory and one or more processors; One or more application programs are stored in the memory, and the one or more application programs are suitable for being executed by the one or more processors to implement the shoulder cyst localization method based on magnetic resonance imaging images as described in any one of claims 1 to 9.
Citation Information
Patent Citations
Prostate cancer automatic detection and classification system based on deep learning
CN111028206A
Image segmentation method, radiotherapy system, computer equipment and storage medium
CN112001925A
Liver CT image multi-lesion classification method based on sample generation and transfer learning
CN112241766A
Medical image brain tumor detection method and system based on improved Faster R-CNN
CN114445328A
Processing method of rotator cuff scanning image, electronic equipment and storage medium
CN116645346A
