CN112150451A - Symmetry information detection method and device, computer equipment and storage medium - Google Patents

Abstract

The present application relates to a symmetrical information detection method, device, computer equipment and storage medium. By acquiring medical image data of a target part and inputting the medical image data into a preset network model, a medical image that can reflect the target part or the target part is obtained. A heat map of the symmetry of the image is obtained, and then the heat map is fitted to obtain the symmetry information of the target part or the medical image of the target part. The method performs fitting based on the actual symmetry distribution of the target part or the medical image of the target part, and can accurately and efficiently determine the symmetry information of the target part or the medical image of the target part.

Description

Symmetric information detection method, device, computer equipment and storage medium

technical field

The present application relates to the field of medical technology, and in particular, to a symmetrical information detection method, device, computer equipment and storage medium.

Background technique

In the process of medical image scanning, when scanning the target part, such as the head, due to the improper placement of the scanned target part, the scanned image will be tilted, and the corresponding images of the sine and coronal planes will not be very good. The anatomical structure of the target site is displayed, so the symmetry of the target site needs to be determined before reading or performing symmetry analysis.

In the traditional technology, the way of adjusting the symmetry of the image includes, but is not limited to, manual adjustment by a doctor, or analysis and adjustment through an algorithm. Manual adjustment by the doctor will seriously waste the doctor's reading time. When analyzing and adjusting through the algorithm, each algorithm also has certain limitations. For example, the plane equation is optimized by calculating the symmetry correlation coefficient of the original image and the flipped image, but The adaptability to the data with lesions in the target site is not particularly good; alternatively, search line by line to find the points where the pixel value change of each line satisfies the statistical curve, and finally perform random sampling on these points to fit a plane, but it is only applicable to the Image of statistical curve.

Therefore, the prior art lacks a method that can efficiently and accurately determine the symmetry plane or symmetry line of the image of the scanned part.

SUMMARY OF THE INVENTION

Based on this, it is necessary to provide a symmetry information detection method, device, computer equipment and storage medium that can efficiently and accurately determine the symmetry information of the image of the scanned part, aiming at the above technical problems.

In a first aspect, an embodiment of the present application provides a symmetrical information detection method, the method comprising:

Obtain medical image data of the target part;

Input the medical image data into a preset network model to obtain a heat map of the symmetry information of the target part or the medical image of the target part; the heat map reflects the symmetry distribution of the target part or the medical image of the target part;

Fit the heat map to obtain the symmetry information of the target part or the medical image of the target part.

In one of the embodiments, the above-mentioned obtaining medical image data of the target part includes:

Acquire the original medical image of the target part;

The original medical image is preprocessed to obtain the medical image data of the target part.

In one of the embodiments, the above-mentioned preprocessing includes adjusting the resolution of the original medical image, normalizing the gray value of each pixel in the original medical image, and cutting out one of the regions of the target portion from the original medical image. species or several.

In one embodiment, before the above-mentioned inputting the medical image data into the preset network model, the method further includes:

Obtaining training sample data; the training sample data includes the medical image data of each part and the heat map corresponding to the medical image data of each part;

According to the training sample data, the initial network is trained, and the network model is obtained after the training is completed.

In one embodiment, the above-mentioned acquiring training sample data includes:

Obtain sample medical images of various parts, and perform incremental processing on each sample medical image to obtain incrementally processed sample medical images;

Obtain the sample heat map corresponding to each incrementally processed sample medical image;

Each incrementally processed sample medical image and the corresponding sample heatmap are determined as training sample data.

In one embodiment, the above-mentioned incremental processing of each sample medical image includes:

Incremental processing is performed on each sample medical image by a preset incremental method; the incremental method includes at least one of data cropping, data rotation, adding Gaussian noise, and data translation.

In one of the embodiments, the above-mentioned medical image data includes two-dimensional medical image data or three-dimensional volume data.

In a second aspect, an embodiment of the present application provides a device for detecting symmetrical information, the device comprising:

an acquisition module for acquiring medical image data of the target part;

The processing module is used to input the medical image data into the preset network model to obtain the target part or the heat map of the medical image symmetry of the target part; the heat map reflects the symmetry of the target part or the medical image of the target part Distribution;

The fitting module is used for fitting the heat map to obtain the symmetry information of the target part or the medical image of the target part.

In a third aspect, an embodiment of the present application provides a computer device, including a memory and a processor, the memory stores a computer program, and the processor implements any one of the method steps in the above first aspect embodiments when the processor executes the computer program.

In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, implements any one of the method steps in the above first aspect embodiment.

The embodiments of the present application provide a symmetrical information detection method, device, computer equipment, and storage medium. By acquiring medical image data of a target part, and inputting the medical image data into a preset network model, a result that can reflect the target part or the target part is obtained. The heat map of the symmetry of the medical image is obtained, and then the heat map is fitted to obtain the symmetry information of the target part or the medical image of the target part. In this method, because the adopted preset network model learns the symmetry information heat map of the target part or the medical image of the target part from the medical image data, that is, the symmetry distribution of the target part is known in advance through the heat map, and then based on the heat map Plane fitting is performed on the symmetry distribution, so that the fitting is performed based on the actual symmetry distribution of the target part or the medical image of the target part, and the symmetry information of the target part can be accurately and efficiently determined; and the network model is in the It has been trained before use. By inputting the medical image data into the preset network model, the heat map of the target part or the medical image of the target part can be quickly and accurately obtained, which further improves the determination of the target part or the medical image of the target part. Efficiency and accuracy of image symmetry information.

Description of drawings

1 is an application environment diagram of a symmetrical information detection method in one embodiment;

2 is a schematic flowchart of a method for detecting symmetrical information in one embodiment;

Figure 2a is a schematic diagram of a heat map in one embodiment;

2b is a schematic diagram of a symmetrical information detection result in one embodiment;

2c is a schematic diagram of a symmetrical information detection result in another embodiment;

3 is a schematic flowchart of a symmetrical information detection method in another embodiment;

4 is a schematic flowchart of a symmetrical information detection method in another embodiment;

5 is a schematic flowchart of a symmetrical information detection method in another embodiment;

5a is a schematic diagram of a data clipping method in another embodiment;

5b is a schematic diagram of a data clipping method in another embodiment;

6 is a flowchart of a method for detecting symmetrical information in one embodiment;

FIG. 7 is a schematic structural diagram of an apparatus for detecting symmetrical information in an embodiment.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present application more clearly understood, the present application will be described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, but not to limit the present application.

Reference herein to an "embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor a separate or alternative embodiment that is mutually exclusive of other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

A symmetrical information detection method provided by the present application can be applied to the application environment shown in FIG. 1 . The computer device in FIG. 1 includes a processor, a memory and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium, an internal memory. The nonvolatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the execution of the operating system and computer programs in the non-volatile storage medium. The database of the computer device is used to store data related to the detection of symmetrical information. The network interface of the computer device is used to communicate with an external terminal through a network connection. The computer program, when executed by the processor, implements a symmetrical information detection method.

The symmetry adjustment of the image plane in the scanning process by the related technology is based on the analysis of traditional algorithms, including but not limited to optimizing the plane equation by calculating the symmetry correlation coefficient of the original image and the flipped image; The change of row pixel value satisfies the points on the statistical curve, and finally these points are randomly sampled to fit a plane; or the points near the sagittal plane in the image are sampled, the features are manually extracted, and the random forest is used for regression to find that the midsagittal plane response is relatively high The value of , so as to fit the final plane; the symmetry information can also be analyzed by segmenting specific structures; however, calculating the symmetry correlation coefficient of the original image and the flipped image to optimize the plane equation is not particularly suitable for data with lesions. The method of line search to find the point on which the pixel value change of each line satisfies the statistical curve is only applicable to the image that meets the statistical curve. The segmentation scheme is limited in some scenarios, for example, if the head symmetry information is obtained by segmenting the eyes, but the brain data does not necessarily have eyes.

Based on this, the embodiments of the present application provide a symmetrical information detection method, device, computer equipment, and storage medium, which can efficiently and accurately determine the symmetrical information of a scanned part image. The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems will be specifically described in detail below with reference to the accompanying drawings. The following specific embodiments may be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. It should be noted that, in a symmetrical information detection method provided by the present application, the execution subject of FIG. 2 to FIG. 6 is a computer device, wherein the execution subject may also be a symmetrical information detection device, wherein the device can be detected by software, hardware or The combination of software and hardware is realized as part or all of the computer equipment.

In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments It is a part of the embodiments of the present application, but not all of the embodiments. The symmetric information detection method provided by the present application will be described below through specific embodiments.

As shown in FIG. 2, in one embodiment, a symmetrical information detection method is provided, and the method includes the following steps:

S101, acquiring medical image data of a target part.

The medical image data includes, but is not limited to, Computed Tomography (CT) images, Magnetic Resonance Imaging (MRI) images, etc. Optionally, the medical image data includes a two-dimensional medical image or a three-dimensional volume The data, for example, may be a two-dimensional tomographic image of an image of a shoulder, a head, or other parts, or three-dimensional volume data, which is not limited in this embodiment of the present application. The target part refers to the part that needs to be scanned currently, for example, the data of scanning the head, and it may also be head images of parts such as the cranial brain, the sinuses, the orbit, the inner ear, etc., which are not limited in this embodiment of the present application.

The acquired medical image data is the medical image to be input into the network model, which can be the original medical image data collected by the imaging device, or the image stored in a workstation or a storage medium such as a PACS (Medical Image Archiving and Communication System).

For example, the computer device may acquire the medical image data of the target part by requesting the computer device to acquire the medical image device, and then the medical image device collects the image of the target part and then sends it back to the computer device; or, the computer device accepts a third-party device. An image input through a connection interface; or, an image downloaded from a network or a medical image database, etc., are not limited in this embodiment of the present application.

S102: Input the medical image data into a preset network model to obtain a heat map of the symmetry information of the target part or the medical image of the target part; the heat map reflects the distribution of the symmetry of the target part or the medical image of the target part.

After obtaining the medical image data of the target part, input the medical image data of the target part into a preset network model, the network model is a model that is pre-trained and used to extract heat maps from medical images, such as , the network model can be a neural network model trained based on deep learning technology, or an algorithm model trained based on mathematical algorithms, etc. The network model trained by deep learning technology can be any one of regression, segmentation or positioning network models. This implementation Examples are not limited to this model.

After the medical image data of the target part is input into the preset network model, the obtained output result is a heat map of the symmetrical information of the target part or the medical image of the target part, wherein the heat map refers to a heat map that can reflect the target part or See Figure 2a for a diagram of the symmetry distribution of the medical image of the target site. The left side is an example of a medical image, and the right side is its corresponding heat map. Commonly used heat map representations include but are not limited to masks. The image can accurately obtain the symmetry information of the target part or the medical image of the target part, and can perform subsequent operations such as symmetry adjustment according to the obtained symmetry information.

S103: Fit the heat map to obtain the target part or the symmetry information of the medical image of the target part.

Fitting the heatmap, for example, performing plane fitting, commonly used plane fitting methods include but are not limited to least squares linear fitting, SVD or principal component analysis (Principal Component Analysis, PCA) to calculate the principal plane, and the like.

Among them, the symmetry information includes a symmetry plane or a symmetry line, that is, after the plane fitting is performed on the heat map, the symmetry plane or symmetry line of the medical image of the target part or the target part is obtained. The symmetry information (symmetry plane or symmetry line) can be used for Adjust the symmetry of the target part in the medical image data. Among them, if it is a three-dimensional medical image, the plane of symmetry is obtained, and if it is a two-dimensional medical image, the line of symmetry is obtained. Please refer to FIG. 2b and FIG. 2c, which illustrate the test results detected by the symmetrical information detection method provided by the embodiment of the present application, wherein, FIG. 2b is a CT image of the head of the soft tissue window through the symmetrical information detection method provided by the embodiment of the present application. Figure 2c is an example result diagram of the symmetry plane detected by the CT image of the head of the bone window by the symmetry information detection method provided by the embodiment of the present application.

In a method for detecting symmetry information provided in this embodiment, by acquiring medical image data of a target part, and inputting the medical image data into a preset network model, a thermal force that can reflect the symmetry of the target part or the medical image of the target part is obtained. Then, the heat map is fitted to obtain the symmetry information of the target part or the medical image of the target part. In this method, since a preset network model is used to learn the heat map of the target part or the medical image of the target part from the medical image data, that is, the symmetry distribution of the target part is known in advance through the heat map, and then based on the symmetry In this way, the fitting is performed based on the actual symmetry distribution of the medical image of the target part or the target part, and the symmetry information of the target part or the medical image of the target part can be accurately and efficiently determined; and The network model has been trained before use. By inputting the medical image data into the preset network model, the heat map of the target part or the medical image of the target part can be quickly and accurately obtained, which further improves the determination of the target part or target. Efficiency and accuracy of symmetry information in medical images of parts.

On the basis of the above embodiment, an embodiment of a symmetrical information detection method is also provided. This embodiment mainly relates to the specific process of obtaining medical image data of a target part by a computer device. As shown in FIG. 3 , this embodiment Include the following steps:

S201, collecting the original medical image of the target part.

This embodiment is described by taking the method of collecting medical image data of the target part by medical equipment as an example; then, the original medical image of the target part refers to the uncropped and unprocessed image collected at the beginning of the scan, that is, the collected image. The original medical image is subject to the range that the device can scan, which may include other parts than the target part. For example, when scanning head data, the target part is the sinuses, and the scanned original medical images include not only the sinuses but also the orbits. , the included eye socket may be part or all, which is not limited.

S202 , preprocessing the original medical image to obtain medical image data of the target part.

After the original medical image of the target part is collected, the original medical image is preprocessed to obtain the medical image of the target part. The preprocessing is to remove the redundant and unnecessary areas in the original image, or to process the size, scale, resolution, etc. of the original medical image to meet the requirements of the network model, so that the network model can be effectively used to determine the target area. Heat map, to avoid the unsatisfactory input of original medical images into the network model.

Optionally, the preprocessing that can be performed includes one of adjusting the resolution of the original medical image, normalizing the gray value of each pixel in the original medical image, and intercepting the region of the target site from the original medical image. several.

The preprocessing is used to process the original medical image to meet the requirements of the input data of the regression network model. For example, the original medical image can be adjusted to a uniform resolution: [1.796mm, 1.796mm, 5mm], and the gray value of the original medical image can be normalized: the value in the gray range of the original image [-1024, 2000] stretch to [0,1]. For another example, assuming that the target part is the entire area of the head, the redundant areas below the head can be cropped to reduce interference: the top layer position of the head can be calculated from the CT image or the maximal intensity projection (mip) map in the sagittal direction of the CT, Then the image is intercepted according to the physical length of the head, the slices below the head are discarded, and only the head area is retained.

The image obtained after the above preprocessing of the original medical image can be used as the medical image data of the target part to be input into the network model, so that the medical image data of the target part can accurately obtain the target part or the target part through the network model. heatmap of medical images.

The function of the network model is generally to carry out targeted training according to its usage scenarios, so as to obtain a network model that is more in line with the effect of each scheme. Therefore, before obtaining the heat map of the target part through the preset network model, it is necessary to pre-collect and train. sample data, and train the above network model according to the training sample data. Based on this, the training process of the network model is described in detail below.

As shown in Figure 4, in one embodiment, the training process of the network model includes:

S301 , acquiring training sample data; the training sample data includes medical image data of each part and a heat map corresponding to the medical image data of each part.

Among them, the training sample data is the data used to train the network. In practical applications, in addition to the training sample data, you can also prepare some training and verification data, that is, first use the training sample data to train the network model, and then use the training and verification data to input the training data. The performance of the network model is verified in the network model. Among them, the training sample data and the training verification data are data of different batches, and the training verification data needs to be data that does not participate in the training of the training network model.

Among them, the training sample data includes the medical image data of each part and the heat map corresponding to the medical image data of each part; the medical image data of each part, such as the medical image data of the eye frame part, the medical image data of the skull part, and the inner ear medical image data, etc. , when collecting sample data, the more parts, the better, so that the training sample data can be more comprehensive and complete, thereby improving the applicability of the network model.

In addition, in addition to the medical image data of each part, it is also necessary to obtain the heat map corresponding to the medical image data of each part. Each training sample data includes a pair of medical image data of the part and the heat map corresponding to the medical image data of the part. The heat map reflects Symmetry information of the target part or the medical image of the target part, for example, in the case of cranial data, the heat map shows the anatomical position of the falx, where the symmetry is more obvious.

S302 , train the initial network according to the training sample data, and obtain a network model after the training is completed.

After acquiring the training sample data, input the training sample data into the initial network, and let the initial network learn the mapping relationship between the medical image data of each part in the training sample data and the heat map corresponding to each medical image data, and learn repeatedly. During the process, the learning direction of the initial network can be adjusted according to the value of the preset loss function, until the network model is obtained after the training is completed.

After the training is completed to obtain the network model, the above-mentioned training and verification data can be used to verify the network model, that is, the medical image data of each part in the training and verification data is input into the network model to verify the effect of the network. A model with better verification effect under one iteration number, or select a network model with better verification effect from multiple different training network models.

In this embodiment, the network model that can directly obtain the heat map of each part based on the medical image data of each part can be obtained by acquiring the training sample data for training, and the heat map of the target part or the medical image of the target part can be quickly and accurately obtained, which improves the Efficiency and accuracy of determining symmetry planes or lines of symmetry of a target site or a medical image of a target site.

For the above network model, the more types and types of training sample data obtained, the more accurate the heatmap determined by the trained network model.

As shown in FIG. 5, in one embodiment, the above-mentioned acquisition of training sample data includes:

S401 , obtaining sample medical images of various parts, and performing incremental processing on each sample medical image to obtain an incrementally processed sample medical image.

In practical applications, it is not always possible to collect enough data for training. Therefore, in order to enrich the training sample data, the sample medical images of each part can be incrementally processed to obtain rich sample medical images, and the CT head can be used to obtain rich sample medical images. Taking scan data as an example, head scan data includes a variety of scans, and the scan range can include brain scan, sinus scan, orbital scan, inner ear scan, etc.

The sample medical image here refers to the pre-processed medical image of the original medical image of each part, and the specific process can refer to the process shown in FIG. 3 above.

Optionally, the above-mentioned incremental processing of each sample medical image includes: performing incremental processing on each sample medical image through a preset incremental method; the incremental method at least includes data cropping, data rotation, adding Gaussian noise, One of the data translations.

Among them, the preset incremental methods include various methods, such as data cropping, data rotation, adding Gaussian noise, data translation, etc. Among them, taking the sinuses as an example, the specific process of data cropping includes: cleaning the collected sample medical images After that, generate a sagittal Mip map, select the data including the nose from the sagittal image, and then mark the cropping frame on the sagittal plane according to the scan frame of the sinus scan, and cut the data according to the marked frame. Please refer to Figure 5a, (I) in Figure 5a is an example of the sagittal Mip of the head without the nose, (II) in Figure 5a is an example of sinus scan clipping, (III) in Figure 5a For another example of the way to scan and crop the sinuses, please refer to Figure 5b, (I) in Figure 5b is the original image of the head including the sinuses from different perspectives, and (II) and (I) in Figure 5b are the top and bottom of each image. One-to-one corresponding cropped schematic diagram; from Figure 5a and Figure 5b, it is shown that the sinus data is simulated by various cropping methods. Likewise, sample medical images from orbital scans and inner ear scans can be incremented in a similar manner.

Among them, the specific process of data rotation includes: in view of the fact that during the scanning process, the patient may have some incorrect positioning, which causes the data to be skewed, and considering that the head data is symmetrical, the data is randomly flipped left and right and in 3D space. Make a certain angle, such as (0°-30°) increments of left and right angular rotation. In this way, the scanned data is rotated at a certain angle, or flipped left and right to form different samples. In addition, Gaussian noise is added to the sample medical images of each part; or data translation is performed. Considering that some data are not fully captured in the scanning field of view, the images are shifted up, down, left and right to a certain degree to increment this type of data. .

S402: Obtain a sample heat map corresponding to each incrementally processed sample medical image.

S403: Determine each incrementally processed sample medical image and the corresponding sample heat map as training sample data.

After the sample medical images are incrementally processed through the above-mentioned various incremental methods, the sample medical images after each incremental processing and the sample heat maps corresponding to the sample medical image data after each incremental processing are obtained, and each incremental processing is performed. The post-sample medical image data and the corresponding sample heatmap are determined as training sample data.

In this embodiment, sample medical images of various parts are acquired, and each sample medical image is incrementally processed to obtain an incrementally processed sample medical image, and then the incrementally processed sample medical images are compared with the corresponding samples. The heat map is determined as the training sample data. Since the sample data is incremented by some incremental methods, the regression network obtained by training the training data obtained in this way is more robust and can be more accurately determined from the medical image data of each part. Heat map of each part.

In one embodiment, a symmetrical information detection method is also provided, as shown in FIG. 6 , the embodiment includes the following steps:

S1, obtain a sample medical image, and perform incremental processing on the image; execute S2;

S2, obtain the heat map corresponding to each incrementally processed image, the incrementally processed image and the corresponding heat map are training data, and perform S3;

S3, train the initial network with the training data, and obtain the network model after the training is completed; execute S4;

S4, collect the medical image of the target part; execute S5;

S5, preprocess the medical image; execute S6;

S6, input the preprocessed image into the trained network model to obtain a heat map of the symmetry information of the medical image of the target part or the target image; the heat map reflects the distribution of the symmetry of the medical image of the target part or the target image ;Execute S7;

S7: Fit the heat map to obtain the symmetry information of the target part or the medical image of the target image.

The implementation principles and technical effects of the steps in the symmetrical information detection method provided in this embodiment are similar to those in the previous embodiments of the symmetrical information detection method, and are not repeated here. The implementation manner of each step in the embodiment of FIG. 6 is only an example, and each implementation manner is not limited, and the sequence of each step can be adjusted in practical application, as long as the purpose of each step can be achieved.

It should be understood that although the steps in the flowcharts of FIGS. 2-6 are shown in sequence according to the arrows, these steps are not necessarily executed in the sequence shown by the arrows. Unless explicitly stated herein, the execution of these steps is not strictly limited to the order, and these steps may be performed in other orders. Moreover, at least a part of the steps in FIGS. 2-6 may include multiple steps or multiple stages. These steps or stages are not necessarily executed and completed at the same time, but may be executed at different times. The execution of these steps or stages The order is also not necessarily sequential, but may be performed alternately or alternately with other steps or at least a portion of the steps or phases within the other steps.

In one embodiment, as shown in FIG. 7 , a device for detecting symmetrical information is provided, and the device includes: an acquisition module 10 , a processing module 11 , and a fitting module 12 , wherein,

an acquisition module 10 for acquiring medical image data of the target part;

The processing module 11 is used for inputting the medical image data into a preset network model to obtain a heat map of the symmetry information of the target part or the medical image of the target part; the heat map reflects the symmetry of the target part or the medical image of the target part distribution;

The fitting module 12 is used for fitting the heat map to obtain the symmetry information of the target part or the medical image of the target part.

In one embodiment, the above-mentioned obtaining module 10 includes:

an acquisition unit for acquiring the original medical image of the target part;

The preprocessing unit is used for preprocessing the original medical image to obtain the medical image data of the target part.

In one embodiment, the above-mentioned preprocessing includes one of adjusting the resolution of the original medical image, normalizing the gray value of each pixel in the original medical image, and cutting out the region of the target portion from the original medical image, or several.

In one embodiment, the apparatus further includes:

The training sample acquisition module is used to acquire the training sample data; the training sample data includes the medical image data of each part and the heat map corresponding to the medical image data of each part;

The training module is used to train the initial network according to the training sample data, and the network model is obtained after the training is completed.

In one embodiment, the above-mentioned training sample acquisition module includes:

The incremental processing unit is used to obtain sample medical images of various parts, and perform incremental processing on each sample medical image to obtain the sample medical images after incremental processing;

an acquisition unit, configured to acquire a sample heat map corresponding to each incrementally processed sample medical image;

A determination unit, configured to determine each incrementally processed sample medical image and the corresponding sample heat map as training sample data.

In one embodiment, the above-mentioned incremental processing unit is specifically configured to perform incremental processing on each sample medical image through a preset incremental method; the incremental method at least includes data cropping, data rotation, adding Gaussian noise, and data translation. one of them.

In one embodiment, the above-mentioned medical image data includes two-dimensional medical image data or three-dimensional volume data.

For the specific limitation of the symmetrical information detection device, reference may be made to the above limitation on the symmetrical information detection method, which will not be repeated here. Each module in the above-mentioned symmetrical information detection apparatus may be implemented in whole or in part by software, hardware and combinations thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

In one embodiment, a computer device is provided, the computer device may be a terminal, and its internal structure diagram may be as shown in FIG. 1 above. The computer equipment includes a processor, memory, a communication interface, a display screen, and an input device connected by a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium, an internal memory. The nonvolatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the execution of the operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for wired or wireless communication with an external terminal, and the wireless communication can be realized by WIFI, operator network, NFC (Near Field Communication) or other technologies. The computer program, when executed by the processor, implements a symmetrical information detection method. The display screen of the computer equipment may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment may be a touch layer covered on the display screen, or a button, a trackball or a touchpad set on the shell of the computer equipment , or an external keyboard, trackpad, or mouse.

Those skilled in the art can understand that the structure shown in FIG. 1 above is only a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer equipment to which the solution of the present application is applied. More or fewer components than shown in the figures may be included, or some components may be combined, or have a different arrangement of components.

In one embodiment, a computer device is provided, including a memory and a processor, a computer program is stored in the memory, and the processor implements the following steps when executing the computer program:

Obtain medical image data of the target part;

Input the medical image data into the preset network model to obtain a heat map of the symmetry information of the target part or the medical image of the target part; the heat map reflects the symmetry distribution of the target part or the medical image of the target part;

Fit the heat map to obtain the symmetry information of the target part or the medical image of the target part.

The implementation principle and technical effect of the computer device provided by the foregoing embodiment are similar to those of the foregoing method embodiment, and details are not described herein again.

In one embodiment, a computer-readable storage medium is provided on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

Obtain medical image data of the target part;

Input the medical image data into the preset network model to obtain a heat map of the symmetry information of the target part or the medical image of the target part; the heat map reflects the symmetry distribution of the target part or the medical image of the target part;

Fit the heat map to obtain the symmetry information of the target part or the medical image of the target part.

The implementation principle and technical effect of a computer-readable storage medium provided by the foregoing embodiments are similar to those of the foregoing method embodiments, and details are not described herein again.

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage In the medium, when the computer program is executed, it may include the processes of the above-mentioned method embodiments. Wherein, any reference to memory, storage, database or other media used in the various embodiments provided in this application may include at least one of non-volatile and volatile memory. The non-volatile memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, and the like. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, the RAM may be in various forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM).

The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, all It is considered to be the range described in this specification.

The above-mentioned embodiments only represent several embodiments of the present application, and the descriptions thereof are relatively specific and detailed, but should not be construed as a limitation on the scope of the invention patent. It should be noted that, for those skilled in the art, without departing from the concept of the present application, several modifications and improvements can be made, which all belong to the protection scope of the present application. Therefore, the scope of protection of the patent of the present application shall be subject to the appended claims.

Claims (10)

1. a symmetrical information detection method, is characterized in that, described method comprises: Obtain medical image data of the target part; Inputting the medical image data into a preset network model to obtain a heat map of the target part or the symmetry information of the medical image of the target part; the heat map reflects the target part or the target part The symmetrical distribution of medical images; Fitting the heat map to obtain the target part or the symmetry information of the medical image of the target part. 2. The method according to claim 1, wherein the acquiring medical image data of the target part comprises: acquiring an original medical image of the target site; The original medical image is preprocessed to obtain medical image data of the target part. 3. The method according to claim 2, wherein the preprocessing comprises performing resolution adjustment on the original medical image, normalizing the gray value of each pixel in the original medical image, One or more of the regions of the target part are cut out from the original medical image. 4. The method according to any one of claims 1-3, wherein before inputting the medical image data into a preset network model, the method further comprises: Obtaining training sample data; the training sample data includes a heat map corresponding to the medical image data of each part and the medical image data of each part; According to the training sample data, the initial network is trained, and the network model is obtained after the training is completed. 5. The method according to claim 4, wherein the acquiring training sample data comprises: Obtaining sample medical images of various parts, and performing incremental processing on each of the sample medical images to obtain incrementally processed sample medical images; Obtain the sample heat map corresponding to each incrementally processed sample medical image; Each incrementally processed sample medical image and the corresponding sample heat map are determined as the training sample data. 6. The method according to claim 5, wherein the performing incremental processing on each of the sample medical images comprises: Incremental processing is performed on each of the sample medical images in a preset incremental manner; the incremental manner includes at least one of data cropping, data rotation, adding Gaussian noise, and data translation. 7. The method according to any one of claims 1-3, wherein the medical image data comprises two-dimensional medical image data or three-dimensional volume data. 8. A symmetrical information detection device, wherein the device comprises: an acquisition module for acquiring medical image data of the target part; a processing module, configured to input the medical image data into a preset network model to obtain a heat map of the target part or the symmetry information of the medical image of the target part; the heat map reflects the target part Or the symmetrical distribution of the medical image of the target part; The fitting module is used for fitting the heat map to obtain the target part or the symmetry information of the medical image of the target part. 9. A computer device, comprising a memory and a processor, wherein the memory stores a computer program, wherein the processor implements the method according to any one of claims 1 to 7 when the processor executes the computer program. step. 10. A computer-readable storage medium on which a computer program is stored, characterized in that, when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images