WO2022033598A1

WO2022033598A1 - Breast x-ray radiography acquisition method and apparatus, and computer device and storage medium

Info

Publication number: WO2022033598A1
Application number: PCT/CN2021/112733
Authority: WO
Inventors: 李伟; 梁侃; 唐定车; 施瑞森; 储冬玮; 胡扬; 杨乐; 张娜
Original assignee: 上海联影医疗科技股份有限公司
Priority date: 2020-08-14
Filing date: 2021-08-16
Publication date: 2022-02-17

Abstract

The present application relates to a breast X-ray radiography acquisition method and apparatus, and a computer device and a storage medium. The method comprises: performing a pressure test on a breast part of an object to be subjected to detection, and determining the type of the breast part according to a pressure test result; acquiring an optical image of the breast part, and performing identification processing on the optical image to obtain a quantization parameter of the breast part; and determining an image collection parameter of the breast part according to the type and the quantization parameter of the breast part, and acquiring X-ray radiography of the breast part according to the image collection parameter. By means of the method, a patient can be prevented from receiving a dose of radiation that is too high, and the problem of a relatively low efficiency and a relatively great error caused by a doctor only being able to empirically determine the position of a lesion relative to a whole breast is solved.

Description

Mammographic image acquisition method, apparatus, computer equipment and storage medium

Related applications

This application requires the application on August 14, 2020, the application number is 202010817353.6, and the title is "Image acquisition parameter acquisition method, device, equipment, system and storage medium" and August 27, 2020 The application number is 202010880432.1, The priority of the Chinese patent application titled "Mammary X-ray Image Display Method, Device and Storage Medium", and the application number 202011630482.0 filed on December 31, 2020, titled "A method, device, device for area detection" and storage media". It is hereby incorporated by reference in its entirety.

technical field

The present application relates to the technical field of medical images, and in particular, to a method, apparatus, computer equipment and storage medium for acquiring a mammogram.

Background technique

Breast disease is an important disease that seriously threatens women's health. In recent years, with the development of medical imaging technology, the detection and diagnosis accuracy of breast disease is gradually improving. X-ray image (X-ray Radiography, XR for short) realization.

In the prior art, when collecting X-ray data of a patient, a pre-exposure method is usually used to obtain a grayscale image of the breast region of the patient, and then the grayscale change trend of the breast region on the grayscale image is analyzed to obtain the second time. The relevant acquisition parameters of data acquisition, so that the X-ray machine can be set by the acquisition parameters to realize data acquisition and image reconstruction. However, the above-mentioned techniques have the problem of long imaging time, resulting in excessive radiation dose received by the patient.

Further, after obtaining the X-ray image, the doctor needs to judge the location of the lesion by viewing the X-ray image, but in the related art, the X-ray imaging device can only obtain the partial image of the breast, and the doctor can only judge the X-ray image based on experience. The location of the lesions in the whole breast is less efficient and has larger errors. At present, X-ray imaging equipment in the related art can only obtain partial images of the breast, which leads to low efficiency and large errors in the process of judging the relative position of the lesion and the entire breast, and no effective solution has been proposed yet.

SUMMARY OF THE INVENTION

Based on this, it is necessary to provide a mammographic X-ray image acquisition method, device, computer equipment and storage medium in view of the above technical problems, which can prevent the patient from receiving excessive radiation dose and solve the problem of doctors in judging the relative position of the lesion and the entire breast. In the process, the efficiency is low and the error is large.

In a first aspect, a method for acquiring a mammogram is provided, the method comprising:

Carry out a stress test on the breast part of the object to be tested, and determine the type of breast part according to the stress test results;

Obtain the optical image of the breast part, and identify and process the optical image to obtain the quantitative parameters of the breast part;

Image acquisition parameters of the breast part are determined according to the type of the breast part and quantification parameters, and an X-ray image of the breast part is acquired according to the image acquisition parameters.

A second aspect provides a mammogram display method, the method comprising:

Obtain optical images of the breast as well as X-ray images of the breast area of interest;

According to the optical image, determine the breast contour;

register the optical image with the X-ray image;

The outline of the breast is displayed on the X-ray image, or the X-ray information of the breast area of interest is displayed on the optical image.

In a third aspect, a method for acquiring a mammogram is provided, the method comprising:

Control the compression plate to compress the breast;

Controlling the optical image acquisition unit to acquire an optical image of the breast in the compressed state, and determining the outline of the breast in the optical image;

Controlling the X-ray image acquisition unit to acquire the X-ray image of the breast under compression;

The optical image is registered with the X-ray image, and the breast contour is displayed on the X-ray image or the X-ray information of the breast part of interest is displayed on the optical image.

In a fourth aspect, an area detection method is provided, the method comprising:

According to the pixel gray level of each gray pixel point in the medical image, each gray pixel point is clustered, and the region where each gray pixel point belonging to the same category is located is used as a candidate region;

A gradient image corresponding to the medical image is generated according to each pixel grayscale, and a gradient edge in the gradient image is determined according to the pixel gradient of each gradient pixel point in the gradient image;

A target region is detected from each of the candidate regions according to the degree of coupling between the region boundary of each of the candidate regions and the gradient edge.

In a fifth aspect, a computer device is provided, including a memory and a processor, the memory stores a computer program, and the processor implements the steps of the first, second, third and fourth aspects when executing the computer program.

In a sixth aspect, 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 steps described in the first aspect, the second aspect, the third aspect and the fourth aspect are implemented.

The above-mentioned method, device, computer equipment and storage medium for obtaining mammograms are performed by performing a stress test on the breast part of the object to be detected, and the type of the breast part is determined according to the stress test result, and the breast part can be characterized by obtaining the optical image of the breast part. The quantification parameters of the size are determined according to the type and quantification parameters of the breast part, and the image acquisition parameters used to acquire the medical image of the breast part are determined, and the X-ray image of the breast part is acquired according to the image acquisition parameters, and then the optical image is compared with the X-ray image. Registration is performed, and the breast contour is displayed on the X-ray image based on the registration results. In the embodiment of the present disclosure, since the corresponding medical image acquisition parameters can be determined by the type and quantification parameters of the breast part, when obtaining the medical image of the breast part, only the breast part needs to be exposed through the determined medical image acquisition parameters The corresponding X-ray image can be obtained once. Compared with the traditional X-ray image that requires two imaging to obtain the X-ray image, the time for obtaining the X-ray image in the embodiment of the present disclosure is shorter, and the radiation dose received by the patient is correspondingly higher. so that the patient can be prevented from receiving excessive radiation doses. In addition, by registering the optical image and the X-ray image, the position information of the breast contour in the X-ray image can be directly obtained. Compared with the doctor's experience marking the relative position of the X-ray image in the entire breast, the embodiment of the present disclosure solves the problem. X-ray imaging equipment can only obtain partial images of the breast, so that doctors can only judge the relative position of the lesion and the entire breast based on experience, which is inefficient and has large errors, which improves the diagnosis efficiency and position of doctors in the diagnosis process. The accuracy of the labeling.

Description of drawings

Fig. 1a is one of the application environment diagrams of the mammography image display method in one embodiment;

Fig. 1b is the second application environment diagram of the mammography image display method in one embodiment;

FIG. 2 is one of the schematic flow charts of a mammogram display method in one embodiment;

3 is a second schematic flowchart of a method for displaying a mammogram in one embodiment;

4 is a schematic flowchart of steps of determining the type of breast part according to the test result in one embodiment;

FIG. 5 is a schematic flowchart of a step of obtaining quantitative parameters of a breast part in one embodiment;

FIG. 6 is a schematic flowchart of a step of obtaining image acquisition parameters of a breast part in one embodiment;

7 is a schematic flowchart of steps of determining a breast contour in one embodiment;

8 is a schematic flowchart of a step of registering an optical image and an X-ray image in one embodiment;

9 is a schematic diagram of a detection area in one embodiment;

10 is a schematic flowchart of a step of detecting a target area in one embodiment;

11 is a schematic flowchart of a step of clustering each grayscale pixel point in one embodiment;

12 is a schematic flowchart of a step of detecting a target region from each candidate region in one embodiment;

13a is a schematic diagram of an optional example of a region detection method in one embodiment;

13b is a schematic diagram of an optional example of a region detection method in one embodiment;

FIG. 14 is a structural block diagram of a mammography image acquisition apparatus in one embodiment;

FIG. 15 is a structural block diagram of a mammography image display apparatus in one embodiment;

16 is a structural block diagram of a region detection apparatus in one embodiment;

Figure 17 is a diagram of the internal structure of a computer device in one embodiment.

detailed description

In order to make the objectives, technical solutions and advantages of the present application clearer, 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.

The breast X-ray image display method provided by the embodiments of the present application can be applied to the medical imaging system 10 shown in FIG.

The above-mentioned medical imaging device 11 may be a breast machine. As shown in FIG. 1 b , the breast machine 11 may include a frame 111 , a rotating support 112 , a medical imaging assembly 113 and a compression assembly 114 . The rotating bracket 112 is rotatably connected to the gantry 111 ; the medical imaging component 113 is installed on the rotating bracket 112 and can rotate with the rotating bracket 112 to obtain X-ray images of different angles; the compression component 114 is installed on the gantry 111 , and the compression component 114 It is used to carry and compress a patient's breast to shape the breast into a relatively thin and uniform shape, thereby facilitating high-quality medical images. The frame 111 is used to carry the rotating bracket 112, the medical imaging assembly 113 and the compression assembly 114. The rotating bracket 112 can rotate relative to the frame 111 and drive the medical imaging assembly 113 installed on the rotating bracket 112 to rotate together, thereby enabling medical imaging An angle is formed between the component 113 and the compression component 114, and the medical imaging component 113 can acquire images of the patient's breast at different angles, so that the physician can accurately determine the location of the lesion. The medical imaging assembly 113 includes a radiation source 1131 , a beam limiter and a detector 1132 . The radiation source 1131 and the detector 1132 are respectively disposed at two ends of the rotating bracket 112 , and a shooting area is formed between the radiation source 1131 and the detector 1132 . The compression assembly 114 includes a compression plate 1141, a compression platform 1142, and a drive member.

The medical imaging device 11 is used for collecting medical images of the breast; the radiation source 1131 may be an array X-ray source, or a conventional single emission source. The array X-ray source may adopt a linear array X-ray source and/or an area array X-ray source. Any X-ray source or single emission source in the array X-ray source can be either a field emission X-ray source or a hot cathode X-ray source. The beam limiter is usually arranged in front of the output window of the radiation source 1131 . The detector 1132 detects (acquires) projection data of the X-rays emitted by the radiation source 1131 after passing through the breast part, which is located between the detector 1132 and the compression component 114, and transmits the projection data to the computer device 13 for processing, the detector 1132 can be a flat panel detector, and of course other types of detectors.

The optical imaging device 12 is used to acquire an optical image of the breast part of the object to be detected; for example, it may be an optical camera or the like. The optical imaging device 12 may transmit the acquired optical images to the computer device 13 for processing.

The computer device 13 may be a server, which may be implemented by an independent server or a server cluster composed of multiple servers. Of course, it can also be a terminal, which can be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, portable wearable devices, and the like.

It should be noted that the execution subject of the following embodiments of the present application may be a computer device or a medical imaging system, and the computer device will be used as an example for description below.

In one embodiment, a method for displaying a mammary X-ray image is provided. This embodiment involves acquiring an optical image and an X-ray image of a breast part, and displaying the X-ray image on the X-ray image according to the registration result of the optical image and the X-ray image. The specific process of showing breast contours. As shown in Figure 2, the method includes the following steps:

S201, perform a stress test on the breast part of the object to be detected, and determine the type of the breast part according to the stress test result.

Wherein, the breast part here can be both breasts or one breast. When performing the stress test in this embodiment, the description is mainly for one breast, and the other breast can be performed after the imaging of the breast is completed. Same operation.

During a stress test, the breast area can be placed between the compression plate and the detector. The computer equipment can control the movement of the driving part by outputting electrical signals from the software. The driving part applies pressure to the compression plate to drive the movement of the compression plate. The compression plate can move up and down, or move left and right, or move back and forth. The movement of the compression plate can change the thickness of the breast area between the compression plate and the detector. At the same time, the force of the compression plate during the movement of the compression plate can also be known, so that the force of the compression plate and the thickness change of the breast part can be analyzed to obtain the type of the breast part. Optionally, the types of breast parts here may include fat type, dense type, and of course other types, which can be set according to actual conditions.

S202 , an optical image of the breast part is acquired, and the optical image is identified and processed to obtain a quantitative parameter of the breast part; the quantitative parameter is used to characterize the size of the breast part in a compressed state.

Wherein, the timing of acquiring the optical image of the breast part in this step may be after the stress test is performed, that is, after the execution of S201. Whether the stress test is completed here can be characterized by a predetermined period of time, for example, pressure is continuously applied to the breast region through the compression plate within the predetermined period of time, and when the preset period of time is reached, the stress test is considered to be completed. Alternatively, the completion of the stress test can also be characterized by the compression thickness of the breast part. For example, when the breast part is continuously pressed by the compression plate, when the thickness of the breast part is compressed to a certain thickness, the stress test can be considered complete. Of course, other characterization methods are also possible, in short, the results of whether the stress test is completed can be obtained.

Of course, the timing of acquiring the optical image of the breast part in this step can also be before S201, that is, the breast part can be compressed in advance, so that the breast part is in a compressed state, and when the breast part is compressed to a certain thickness, the breast part can be obtained in this Optical image under compression.

In a word, whether after the above stress test is completed or before the stress test, the breast will be compressed to a relatively thin thickness. At this time, the breast is thin and uniform, so that the overlapping soft tissue in the breast structure is easier to separate. In order to be able to obtain high-quality medical or optical images of the breast region.

Specifically, when the breast part is in a compressed state, an optical imaging device can be used to capture an image of the breast part to obtain an optical image of the breast part, and then the optical imaging device can transmit the optical image to the computer device, so that the computer device can An optical image of the breast area can be obtained. Wherein, the optical imaging device may include a camera, and the camera may be a 2D or 3D camera.

After that, the computer equipment can identify the optical image through the image recognition algorithm, identify the breast part in it, and obtain the contour and shape of the breast part, and then analyze and calculate the contour and shape of the breast part to obtain the breast part. The size, volume, area and other parameters of the breast are called quantitative parameters of the breast part.

Wherein, the image recognition algorithm can be an algorithm based on geometric features, a template-based algorithm, a model-based algorithm, etc.; for example, a model-based algorithm can be a recognition algorithm based on a neural network model.

Taking a model-based algorithm as an example, the model may be a neural network model, and before using the neural network model for identification, the neural network model may be trained.

The training process may include: acquiring a training image set, each training image in the training image set is an optical image of a breast part, and each training image includes labeling position information of the breast; after that, each training image in the training image set may be The image is used as the input of the initial neural network model, and the labeled position information of the breast corresponding to each training image is used as the reference output of the initial neural network model, and the initial neural network model is trained to obtain a trained neural network model.

After the neural network model is trained, the optical image of the breast part under test can be input into the neural network model to obtain the position information of the breast on the optical image of the test, then the contour and shape of the breast can be obtained. information.

S203, according to the type of the breast part and the quantification parameter, determine the image acquisition parameter of the breast part, and acquire the X-ray image of the breast part according to the image acquisition parameter.

The image acquisition parameters may be parameters set for components in the medical imaging system when collecting medical images, for example, may include setting parameters of a ray source, setting parameters of a beam limiter, and the like.

After obtaining the type of the breast part and the relevant quantitative parameters, the image acquisition parameters of the breast part can be obtained by means of table look-up, calculation and the like. It may include setting parameters of the ray source, setting parameters of the beam limiter, etc., then the ray source can be set according to the setting parameters of the ray source, and the beam limiter can be set according to the setting parameters of the beam limiter.

After the parameters of the ray source and the beam limiter are set, the ray source and the beam limiter with the set parameters can be used to emit X-rays to the breast part, and the data can be collected by the detector, and the collected data can be collected by Image reconstruction, an X-ray image of the breast area can be obtained.

In the above embodiment, a stress test is performed on the breast part of the object to be detected, and the type of the breast part is determined according to the result of the stress test, and a quantitative parameter that can characterize the size of the breast part is obtained through the optical image of the breast part. The quantification parameter determines the image acquisition parameters used to acquire the medical image of the breast part, and obtains the X-ray image of the breast part according to the image acquisition parameters, then, the optical image is registered with the X-ray image, and the X-ray image is registered according to the registration result. The outline of the breast is shown on the image. In the embodiment of the present disclosure, since the corresponding medical image acquisition parameters can be determined according to the type and quantification parameters of the breast part, when a medical image of the breast part needs to be obtained, it is only necessary to use the determined medical image acquisition parameters for the breast part. The corresponding X-ray image can be obtained with one exposure. Compared with the traditional X-ray image that requires two imaging to obtain the X-ray image, the time for obtaining the X-ray image in the embodiment of the present disclosure is shorter, and the radiation dose received by the patient is also reduced accordingly. less, so that the patient can be prevented from receiving excessive radiation doses.

In one embodiment, as shown in FIG. 3 , on the basis of the foregoing embodiment, the embodiment of the present disclosure may further include:

S204 , register the optical image with the X-ray image, and display the breast contour on the X-ray image or display mammography image information on the optical image according to the registration result.

Among them, registration refers to matching images collected by different imaging methods in the same coordinate system, and the specific matching methods include geometric correction, projection transformation and unified scale.

After the optical image and the X-ray image are acquired, the position information of the breast in the X-ray image can be determined through image registration. Optionally, the position information may be coordinate information of the breast contour in the X-ray image. In one example, the imaging area of the X-ray image is small, and only the position of the breast lesion is captured, so the outline of the breast can be displayed in the X-ray image according to the position information.

After the optical image and the X-ray image are acquired, the position information of the breast in the optical image can also be determined by image matching, and the outline of the breast can be displayed in the optical image according to the position information. Optionally, the position information may be coordinate information of the breast contour in the optical image.

In the above embodiment, the optical image and the X-ray image are registered, and the position information of the breast contour in the X-ray image can be directly obtained, and the relative position of the X-ray image in the whole breast is marked by the doctor according to the experience. The embodiment solves the problem that when the doctor obtains the information of the breast part that needs to be irradiated through experience (such as the way of asking and touching the breast), after setting the beam limiter opening to an appropriate size, and only obtaining the partial image of the breast through the X-ray imaging device, the doctor only needs to The relative position of the lesion and the entire breast can be judged based on experience, which has low efficiency and large error, which improves the doctor's diagnostic efficiency and the accuracy of position labeling during the diagnosis process.

In one embodiment, a specific process of stress testing a breast area and determining the type of breast area based on the test results is involved. On the basis of the foregoing embodiment, as shown in FIG. 4 , the foregoing S201 may include the following steps:

S2011, use the compression component to perform a pressure test on the breast part, and obtain the changing relationship between the compression force generated during the pressure test and the compression thickness of the breast part; the compression thickness of the breast part is the thickness of the breast part under the action of the compression force .

During the stress test on the breast part, when the compression plate presses the breast part, a pressure force will be generated. When different pressure forces are used to press the breast part, the thickness of the breast part will also change.

It is understandable that during the stress test, by using different pressures to compress the breast, the thickness of the breast after being compressed under different compression forces can be obtained, which is recorded as the compression thickness. The change trend of the compression thickness is recorded as the relationship between the compression force and the compression thickness.

Of course, the changing relationship between the pressing force and the pressing thickness can be represented by a curve. For example, the horizontal axis of the coordinate axis is the pressing thickness, and the vertical axis is the pressing force. By filling the pressing thickness corresponding to different pressing forces under the coordinate axis The corresponding position of the compressive force and the compressed thickness can be obtained by connecting lines, or linear fitting, or curve fitting to each point of the filling.

S2012, determining the type of the breast part according to the changing relationship between the compression force and the compression thickness.

After obtaining the variation curve between the compression force and compression thickness, that is, after the variation relationship, optionally, the variation relationship includes the variation curvature, then the variation curvature can be compared with the preset curvature threshold; if the variation curvature is greater than the curvature threshold , the type of the breast part is determined to be fat; otherwise, the type of breast part is determined to be dense.

That is to say, after obtaining the change curve between the compression force and compression thickness of the breast part, the curvature of any point on the curve can be calculated, and the calculated curvature of a point can be compared with the curvature threshold. When the calculated curvature is greater than When the curvature threshold is set, it can be considered that the compression force and compression thickness of the breast part change rapidly, then the type of the breast part can be considered to be fat type (here, the fat type can be understood as more fat in the breast part, and under the same compression force, its thickness is thinner).

Of course, if the calculated curvature is not greater than the curvature threshold, that is, less than or equal to the curvature threshold, it can be considered that the compression force and compression thickness of the breast part change slowly, then the type of the breast part can be considered to be dense (the dense type in It can be understood here that there is less fat in the breast area, and the thickness is thicker under the same pressure).

It should be noted that the classification types of fat type and dense type here are just an example, of course, there may be more classification types, for example, classification can be performed according to how much the curvature is greater than the curvature threshold.

In the above embodiment, the variation relationship between the compression force and compression thickness on the breast part during the stress test is obtained, and the type of the breast part is obtained through the variation relationship. Through the embodiment of the present disclosure, the type of the breast part can be determined through the changing relationship between the compression force and the compression thickness, the determination method is relatively simple, and the obtained type result is relatively accurate.

In one embodiment, it involves a specific process of identifying optical images to obtain quantitative parameters of the breast region. On the basis of the foregoing embodiment, as shown in FIG. 5 , the foregoing S202 may include the following steps:

S2021, performing identification processing on the optical image to obtain a breast contour; the breast contour includes the positions of multiple points.

S2022, perform mathematical operation processing on the positions of each point on the breast contour to obtain the volume of the breast part; or, perform mathematical operation processing on the positions of each point on the breast contour to obtain the projected area of the breast part on the compression device.

In this embodiment of the present disclosure, the image recognition algorithm in S202 above may be used for the recognition processing, which will not be repeated here.

It should be noted that the optical images here are also acquired after the stress test is completed and the breast area is in a thin and uniform state.

Specifically, by using an image recognition algorithm to recognize the obtained optical image, the breast contour of the breast part can be identified, and the position information of each point on the breast contour can be obtained at the same time.

After the position information of each point on the breast contour is obtained, the length and width of the breast part can be calculated according to the bottommost edge in the horizontal direction of the breast contour, and the vertical direction is used as the height direction of the breast part to calculate the height of the breast part, and calculate the height of the breast part. Multiply the calculated length, width and height to get the volume of the breast part.

In addition, the projected area of the breast part on the compression assembly refers to the projected area of the breast part on the compression plate in the compression assembly. Usually, when the breast is compressed, the compression plate is generally perpendicular to the vertical direction of the human body, that is, the projected area here is the projected area of the breast in the horizontal direction.

Then, after obtaining the volume of the breast part above, since the volume of the breast part is obtained by calculating the length, width and height above, then when calculating the projected area in the horizontal direction, the above volume can be used to calculate the middle length and width, Get the projected area.

It should be noted that the projected area here may be calculated when the breast part is in a compressed state and the breast part is in a thin and uniform state.

In the above embodiment, the position of each point on the breast contour is obtained by identifying the optical image, thereby calculating the volume of the breast part or the projected area of the breast part on the compression device. In the embodiment of the present disclosure, the volume or projected area of the breast part is calculated based on the contour position of the breast part, so that the calculated volume or projected area of the breast part is more in line with the actual situation and more accurate.

In one embodiment, a possible implementation of obtaining image acquisition parameters of a breast part based on the type and quantification parameters of the breast part is involved. On the basis of the foregoing embodiment, the foregoing S203 may include the following steps:

Step 1: Determine the radiation field corresponding to the breast part according to the quantitative parameter.

Step 2: Determine the image acquisition parameters of the breast part according to the type of the breast part and the radiation field corresponding to the breast part.

The quantification parameters may include parameters such as size, volume, and area of the breast part. After obtaining the quantitative parameters, the size, volume, or area of the breast part can be directly used as the radiation field of the breast part, that is, the size of the radiation range of the breast part.

After that, the radiation field of the breast can be used as the setting parameter of the beam limiter to set the beam limiter. At the same time, the image acquisition parameters of the breast part can also be obtained by means of table look-up, calculation and the like. It may include setting parameters of the ray source, setting parameters of the beam limiter, etc., then the ray source can be set according to the setting parameters of the ray source, and the beam limiter can be set according to the setting parameters of the beam limiter.

In the above embodiment, the radiation field of the breast part is obtained by quantifying parameters such as the size, area, and volume of the breast part, and the image acquisition parameters of the breast part are determined according to the type of the breast part and the radiation field. Through the embodiments of the present disclosure, the radiation field and image acquisition parameters of the breast can be obtained relatively simply and accurately, so that the object to be detected can be prevented from receiving unnecessary radiation as much as possible during the medical imaging process.

In one embodiment, the above-mentioned image acquisition parameters include a first image acquisition parameter and a second image acquisition parameter; the embodiment of the present disclosure relates to a method for obtaining the image acquisition parameters of the breast part according to the type and quantification parameter of the breast part. implementation. On the basis of the foregoing embodiment, as shown in FIG. 6 , the foregoing S203 may include the following steps:

S2031, according to the volume or projected area of the breast part, obtain the radiation field corresponding to the breast part.

S2032: Determine the radiation field corresponding to the breast part as the first image acquisition parameter.

After obtaining the volume of the breast part or the projected area of the breast part on the compression plate, taking the projected area as an example, the projected area can be used as the size of the imaging area during X-ray imaging of the breast part, that is, when the breast part is imaged. The radiation field of the X-rays received by the breast during X-ray imaging. The radiation field is usually expressed according to the opening size of the beam limiter, so the projected area here is also the opening size of the beam limiter. For example, it can be 18*25cm, etc.

S2033, according to the type of the breast part, determine the second image acquisition parameter.

First, the actual compression thickness of the breast part under a preset condition is obtained; the preset condition is related to the tolerance of the object to be detected to the compression force. Then, according to the type of the breast part and the actual compression thickness, the second image acquisition parameters are determined.

When actually performing X-ray image imaging on the breast of the object to be detected, taking into account the tolerance of the individual to be detected to the compression force, the breast can be obtained when the object to be detected is in a critical state of resistance to compression force. The compression thickness of the part is determined, and the compression thickness at this time is regarded as the compression thickness when the X-ray imaging is actually performed, that is, the actual compression thickness.

After the actual compression thickness of the breast part and the type of the breast part are obtained, optionally, according to the type of the breast part and the actual compression thickness, a second mapping table corresponding to the type of the breast part and the actual compression thickness can be obtained in the preset mapping table Image acquisition parameters; wherein, the mapping table includes the types and compression thicknesses of multiple groups of parts, and the second image acquisition parameters corresponding to the types and compression thicknesses of each group of parts; the second image acquisition parameters are used to characterize the power of the ray source and beam limiter filtering.

That is to say, a mapping table can be preset, which includes different types (ie, types of breast parts) and the corresponding relationship between different compression thicknesses, powers of radiation sources, and filtering methods of the beam limiter. Wherein, the power of the ray source can be expressed by current and voltage, for example, it can be expressed by KV and mas.

For example, the mapping table can be referred to as shown in Table 1 below:

Then, after the actual compression thickness and the type of breast part are obtained above, the power of the ray source and the filtering method of the beam limiter corresponding to the actual compression thickness and the type of breast part can be obtained by looking up the table.

After that, you can use the power parameters of the ray source obtained by looking up the table to set the voltage and current of the ray source, and you can also use the filtering method obtained by looking up the table to set the filtering of the beam limiter. After setting the ray source and the beam limiter, you can use the ray source and the beam limiter with the set parameters to emit X-rays to the breast, and perform data acquisition, and by reconstructing the collected data, you can get X-ray image of the breast area.

In the above embodiment, the radiation field of the breast is obtained by the volume of the breast or the projected area on the compression device, and the power of the radiation source and the filtering method of the beam limiter are obtained by the type of the breast and the actual compression thickness. . In the embodiment of the present disclosure, the radiation field of the breast is obtained through a specific volume or projected area, and the radiation field obtained in this way is relatively accurate, so that the object to be detected can be prevented from receiving unnecessary radiation as much as possible during the X-ray imaging process. In addition, the power of the radiation source and the filtering method of the beam limiter are determined by the type of the breast part and the actual compression thickness, which can further refine the imaging parameters and further reduce the radiation dose of the object to be detected.

In one embodiment, the process involving identifying and processing an optical image to obtain a breast contour may include: identifying and processing the optical image to obtain the pixel value of each pixel in the optical image; A segmentation algorithm determines the breast contour. Among them, the optical image is an optical image; the image segmentation algorithm is the technology and process of dividing the captured image into several specific regions with unique properties, and extracting the target of interest, which is a key step from image processing to image analysis. . Image segmentation algorithms mainly include: threshold-based segmentation algorithms, region-based segmentation algorithms, edge-based segmentation algorithms, and segmentation algorithms based on specific theories, etc. Based on the image segmentation algorithm, the embodiment of the present disclosure can segment the breast from other regions to obtain the breast contour. Further, the doctor can obtain the position information of the breast in the X-ray image more clearly according to the breast contour, which further improves the diagnosis efficiency. and location labeling accuracy.

In one embodiment, it involves the process of determining the breast contour through an image segmentation algorithm according to the pixel value of each pixel point. As shown in FIG. 7 , on the basis of the foregoing embodiments, the embodiments of the present disclosure may include the following steps:

S301, obtaining a segmentation threshold.

A threshold-based segmentation algorithm is used to segment the optical image to obtain the breast contour, and the segmentation threshold in this embodiment can be obtained according to the grayscale feature of the optical image.

S302, according to the segmentation threshold and the pixel value of each pixel point, segment the optical image into a breast region image and a background region image.

After the segmentation threshold is obtained, the gray value of each pixel is obtained through the pixel value of each pixel in the optical image, and the gray value is compared with the segmentation threshold. The gray value is greater than the segmentation threshold. Values less than or equal to the segmentation threshold are classified into another category.

Specifically, the X-ray image of the breast is assisted by a detector, and the detector converts the light signal and the electrical signal to make the image quality of the X-ray image clearer. Usually, the surface of the detector is dark, such as black, while the color of human tissue is lighter, so after the optical image is acquired, the optical image can be segmented according to the color difference between the detector and the breast, so as to obtain Breast contour. In the optical image of the embodiment of the present disclosure, pixels with a grayscale value greater than a segmentation threshold are classified as a background area image, and pixels with a grayscale value less than or equal to the segmentation threshold are classified as a breast area image.

S303, taking the contour of the breast region image as a breast contour.

The embodiment of the present disclosure adopts the segmentation algorithm based on segmentation threshold to segment the optical image, the calculation method is simple, the amount of data to be calculated is small, and the calculation efficiency is high, which can effectively improve the segmentation efficiency of the optical image and further improve the diagnosis efficiency.

In one embodiment, a process involving registering an optical image with an X-ray image may include: extracting a plurality of first features in the optical image and a plurality of second features in the X-ray image through a feature extraction algorithm, according to The first feature and the second feature register the optical image with the X-ray image. Among them, the feature extraction algorithm is used to extract the image features in the optical image and the X-ray image. The feature can be a point, line or area in the breast region image, the first feature in the optical image and the second feature in the X-ray image. corresponds to the same features in the breast region image. Since the first feature belongs to the optical image, the second feature belongs to the X-ray image, and the first feature and the second feature correspond to the same features of the breast region image, so through the correspondence between the first feature and the second feature, the The optical image is registered with the X-ray image. The embodiment of the present disclosure extracts the first feature in the optical image and the second feature in the X-ray image based on the feature extraction algorithm, and realizes the matching between the optical image and the X-ray image through the correspondence between the first feature and the second feature The accuracy of the judgment of the position information of the breast in the X-ray image is improved.

In one embodiment, a process of registering an optical image with an X-ray image is involved. On the basis of the foregoing embodiments, as shown in FIG. 8 , the embodiments of the present disclosure may include the following steps:

S401, perform feature matching on the first feature and the second feature through similarity measurement and cluster analysis.

The similarity measure is used to evaluate the similarity between the first feature and the second feature. The closer the first feature is to the second feature, the greater the similarity measure value between the first feature and the second feature, the more distant the first feature and the second feature are, and the smaller the similarity measure value.

Optionally, the similarity measurement algorithm includes: a correlation coefficient algorithm, a similarity coefficient algorithm, a sample matching coefficient algorithm or a sample matching consistency degree algorithm, and the like. Cluster analysis is used to classify data and is an unsupervised learning process. The algorithms of cluster analysis include systematic clustering, decomposition, joining, dynamic clustering, ordered sample clustering, and overlapping clustering. class and fuzzy clustering, etc.

S402: Obtain a coordinate mapping function according to the coordinates of the first feature in the optical image and the coordinates of the second feature in the X-ray image. The coordinate mapping function is obtained according to the correspondence between the coordinates of the first feature and the coordinates of the second feature.

S403, according to the coordinate mapping function, register the optical image and the X-ray image.

The coordinate mapping function can realize the coordinate conversion between the pixels of the optical image and the X-ray image, and then realize the registration of the optical image and the X-ray image.

The embodiment of the present disclosure realizes the registration between the optical image and the X-ray image based on the coordinate transformation, and further improves the judgment accuracy of the position information of the breast in the X-ray image.

In one embodiment, it involves a process of displaying a breast contour on an X-ray image, which may include: obtaining position information of the breast contour in the X-ray image according to the breast, its optical image and the X-ray image, and in the X-ray image The outline of the breast is shown on the image. By registering the optical image and the X-ray image, the X-ray image and the optical image can be superimposed to the same coordinate system. Therefore, after the X-ray image and the optical image are registered, it is possible to more accurately obtain the image of the breast shooting area in the X-ray image. The position information in the X-ray image can be used to display the breast contour on the X-ray image according to the position information. During the diagnosis process, the doctor can also make a more accurate determination of the position of the lesion.

In another embodiment, after the X-ray image and the optical image are registered, the mammogram in the X-ray image may be superimposed on the optical image.

In one embodiment, acquiring the coordinate mapping function also requires camera parameters, where the camera is used to acquire the optical image, and the camera parameters of the camera include camera intrinsic parameters and camera extrinsic parameters, specifically, the camera intrinsic parameters include the focal length of the camera and the image plane. The center point is offset, and the camera extrinsic parameters include the installation position information and pitch angle of the camera.

In one embodiment, it involves the process of obtaining the position information of the breast contour in the X-ray image according to the breast, its optical image and the X-ray image, which may include: determining the position information of the breast in the optical image according to the optical image , where the position information may specifically be coordinate information; according to the position information of the breast in the optical image and the registered X-ray image, the position information of the breast in the X-ray image is obtained, and the registered X-ray image is obtained. Coordinate correspondence with the optical image can be achieved, and through the coordinate correspondence, the accuracy of the acquired position information of the breast in the X-ray image can be improved.

In one embodiment, as shown in FIG. 9 , the area enclosed by the solid line frame is the detector area and the imaging area of the optical image, and the dashed line is the breast contour obtained by the image segmentation algorithm. The imaging area is divided into the background area image and the breast area image. In the breast area image, the rectangle enclosed by the dotted line is the compression plate area and the imaging area of the X-ray image. Due to the existence of the breast contour, the doctor can obtain the position information of the breast in the X-ray image in FIG. 9 , thereby improving the diagnostic efficiency and the accuracy of judging the location of the lesion.

In one embodiment, the embodiments of the present disclosure may further include:

Step 1: Control the compression plate to compress the breast.

Step 2, controlling the optical image acquisition unit to collect the optical image of the breast in the compressed state, and determine the breast contour in the optical image;

Wherein, the optical image is an optical image including at least part of the breast contour. The optical image acquisition unit may be a camera, and the completeness of the breast contour in the optical image is determined according to the detection range of the patient by the doctor. Therefore, the breast contour in the optical image may be a complete breast contour or only a partial breast contour.

Step 3, controlling the X-ray image acquisition unit to acquire the X-ray image of the breast in the compressed state, wherein the X-ray image acquisition unit may be an X-ray imager.

Step 4, register the optical image with the X-ray image, and display the outline of the breast on the X-ray image or display the X-ray information of the breast part of interest on the optical image.

Specifically, by registering the X-ray image with the optical image, the position of the breast contour in the X-ray image can be acquired, thereby displaying the breast contour in the X-ray image. It is also possible to acquire X-ray information of the breast part of interest, thereby displaying the X-ray information of the breast part of interest on the optical image.

In the above-mentioned embodiment, during the breast detection process of the patient, the doctor can directly see the outline of the breast on the X-ray image, which solves the problem that the X-ray imaging device can only obtain the partial image of the breast, so that the doctor can only judge the lesion based on experience. The relative position of the whole breast has low efficiency and large error, which improves the diagnosis efficiency and the accuracy of position labeling in the diagnosis process of doctors.

Low-gray areas with too low gray values and/or high-gray areas with too high gray values often appear in X-ray images; the low-gray areas may come from implants in the subject (eg Artificial joints, stents, pacemakers, plates, screws, etc.), fixation devices (such as external fixation devices) used to fix the subject during surgery, positioning devices used to locate lesions (such as positioning pins, clips, etc.) ), imaging devices in some medical devices (such as needle-holding devices in breast puncture images, etc.), etc., the low-gray area can also be called a high-attenuation area; The area through the body, which may also be referred to as the direct exposure area.

Typically, areas of high attenuation and direct exposure in X-ray images are not areas of concern for doctors, and they interfere with the imaging of normal human tissue. Specifically, the problems and phenomena fed back from the clinic show that if there are high attenuation areas and/or direct exposure areas in the X-ray image, this will directly affect the display effect of normal human tissue under the default window width and window level. Therefore, after obtaining the X-ray image of the breast part according to the image acquisition parameters, as shown in FIG. 10 , the embodiment of the present disclosure may further include the following steps:

S501 , cluster each grayscale pixel point according to the pixel grayscale of each grayscale pixel point in the medical image, and use the region where each grayscale pixel point belonging to the same category is located as a candidate region.

The medical image may be an image obtained after image acquisition of human tissue based on medical imaging technology, such as X-ray image, Computed Tomography (CT) image, Digital Radiography (DR) image, B-scan images, etc.

For the imaging object in the medical imaging process, it can include human tissue and other imaging objects except human tissue. Usually, the attenuation degree of rays involved in X-ray imaging by human tissue and other imaging objects There are differences, the higher the attenuation degree, the lower the pixel gray level in the medical image; the lower the attenuation degree, the higher the pixel gray level in the medical image, so the imaging area of the remaining imaging objects in the medical image can be used as The target area to be detected, the target area can be a low-gray area (that is, a high-attenuation area) where the pixel grayscale is significantly lower than that of human tissue, or a high-gray area (that is, a direct exposure area) where the pixel grayscale is significantly higher than that of human tissue. , or a combination of the two, etc.

The medical image is a grayscale image, the grayscale pixel is the pixel in the medical image, and the pixel grayscale is the pixel value (ie, grayscale value) of the grayscale pixel. The grayscale pixels are clustered according to the pixel grayscale of each grayscale pixel, that is, the grayscale pixels that are relatively similar in pixel grayscale are classified into the same category, and the grayscale pixels that are relatively different in grayscale are classified into the same category. Large gray-scale pixels are classified into different categories, and there are many ways to implement clustering, such as LBG clustering algorithm, k-means clustering algorithm, etc., which are not specifically limited here. After clustering each gray-scale pixel point, each gray-scale pixel point can be clustered into the corresponding category. At this time, the region where each gray-scale pixel point belonging to the same category is located can be used as a candidate region. In other words, each gray pixel point belonging to the same candidate region is a gray pixel point belonging to the same category, and these gray pixel points have strong similarity in pixel gray level. It should be noted that the number of candidate regions formed by gray-scale pixels belonging to the same category may be one, two or more, which is not specifically limited here.

On this basis, optionally, if each grayscale pixel is directly clustered, it may appear that grayscale pixels belonging to the same category cannot form a candidate area because they are scattered points, and there are holes in the formed candidate area. (that is, most of the grayscale pixels in a certain area belong to the same category and very few grayscale pixels do not belong to this category) and so on. On this basis, in order to ensure that the candidate region can be obtained after clustering and the obtained candidate region is a complete region, the following optional processing methods can be performed: noise reduction, smoothing, etc. are performed on the medical image before clustering to remove the medical image. The noise in the image, or the region filling of the candidate regions obtained after clustering, etc., are not specifically limited here.

S502 , a gradient image corresponding to the medical image is generated according to the grayscale of each pixel, and a gradient edge in the gradient image is determined according to the pixel gradient of each gradient pixel point in the gradient image.

Among them, the gradient image is an image corresponding to the medical image generated by performing gradient operation on the pixel grayscale of each grayscale pixel in the medical image. The gradient operation process can be understood as the gradient involved in the edge extraction process of the medical image. The calculation process of information, such as the calculation process of gradient information in the edge extraction process based on local variance, sobel, prewitt, canny, etc., exemplarily, take the calculation of the local variance in the medical image to generate the gradient image as an example, obtain the preset value. A sliding window with a window size of W*H, the sliding window is continuously moved, the variance of the pixel gray levels of all gray pixels in the sliding window after each movement is calculated in turn, and the variance is used as a certain value in the sliding window. The pixel gradient (ie pixel value) of the gradient pixel point corresponding to the grayscale pixel point, the gradient pixel point is the pixel point in the generated gradient image, that is, each gradient pixel point in the gradient image has the same pixel position as the pixel point. The corresponding unique grayscale pixel points, thereby generating a gradient image corresponding to the medical image.

The gradient edge is an edge composed of multiple gradient pixels in the gradient image. The gradient pixels on the gradient edge can be called edge pixels, and the edge pixels can be gradient pixels with larger pixel gradients. The meaning of the edge setting is that there is a strong grayscale difference in the pixel grayscales of the grayscale pixels on both sides at the area boundary of the target area of the medical image, and the area boundary is the boundary of the target area, which means that the The pixel gradient of the gradient pixel corresponding to the grayscale pixel is relatively large, that is, the gradient edge is likely to be the region boundary of the target region, then the candidate region where the region boundary similar to the gradient edge is located can be used as the target region.

On this basis, there are many ways to determine the gradient edge in the gradient image according to the pixel gradient of each gradient pixel in the gradient image. For example, the gradient image can be binarized to obtain a binary image, and the binary image The edge formed by each binary pixel with the median pixel value of 1 is used as the gradient edge; for another example, taking the high attenuation area as an example, the edge pixels on the gradient edge corresponding to the high attenuation area are mostly gray with the smaller area. It corresponds to the grayscale pixels in the candidate area of the degree of The gray level area is selected from each candidate area. The area attribute can represent the overall level of pixel gray level of each gray pixel point in the target area, which can reflect whether the target area is a direct exposure area or a high attenuation area. , and then filter out the edge pixels from the gradient pixels corresponding to the pixels of each region in the gray-scale category area, and then use the edge formed by each edge pixel as the gradient edge; etc., no specific limitation is made here. .

S503, according to the coupling degree between the region boundary and the gradient edge of each candidate region, detect the target region from each candidate region.

Wherein, for each candidate region, the coupling degree can be the similarity between the region boundary of the candidate region and the gradient edge. Similarity in pixel location. The coupling degree can be calculated in various ways, such as obtaining the first number of each boundary pixel on the area boundary, and the second number of boundary pixels that can be the same as or similar to a certain edge pixel in pixel position, according to the first number of boundary pixels. The ratio of the numbers between the second number and the first number determines the degree of coupling. Further, the candidate region corresponding to the region boundary with higher coupling degree can be used as the target region. Of course, in addition to considering the coupling degree between the region boundary and the gradient edge, considering that if there is a certain degree of coupling between a region boundary and the gradient edge and the target boundary of the candidate region detected as the target region, such a candidate The region may also be the target region, so it can also be judged whether the candidate region corresponding to the region boundary is the target region according to the coupling degree between the region boundary and the gradient edge, and the coupling degree between the region boundary and the target boundary; and so on, here Not specifically limited.

It should be noted that the reason for jointly selecting the target area according to the gradient edge and the area boundary, rather than directly selecting the target area according to the gradient edge, is that there is likely to be overlap between the imaging object with high attenuation and human tissue, which makes A highly attenuated imaging object has a breakpoint on the gradient edge corresponding to the high-attenuation region in the medical image, which means that the target region cannot be detected by filling the gradient edge with the breakpoint. Correspondingly, the above-mentioned embodiment can obtain a complete candidate region based on the clustering algorithm, and the region boundary and the gradient edge do not need to completely overlap, thus achieving the ability to detect the complete target region even when there is a breakpoint in the gradient edge. Effect.

In the above embodiment, each grayscale pixel is clustered according to the pixel grayscale of each grayscale pixel in the medical image, and the region where each grayscale pixel belonging to the same category is located is used as a candidate region; The gradient image corresponding to the medical image is generated, and the gradient edge in the gradient image is determined according to the pixel gradient of each gradient pixel in the gradient image. The target area is detected in . Through the embodiments of the present disclosure, the clustering of regions solves the problem of low detection accuracy of the target region due to uncertain grayscale differences between the target region and human tissue, image noise, and grayscale transitions within the target region. The effect of accurate detection of the target area is achieved by comparing the complete area boundary of the candidate area thus obtained with the gradient edge that is more likely to be the area boundary of the target area.

In one embodiment, it involves the process of determining the gradient edge in the gradient image according to the pixel gradient of each gradient pixel point in the gradient image, which may specifically include: according to the regional gray level of each candidate area, screening out the candidate areas from each candidate area with the The gray-scale category area corresponding to the regional attribute of the detected target area; the gradient edge in the gradient image is determined according to the pixel gradient of the gradient pixel points in the gradient image corresponding to the pixel points of each area in the gray-scale category area. Among them, considering that the target area can be a high-attenuation area or a direct exposure area, the high-attenuation area is mostly detected from those candidate areas with smaller area grayscales, while the direct exposure area is mostly detected from those with larger area grayscales. Therefore, the gray level area corresponding to the area attribute of the target area to be detected can be selected from each candidate area first, that is, the gray level area is the same as the target area in the pixel gray area. Then, the gradient edge in the gradient image is determined according to the pixel gradient of the gradient pixel points in the gradient image corresponding to the pixel points of each area in the gray-scale category area, for example, considering the target area The pixel gradient of the gradient pixels corresponding to the gray pixels on the region boundary is relatively large, so the edge formed by the gradient pixels with relatively large pixel gradients in the gray level region can be used as the gradient edge. The gradient edge is determined on the pixel gradient corresponding to the gray-scale category area with similar regional attributes to the target area, thereby improving the determination speed and accuracy of the gradient edge; The detection of the direct exposure area is realized on the gray-level candidate area, and the detection of the high-attenuation area is realized on the low-level gray-level candidate area, because both the high-level gray-level candidate area and the low-level gray level candidate area are self-adaptive and unaffected by the dose. As a result, the above embodiments have better adaptability to different doses of medical images and different types of target areas, thereby achieving the effect of accurate detection of target areas.

In one embodiment, considering that the pixel gradient of the gradient pixel points corresponding to the gray-scale pixels on the region boundary of the target area is usually relatively large, the above-mentioned determination of the gradient edge in the gradient image according to the pixel gradient of each gradient pixel point in the gradient image Specifically, the process may include: sorting each gradient pixel point based on the pixel gradient of each gradient pixel point in the gradient image, and screening each gradient pixel point according to the sorting position of each gradient pixel point in the sorting result, Generate a binary image corresponding to the gradient image, that is, keep the gradient pixels with large pixel gradients in the gradient image, and discard the gradient pixels with small pixel gradients in the gradient image, the sorting position can reflect a gradient pixel point is the relative size of the pixel gradient in all gradient pixels; the binary edge formed by each binary pixel in the binary image is used as the gradient edge of the gradient image, and the binary pixel is the pixel of the binary image, Since the pixel value of the binary pixel is 1 or 0, the binary edge formed by each binary pixel with the pixel value of 1 can be regarded as a gradient edge. In the above embodiment, by sorting each gradient pixel point based on the pixel gradient, the gradient pixel points belonging to the gradient edge can be quickly and accurately screened out from each gradient pixel point, and then the gradient based on these gradient pixel points can be obtained. edge, to achieve the effect of accurate determination of gradient edge.

In order to better understand the specific implementation process of the above gradient edge determination, an exemplary description is given below with reference to specific examples. Under normal circumstances, after clustering each grayscale pixel point, the regional grayscale of each candidate area can be obtained. The regional grayscales of the candidate regions belonging to the same category are similar, and the regional grayscales of the candidate regions belonging to different categories can be obtained. The difference is large. Assuming that the number of preset categories is N, 1/2 of which is used as a low grayscale category, and the other 1/2 is used as a high grayscale category, then each candidate can be determined according to the value of the regional grayscale of each candidate area. The category of the region is obtained, thereby obtaining a low grayscale candidate region belonging to the low grayscale category and a high grayscale candidate region belonging to the high grayscale category. When the target area to be detected is a high attenuation area, the low grayscale candidate area at this time is the grayscale category area described above. Based on the pixel gradient of the gradient pixels, sort these gradient pixels, and extract a certain proportion of edge pixels according to the sorting result to generate a binary image corresponding to the gradient sub-image, which is the gradient sub-image in the gradient image. The part of the image corresponding to the gray-scale category area, so that the target area is subsequently detected from the low-gray-level candidate area, and the detection speed and detection efficiency of the target area are improved. Certainly, when the target area to be detected includes the direct exposure area, the corresponding steps can be performed with the high grayscale candidate area as the grayscale category area, which is similar to the execution process of the high attenuation area, and will not be repeated here.

In one embodiment, it involves a process of clustering each grayscale pixel point according to the pixel grayscale of each grayscale pixel point in the medical image, which may include: sorting the pixel grayscale of each grayscale pixel point in the medical image Sort, and determine the gray-scale category split point in the gray-scale sorting result based on the preset number of categories; take the gray-scale category split point as the initial cluster center, based on the cluster center and the pixels of each gray pixel point Grayscale, cluster each grayscale pixel point. On the basis of the foregoing embodiment, as shown in FIG. 11 , the embodiment of the present disclosure may include the following steps:

S601: Sort the pixel gray levels of each gray level pixel point in the medical image, and determine the gray level category split point in the gray level sorting result based on a preset number of categories.

Among them, in the process of clustering each gray pixel point, the gray class split point that can be used as the initial cluster center will have a certain impact on the clustering speed and clustering accuracy. Therefore, in order to improve the clustering As a result, the pixel grayscale of each grayscale pixel point in the medical image can be sorted first, and then the grayscale category splitting point can be determined in the grayscale sorting result based on the preset number of categories. Exemplarily, assuming that a medical image includes 900 grayscale pixels and the number of categories is 90, then the pixel grayscales ranked at the 10th, 20th, ..., 900th can be used as grayscale category splitting points. In practical applications, optionally, since the region boundary obtained by clustering is subsequently delineated by the gradient edge to detect the target region, then when the number of categories is very small, the region boundary is likely to be larger than the gradient edge. It means that the region boundary cannot be delineated based on the gradient edge, that is, the candidate region corresponding to the target region obtained by clustering can be the under-segmentation result of the target region, which facilitates the subsequent traversal of the region boundary of each candidate region to make it The gradient edge is gradually approached, so the number of classes can be a large number.

S602, using the gray-scale category split point as the initial clustering center, clustering each gray-scale pixel point based on the clustering center and the pixel gray level of each gray-scale pixel point, and classifying each gray-scale pixel belonging to the same category The region where the pixel points are located is used as a candidate region.

Among them, the gray level split point can be used as the initial cluster center, and each gray pixel point is clustered based on the cluster center and the pixel gray level of each gray pixel point. For example, for each gray pixel point, Compare the grayscale distance between the grayscale pixel point and each cluster center, and take the category of the cluster center corresponding to the smallest grayscale distance as the clustering result of the grayscale pixel point, that is, the grayscale pixel point It is classified into the category where the cluster center corresponding to the smallest grayscale distance is located, thereby realizing the effect of accurate clustering of each grayscale pixel point.

On this basis, in order to further improve the clustering accuracy, the following scheme can be used for clustering: for each grayscale pixel point, determine the pixel grayscale of the grayscale pixel point and the grayscale distance between each cluster center, and The grayscale pixels are clustered into the category of the cluster center corresponding to the minimum grayscale distance; the grayscale distortion is determined according to the minimum grayscale distance corresponding to each grayscale pixel, and whether the grayscale distortion is satisfied The preset clustering end condition, which may be whether the grayscale distortion is less than a preset threshold, and whether the absolute value of the difference between the grayscale distortion of this iteration and the grayscale distortion of the previous iteration is less than the relative Error threshold, etc.; if not, then for each category, re-determine the cluster center of the category according to the pixel gray level of each gray pixel point belonging to the category after clustering, and repeat the execution of determining the pixels of the gray pixel point The steps of the grayscale and the grayscale distance between each cluster center, until the grayscale distortion satisfies the clustering end condition, the clustering ends.

S603 , a gradient image corresponding to the medical image is generated according to the grayscale of each pixel, and a gradient edge in the gradient image is determined according to the pixel gradient of each gradient pixel point in the gradient image.

S604, according to the coupling degree between the region boundary and the gradient edge of each candidate region, detect the target region from each candidate region.

In the embodiment of the present disclosure, the gray-scale category splitting point (ie, the initial cluster center) is determined in the sorting result of the pixel gray levels of each gray-scale pixel point by using a preset number of categories, and based on the cluster center and each The pixel gray level is used to cluster each gray level pixel point, thereby realizing the effect of adaptive gray level clustering for each gray level pixel point based on a relatively accurate initial cluster center.

In order to better understand the specific implementation process of the above clustering, the following takes the LBG algorithm as an example to illustrate it exemplarily. Exemplarily, the LBG algorithm is a relatively classic algorithm for image compression based on vector quantization, and uses Lloyd iteration to find an optimal solution, which can effectively divide the training vector set. The steps to design vector quantization are as follows: Suppose the training vector set is,

1) Set the initial codebook, B ⁰ ={y ₀ ⁽⁰⁾ ,y ₁ ⁽⁰⁾ ,...y _N-1 ⁽⁰⁾ }, let the number of iterations n=0, the average distortion D ⁰ →∞, and the relative error threshold ε (0≤ε≤1);

2) According to the nearest neighbor condition, take each codeword in the codebook B ⁿ as the centroid (ie, the cluster center), and divide the training vector set into N cells ^Sn = {S ₀ ⁿ , S ₁ ⁿ , . . . .S _N-1 ⁿ }, S _i ⁿ satisfies

3) Calculate the average distortion (that is, grayscale distortion) generated after cell division. The definition of the average distortion of each cell is shown in the following formula, where the calculation process of min is the process of cell division, and y _j ⁿ is the first The centroids of the j cells before the update, D ⁿ is the average value of the grayscale distances between all _xi and the y _j ⁿ of the divided cells;

4) Determine whether the relative error between the average distortion of this iteration and the average distortion of the previous iteration is less than ε;

5) If so, the entire iterative algorithm stops, otherwise, according to the centroid condition, add the values in the corresponding dimensions of all vectors in each cell, and then divide by the number of all vectors in the cell, as the centroid of the cell, using The re-determined centroid (ie codeword) of each cell updates the codebook, let the number of iterations n=n+1, and then jump to step 2).

On this basis, in combination with the application scenarios that may be involved in the embodiments of the present disclosure, each grayscale pixel point can be used as a vector to perform training based on the LBG algorithm. Since the LBG algorithm is relatively dependent on the initial codebook, the pixel gray levels of each gray pixel point can be sorted here, and the gray level split points can be determined in the gray level sorting result based on the preset number of categories, and then the gray level can be sorted This grayscale class split point serves as the initial codebook for LBG iteration.

In one embodiment, the process of detecting the target region from each candidate region according to the coupling degree between the region boundary and the gradient edge of each candidate region for the current region in each candidate region may specifically include: Obtain the first similarity of each current pixel in the current boundary of the current area and each edge pixel in the gradient edge at the pixel position, and each current pixel and each target pixel in the target boundary of the candidate area that has been detected as the target area. The second similarity degree of the point on the pixel position; according to the first similarity degree, or the first similarity degree and the second similarity degree, it is judged whether the current area is the target area to realize the detection of the target area. Wherein, the explanations of terms that are the same as or corresponding to the above embodiments are not repeated here. On the basis of the foregoing embodiment, as shown in FIG. 12 , the embodiment of the present disclosure may include the following steps:

S701 , cluster each grayscale pixel point according to the pixel grayscale of each grayscale pixel point in the medical image, and use the region where each grayscale pixel point belonging to the same category is located as a candidate region.

S702 , a gradient image corresponding to the medical image is generated according to the grayscale of each pixel, and a gradient edge in the gradient image is determined according to the pixel gradient of each gradient pixel point in the gradient image.

S703, for the current area in each candidate area, obtain the first similarity on the pixel position of each current pixel point in the current boundary of the current area and each edge pixel point in the gradient edge, and each current pixel point and the detected target The second similarity at the pixel position of each target pixel in the target boundary of the candidate region of the region.

Among them, the current area is the candidate area to be detected at the current moment, the current boundary is the area boundary of the current area, the current pixel point is the pixel point on the current boundary, and the first similarity can indicate each current pixel point and each edge The similarity of the pixel points in the pixel position, which can be determined by the ratio between the first number of each current pixel point and the second number of the current pixel point that can be the same or similar to an edge pixel point in the pixel position Sure. Correspondingly, since each time the current area is selected from each candidate area, it can be detected whether the current area is the target area, which means that after the current area is updated, there may be candidate areas that have been detected as the target area. This candidate region is the previous current region. The target boundary is the area boundary of the candidate area that has been detected as the target area, and the target pixel point is the pixel point on the target boundary. Similarly, the second similarity can indicate the pixel position of each target pixel point and each edge pixel point. similarity.

S704, according to the first degree of similarity, or the first degree of similarity and the second degree of similarity, determine whether the current area is a target area to detect the target area.

Wherein, if the first similarity is high, that is, the coupling degree between each current pixel and each edge pixel is high, the current area can be determined as the target area. Exemplarily, as shown in FIG. 13a, if the first similarity The degree is determined by the number ratio between the first number of each current pixel point and the second number of current pixel points that can be the same or similar to an edge pixel in pixel position, then the first similarity may be 50% ; Of course, if the first similarity is not very high, it cannot be directly considered that the current area is not the target area at this time, because part of the current boundary of the current area may coincide with the gradient edge, and another part coincides with the target boundary. The area can also be used as the target area. Exemplarily, as shown in FIG. 13b, if the second similarity degree is expressed in the same manner as the first similarity degree, the first similarity degree and the second similarity degree are both 25%. That is, whether the current area is the target area can be determined according to the first similarity, or the first similarity and the second similarity. It should be noted that, each current region can be detected by using the above steps, thereby realizing the effect of detecting the target region from each candidate region.

In practical applications, optionally, on the basis of the above-mentioned embodiment, the detection process of the target area may be to determine the area detection order of each candidate area according to the area attribute of the target area to be detected, and select the area detection order from each candidate area according to the area detection order. The current area is filtered out of the area. As an example, the target area is determined to be a high-attenuation area according to the area attribute as an example. The candidate area corresponding to the high-attenuation area is usually a low-gray candidate area. The detection of the region starts; according to the coupling degree between the region boundary and the gradient edge of the current region, it is judged whether the current region is used as the target region, and the detection of the target region is realized according to the judgment result; the next region located in the current region in the region detection sequence is used as the current region. area, and repeat the step of judging whether to use the current area as the target area according to the coupling degree between the area boundary of the current area and the gradient edge, until the detected candidate area and/or gradient edge meet the preset judgment end condition, the judgment The ending condition may include that the number of detected candidate regions is greater than a preset number threshold, gradient edges no longer exist, etc., which are not specifically limited here.

In the embodiment of the present disclosure, the first similarity of each current pixel point in the current boundary and each edge pixel point in the gradient edge in the pixel position, and each current pixel point and the target boundary of the candidate area that has been detected as the target area The second similarity of each target pixel in the pixel position determines whether the current area is the target area, thereby achieving the effect of accurate detection of the target area in different situations.

On this basis, in order to better understand the specific implementation process of the above-mentioned target area detection, an exemplary description is given below with reference to specific examples. Exemplarily, taking the target area as a high-attenuation area as an example, step by step traverse each category of the area grayscale from low to high in the LBG classification, take the currently undetected candidate area of the lowest category as the current area, and perform a calculation on the current area. The current boundary is compared with the gradient edge. If the number of current pixels in the current boundary that are the same or similar to the edge pixels on the gradient edge at the pixel position reaches a certain proportion of the total number of current pixels, then the current boundary is considered to be the current boundary. The area belongs to a part of the high attenuation area; if the above conditions are not met, it can be further judged whether a part of the current boundary is relatively close to the gradient edge, and the other part is the area boundary (ie the target boundary) that has been detected as a high attenuation area. If it is relatively close, the current area can also be judged as a high attenuation area. Further, set the edge pixels on the gradient edge that is close to the high attenuation area as 0, that is, set the edge pixels occupied by the target boundary as 0. Repeat the above steps until you reach the category where the gray level of the relatively large area in the LBG classification is located and/or the gradient edge no longer exists in the gradient image. The reason for this setting is that the former should theoretically have completed the goal at this time. The detection of the area, but it has not been completed in the actual application, it may be that there is an error in the area detection process, and the loss should be stopped in time; the latter means that all the high attenuation areas have been detected, and the detection can be stopped at this time.

Correspondingly, the detection process of the direct exposure area and the high attenuation area can be changed from low to high area detection order to high to low area detection order, changing the lowest category to the highest category, and traversing to the LBG classification. The category where the gray level of the relatively large area is located in the LBG classification is changed to the category where the gray level of the relatively small area in the LBG classification is traversed, and the rest of the steps are the same, and will not be repeated here.

It should be noted that, in practical applications, it is not necessary to pre-determine whether there is a high attenuation area or a direct exposure area in the X-ray image. It is only necessary to detect based on the detection steps related to the high attenuation area when the detection target is a high attenuation area. , and when the detection target is the direct exposure area, the detection is performed based on the detection steps related to the direct exposure area. That is, the above-mentioned embodiments can extract high attenuation regions, directly exposed regions, or both high attenuation regions and directly exposed regions from the X-ray image.

It should be understood that although the steps in the flowcharts of FIG. 2 to FIG. 13 are displayed in sequence according to the arrows, these steps are not necessarily executed in the sequence indicated 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 FIG. 2 to FIG. 13 may include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but may be executed at different times. The order of execution is also not necessarily sequential, but may be performed alternately or alternately with other steps or at least a portion of the steps or stages within the other steps.

In one embodiment, as shown in FIG. 14, a mammography image acquisition apparatus is provided, the apparatus comprising:

a type determination module 801, configured to perform a stress test on the breast part of the object to be detected, and determine the type of the breast part according to the stress test result;

The quantification parameter determination module 802 is used for acquiring the optical image of the breast part, and performing identification processing on the optical image to obtain the quantification parameter of the breast part;

The acquisition parameter determination module 803 is configured to determine the image acquisition parameter of the breast part according to the type of the breast part and the quantification parameter, and acquire the X-ray image of the breast part according to the image acquisition parameter.

In one embodiment, the device further includes:

The registration module 804 is configured to register the optical image with the X-ray image, and display the breast contour on the X-ray image or display mammography image information on the optical image according to the registration result.

In one embodiment, the above-mentioned type determination module 801 includes:

The variation relationship determining unit is used to perform a pressure test on the breast part by using the compression component, and obtain the variation relationship between the compression force generated during the pressure test and the compression thickness of the breast part;

The type determination unit is used for determining the type of the breast part according to the changing relationship between the compression force and the compression thickness.

Optionally, the above-mentioned changing relationship includes changing curvature; the above-mentioned type determining unit includes:

The comparison subunit is used to compare the changing curvature with the preset curvature threshold;

The determining subunit is used for determining that the type of the breast part is fat when the change curvature is greater than the curvature threshold value; when the changing curvature is not greater than the curvature threshold value, determining the type of the breast part is the dense type.

In one embodiment, the above-mentioned acquisition parameter determination module 803 includes:

a first radiation field determining unit, configured to determine the radiation field corresponding to the breast part according to the quantitative parameter;

The acquisition parameter determination unit is used for determining the image acquisition parameters of the breast part according to the type of the breast part and the radiation field corresponding to the breast part.

The identification unit is used for identifying and processing the optical image to obtain the breast contour; the breast contour includes the positions of a plurality of points;

The calculation unit is used to perform mathematical operation processing on the position of each point on the breast contour to obtain the volume of the breast part; or, perform mathematical operation processing on the position of each point on the breast contour to obtain the projected area of the breast part on the compression device.

In one embodiment, the image acquisition parameters include a first image acquisition parameter and a second image acquisition parameter; the acquisition parameter determination module 803 includes:

The second radiation field determining unit is configured to obtain the radiation field corresponding to the breast part according to the volume or projected area of the breast part;

a first acquisition parameter determination unit, configured to determine the radiation field corresponding to the breast part as the first image acquisition parameter;

The second acquisition parameter determination unit is configured to determine the second image acquisition parameter according to the type of the breast part.

In one of the embodiments, the above-mentioned second acquisition parameter determination unit includes:

The compression thickness obtaining subunit is used to obtain the actual compression thickness of the breast part under preset conditions;

The acquisition parameter determination subunit is used for determining the second image acquisition parameter according to the type of the breast part and the actual compression thickness.

Optionally, the above acquisition parameter determination subunit is further configured to obtain the second image acquisition parameter corresponding to the type of the breast part and the actual compression thickness in a preset mapping table according to the type of the breast part and the actual compression thickness; wherein, The mapping table includes the types and compression thicknesses of multiple groups of parts, and the second image acquisition parameters corresponding to the types and compression thicknesses of each group of parts; the second image acquisition parameters are used to characterize the power of the radiation source and the filter of the beam limiter. over the way.

In one of the embodiments, the above-mentioned identification unit is specifically configured to perform identification processing on the optical image to obtain the pixel value of each pixel in the optical image; and determine the breast contour through an image segmentation algorithm according to the pixel value of each pixel.

In one embodiment, the above-mentioned identification unit is specifically used to obtain a segmentation threshold; according to the segmentation threshold and the pixel value of each pixel point, the optical image is segmented into a breast area image and a background area image; the outline of the breast area image is used as the breast area image. contour.

In one embodiment, the above-mentioned registration module 804 is specifically configured to extract multiple first features in the optical image and multiple second features in the X-ray image through a feature extraction algorithm; according to the first features and the second features , to register the optical image with the X-ray image.

In one embodiment, the above-mentioned registration module 804 is specifically configured to perform feature matching between the first feature and the second feature through similarity measurement and cluster analysis; according to the coordinates of the first feature in the optical image, the second feature The coordinates of the feature in the X-ray image are obtained, and the coordinate mapping function is obtained; according to the coordinate mapping function, the optical image and the X-ray image are registered.

In one embodiment, the above-mentioned registration module 804 is specifically configured to obtain position information of the breast contour in the X-ray image according to the breast contour, the optical image and the registered X-ray image; and display the breast on the X-ray image. contour.

In one embodiment, the above-mentioned registration module 804 is specifically configured to determine the position information of the breast in the optical image according to the outline of the breast and the optical image; according to the position information of the breast in the optical image and the registered X-ray image , to obtain the position information of the breast in the X-ray image.

In one embodiment, the device further includes:

The candidate region determination module is used to cluster each grayscale pixel point according to the pixel grayscale of each grayscale pixel point in the X-ray image, and use the region where each grayscale pixel point belonging to the same category is located as a candidate region;

The gradient edge determination module is used to generate a gradient image corresponding to the X-ray image according to the gray level of each pixel, and determine the gradient edge in the gradient image according to the pixel gradient of each gradient pixel point in the gradient image;

The target area detection module is used to detect the target area from each candidate area according to the coupling degree between the area boundary and the gradient edge of each candidate area.

In one of the embodiments, the above-mentioned candidate region determination module includes:

The gray-scale category split point determination unit is used to sort the pixel gray levels of each gray-scale pixel point in the X-ray image, and determine the gray-scale category split point in the gray-scale sorting result based on the preset number of categories;

The gray-scale pixel point clustering unit is used to use the gray-scale category split point as the initial cluster center, and cluster each gray-scale pixel point based on the cluster center and the pixel gray level of each gray-scale pixel point.

In one embodiment, the gray-scale pixel point clustering unit is specifically configured to, for each gray-scale pixel point, determine the pixel gray-scale of the gray-scale pixel point and the gray-scale distance between each cluster center, and calculate the gray-scale pixel point Pixels are clustered into the category of the cluster center corresponding to the minimum grayscale distance; grayscale distortion is determined according to the minimum grayscale distance corresponding to each grayscale pixel point, and whether the grayscale distortion meets the preset value is determined. Clustering end condition; if not, then for each category, re-determine the cluster center of the category according to the pixel grayscale of each grayscale pixel point belonging to the category after clustering; repeat the execution to determine the pixel of the grayscale pixel point The steps of the grayscale and the grayscale distance between each cluster center, until the grayscale distortion satisfies the clustering end condition, the clustering ends.

In one embodiment, as shown in FIG. 15, there is provided a mammography image display device, the device comprising:

an acquisition module 805, configured to acquire an optical image of the breast and an X-ray image of the breast part of interest; wherein, the optical image is a visible light image;

a contour calculation module 806, configured to determine the breast contour according to the optical image;

a registration module 807, configured to register the optical image with the X-ray image;

The position calculation module 808 is used for displaying the outline of the breast on the X-ray image, or displaying the X-ray information of the part of interest of the breast on the optical image.

Specifically, the contour calculation module identifies the breast contour in the optical image, and then the registration module registers the optical image and the X-ray image, and the position calculation module can obtain the position information of the breast contour in the X-ray image, so as to The breast contour is displayed in the X-ray image, and the relative position of the X-ray image in the whole breast is marked with respect to the doctor's experience.

The method in this embodiment solves the problem that the X-ray imaging device can only obtain a partial image of the breast, so that the doctor can only judge the relative position of the lesion and the entire breast according to experience, the efficiency is low and the error is large, which improves the doctor's ability to diagnose Diagnostic efficiency in the process and accuracy of location annotation.

In one embodiment, as shown in FIG. 16, an area detection device is provided, and the device includes:

The candidate area determination module 809 is used to cluster each grayscale pixel point according to the pixel grayscale of each grayscale pixel point in the medical image, and take the area where each grayscale pixel point belonging to the same category is located as a candidate area;

The gradient edge determination module 810 is used to generate a gradient image corresponding to the medical image according to the gray level of each pixel, and determine the gradient edge in the gradient image according to the pixel gradient of each gradient pixel point in the gradient image;

The target area detection module 811 is configured to detect the target area from each candidate area according to the coupling degree between the area boundary and the gradient edge of each candidate area.

In one of the embodiments, the candidate region determination module includes:

The gray-scale category split point determination unit is used to sort the pixel gray levels of each gray-scale pixel point in the medical image, and determine the gray-scale category split point in the gray-scale sorting result based on the preset number of categories;

On this basis, the optional gray-scale pixel clustering unit can be used for:

For each grayscale pixel point, determine the pixel grayscale of the grayscale pixel point and the grayscale distance between each cluster center, and cluster the grayscale pixel points to the cluster center corresponding to the smallest grayscale distance. in the category;

Determine the grayscale distortion according to the minimum grayscale distance corresponding to each grayscale pixel, and judge whether the grayscale distortion satisfies the preset clustering end condition;

If not, for each category, re-determine the cluster center of the category according to the pixel grayscale of each grayscale pixel point belonging to the category after clustering;

The steps of determining the pixel grayscale of the grayscale pixel point and the grayscale distance between each cluster center are repeated until the grayscale distortion satisfies the clustering end condition, and the clustering ends.

In one embodiment, the above-mentioned gradient edge determination module includes:

The grayscale category area screening unit is used to screen out the grayscale category area corresponding to the regional attribute of the target area to be detected from each candidate area according to the regional grayscale of each candidate area;

The first gradient edge determination unit is configured to determine the gradient edge in the gradient image according to the pixel gradient of the gradient pixel points in the gradient image corresponding to the pixel points of each region in the gray-scale category area.

The binary image generation unit is used to sort each gradient pixel point based on the pixel gradient of each gradient pixel point in the gradient image, and filter each gradient pixel point according to the sorting position of each gradient pixel point in the sorting result , generate a binary image corresponding to the gradient image;

The second gradient edge determination unit is configured to use the binary edge formed by each binary pixel in the binary image as the gradient edge of the gradient image.

In one embodiment, for the current region in each candidate region, the above-mentioned target region detection module includes:

The similarity obtaining unit is used to obtain the first similarity at the pixel position of each current pixel in the current boundary of the current area and each edge pixel in the gradient edge, as well as each current pixel and the candidate area that has been detected as the target area The second similarity of each target pixel at the pixel position in the target boundary of ;

The first target area detection unit is configured to determine whether the current area is the target area according to the first similarity, or the first similarity and the second similarity, so as to detect the target area.

In one embodiment, the above-mentioned target area detection module includes:

The current area screening unit is used to determine the area detection order of each candidate area according to the area attribute of the target area, and screen out the current area from each candidate area according to the area detection order;

The second target area detection unit is used to judge whether the current area is used as the target area according to the coupling degree between the area boundary of the current area and the gradient edge, and to detect the target area according to the judgment result;

The iterative execution unit is used for taking the next area located in the current area in the area detection sequence as the current area, and repeating the steps of judging whether to use the current area as the target area according to the coupling degree between the area boundary and the gradient edge of the current area, Until the detected candidate region and/or the gradient edge satisfies the preset judgment end condition.

For the specific limitation of the foregoing apparatus, reference may be made to the foregoing limitation of the foregoing method, which will not be repeated here. Each module in the above 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, and the computer device may be a terminal, and its internal structure diagram may be as shown in FIG. 17 . 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 equipment is used for wired or wireless communication with external terminals, and the wireless communication can be realized by WIFI, operator network, NFC (Near Field Communication) or other technologies. The computer program, when executed by a processor, implements a mammography image display 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. 17 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. Include more or fewer components than shown in the figures, or combine certain components, or have a different arrangement of components.

In one embodiment, a computer device is provided, including a memory and a processor, where a computer program is stored in the memory, and the processor implements the steps in the above method when executing the computer program.

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, implements the steps in the above method.

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 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. Non-volatile memory may include read-only memory (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 (SRAM) or 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 examples 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 pointed out 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

A method for acquiring a mammogram, characterized in that the method comprises:

Perform a stress test on the breast part of the object to be tested, and determine the type of the breast part according to the stress test result;

Obtaining an optical image of the breast part, and performing identification processing on the optical image to obtain a quantitative parameter of the breast part;

Image acquisition parameters of the breast part are determined according to the type of the breast part and quantification parameters, and an X-ray image of the breast part is acquired according to the image acquisition parameters.
The method according to claim 1, wherein the method further comprises: registering the optical image with the X-ray image, and displaying a breast contour or a breast contour on the X-ray image according to the registration result. Mammographic image information is displayed on the optical image.
The method according to claim 1, wherein, performing a stress test on the breast part of the object to be tested, and determining the type of the breast part according to the stress test result, comprising:

Use the compression component to perform a pressure test on the breast part, and obtain the changing relationship between the compression force generated during the pressure test and the compression thickness of the breast part;

The type of the breast part is determined according to the changing relationship between the compression force and the compression thickness.
The method according to any one of claims 1-3, wherein the determining the image acquisition parameter of the breast part according to the type of the breast part and the quantification parameter comprises:

determining the radiation field corresponding to the breast part according to the quantitative parameter;

Image acquisition parameters of the breast part are determined according to the type of the breast part and the radiation field corresponding to the breast part.
The method according to any one of claims 1-3, wherein the identifying and processing the optical image to obtain the quantitative parameter of the breast part comprises:

Performing identification processing on the optical image to obtain the breast contour; the breast contour includes the positions of a plurality of points;

Perform mathematical operation processing on the position of each point on the breast contour to obtain the volume of the breast part; or,

Perform mathematical operation processing on the positions of each point on the breast contour to obtain the projected area of the breast part on the compression device.
The method according to claim 5, wherein the image acquisition parameters include a first image acquisition parameter and a second image acquisition parameter; the determination of the breast part according to the type and quantification parameters of the breast part Image acquisition parameters, including:

According to the volume or projected area of the breast part, the radiation field corresponding to the breast part is obtained;

determining the radiation field corresponding to the breast part as the first image acquisition parameter;

The second image acquisition parameters are determined based on the type of the breast part.
The method according to claim 6, wherein the determining the second image acquisition parameter according to the type of the breast part comprises:

obtaining the actual compression thickness of the breast part under preset conditions;

The second image acquisition parameter is determined according to the type of the breast part and the actual compression thickness.
The method according to claim 5, wherein the identifying and processing the optical image to obtain the breast contour comprises:

performing identification processing on the optical image to obtain the pixel value of each pixel in the optical image;

According to the pixel value of each pixel point, the breast contour is determined through an image segmentation algorithm.
The method according to claim 8, wherein, determining the breast contour by an image segmentation algorithm according to the pixel value of each pixel point, comprising:

Get segmentation threshold;

According to the segmentation threshold and the pixel value of each pixel point, the optical image is segmented into a breast area image and a background area image;

The outline of the breast region image is used as the breast outline.
The method of claim 2, wherein the registering the optical image with the X-ray image comprises:

Extracting a plurality of first features in the optical image and a plurality of second features in the X-ray image by a feature extraction algorithm;

The optical image is registered with the X-ray image based on the first feature and the second feature.
The method according to claim 10, wherein the registering the optical image and the X-ray image according to the first feature and the second feature comprises:

Perform feature matching on the first feature and the second feature through similarity measurement and cluster analysis;

obtaining a coordinate mapping function according to the coordinates of the first feature in the optical image and the coordinates of the second feature in the X-ray image;

The optical image is registered with the X-ray image according to the coordinate mapping function.
The method according to claim 1, wherein after acquiring the X-ray image of the breast part according to the image acquisition parameter, the method further comprises:

Clustering each of the gray-scale pixel points according to the pixel gray level of each gray-scale pixel point in the X-ray image, and using the region where each of the gray-scale pixel points belonging to the same category is located as a candidate region;

A gradient image corresponding to the X-ray image is generated according to each pixel grayscale, and a gradient edge in the gradient image is determined according to the pixel gradient of each gradient pixel point in the gradient image;

According to the coupling degree between the area boundary of each candidate area and the gradient edge, a target area is detected from each of the candidate areas; the target area includes a high attenuation area and/or a direct exposure area.
A mammogram display method, characterized in that the method comprises:

Obtain optical images of the breast as well as X-ray images of the breast area of interest;

determining a breast contour from the optical image;

registering the optical image with the X-ray image;

The breast contour is displayed on the X-ray image, or the X-ray information of the breast region of interest is displayed on the optical image.
The method of claim 13, wherein the determining a breast contour according to the optical image comprises:

obtaining the pixel value of each pixel point according to the optical image;

According to the pixel value of each pixel point, the breast contour is determined through an image segmentation algorithm.
The method according to claim 14, wherein the determining the breast contour through an image segmentation algorithm according to the pixel value of each pixel point comprises:

Get segmentation threshold;

According to the segmentation threshold and the pixel value of each pixel point, the optical image is segmented into a breast area image and a background area image;

The contour of the breast region image is taken as the breast contour.
The method of claim 13, wherein the registering the optical image with the X-ray image comprises:

Extracting a plurality of first features in the optical image and a plurality of second features in the X-ray image by a feature extraction algorithm;

The optical image is registered with the X-ray image based on the first feature and the second feature.
The method according to claim 16, wherein the registering the optical image and the X-ray image according to the first feature and the second feature comprises:

Perform feature matching on the first feature and the second feature through similarity measurement and cluster analysis;

obtaining a coordinate mapping function according to the coordinates of the first feature in the optical image and the coordinates of the second feature in the X-ray image;

The optical image is registered with the X-ray image according to the coordinate mapping function.
A breast X-ray image acquisition method, characterized in that the method comprises:

Control the compression plate to compress the breast;

Controlling the optical image acquisition unit to acquire an optical image of the breast in the compressed state, and determining the outline of the breast in the optical image;

Controlling the X-ray image acquisition unit to acquire the X-ray image of the breast under compression;

The optical image is registered with the X-ray image, and the breast contour is displayed on the X-ray image or the X-ray information of the breast part of interest is displayed on the optical image.
A method for region detection, comprising:

According to the pixel gray level of each gray pixel point in the medical image, each gray pixel point is clustered, and the region where each gray pixel point belonging to the same category is located is used as a candidate region;

A gradient image corresponding to the medical image is generated according to each pixel grayscale, and a gradient edge in the gradient image is determined according to the pixel gradient of each gradient pixel point in the gradient image;

A target region is detected from each of the candidate regions according to the degree of coupling between the region boundary of each of the candidate regions and the gradient edge.
The method according to claim 19, wherein the clustering each of the grayscale pixels according to the pixel grayscale of each grayscale pixel in the medical image comprises:

Sort the pixel gray level of each gray level pixel point in the medical image, and determine the gray level category split point in the gray level sorting result based on the preset number of categories;

The grayscale category split point is used as an initial cluster center, and each grayscale pixel point is clustered based on the cluster center and the pixel grayscale of each grayscale pixel point.
The method according to claim 20, wherein the clustering of each of the gray-scale pixel points based on the cluster center and the pixel gray level of each of the gray-scale pixel points comprises:

For each grayscale pixel point, determine the pixel grayscale of the grayscale pixel point and the grayscale distance between each of the cluster centers, and cluster the grayscale pixel points to the smallest grayscale pixel. In the category where the cluster center corresponding to the grayscale distance is located;

Determine grayscale distortion according to the minimum grayscale distance corresponding to each of the grayscale pixels, and determine whether the grayscale distortion satisfies a preset clustering end condition;

If not, for each of the categories, re-determine the cluster center of the category according to the pixel grayscale of each grayscale pixel point belonging to the category after clustering;

The step of determining the pixel grayscale of the grayscale pixel point and the grayscale distance between each of the cluster centers is repeated until the grayscale distortion satisfies the clustering termination condition, and the clustering ends.
The method according to claim 19, wherein the determining the gradient edge in the gradient image according to the pixel gradient of each gradient pixel in the gradient image comprises:

According to the regional grayscale of each of the candidate regions, screen out a grayscale category region corresponding to the regional attribute of the target region to be detected from each of the candidate regions;

The gradient edge in the gradient image is determined according to the pixel gradients of the gradient pixel points in the gradient image corresponding to the pixel points of each region in the gray-scale category area.
The method according to claim 19, wherein the determining the gradient edge in the gradient image according to the pixel gradient of each gradient pixel in the gradient image comprises:

Sorting the gradient pixel points based on the pixel gradient of each gradient pixel point in the gradient image, and screening each gradient pixel point according to the sorting position of each gradient pixel point in the sorting result , generating a binary image corresponding to the gradient image;

The binary edge formed by each binary pixel in the binary image is used as the gradient edge of the gradient image.
The method according to claim 19, wherein, for the current region in each of the candidate regions, according to the degree of coupling between the region boundary of each of the candidate regions and the gradient edge, from each of the candidate regions The target area is detected in the candidate area, including:

Obtain the first similarity in pixel positions of each current pixel in the current boundary of the current area and each edge pixel in the gradient edge, and each of the current pixels and the candidate detected as the target area the second similarity of each target pixel at the pixel position in the target boundary of the area;

According to the first similarity, or the first similarity and the second similarity, it is determined whether the current area is the target area to detect the target area.
The method according to claim 19, wherein the detecting a target region from each candidate region according to the coupling degree between the region boundary of each candidate region and the gradient edge comprises:

Determine the region detection order of each of the candidate regions according to the region attribute of the target region to be detected, and filter out the current region from each of the candidate regions according to the region detection order;

Judging whether to use the current region as the target region according to the coupling degree between the region boundary of the current region and the gradient edge, and detecting the target region according to the judgment result;

Taking the next region located in the current region in the region detection sequence as the current region, and repeating the process of determining whether to use the The step of using the current area as the target area until the detected candidate area and/or the gradient edge satisfies a preset judgment end condition.