Body cavity bleeding detection method and device

Abstract

The invention discloses a body cavity bleeding detection method and device, and the method comprises the steps: obtaining an ultrasonic effective region of an ultrasonic image, and obtaining the physical length and pixel length of a center line of the ultrasonic effective region; acquiring a bleeding area of the ultrasonic image; acquiring an overlapping area of the ultrasonic effective area and the bleeding area, and acquiring a minimum inscribed rectangular frame of the overlapping area; the length of the upper boundary of the minimum inscribed rectangular frame serves as the pixel value of the bleeding depth, and the physical value of the bleeding depth is calculated based on the ratio of the physical length of the center line of the ultrasonic effective area to the pixel length; calculating a weighted average value of the physical values of the bleeding depths based on the physical values of the bleeding depths of the ultrasonic images under different breathing phases and the weights corresponding to the breathing phases; the amount of bleeding is calculated based on a weighted average of the bleeding depth physical values. According to the body cavity bleeding detection method and device, the bleeding amount is accurately calculated, and the technical problem that in the prior art, the bleeding amount detection accuracy is low is solved.

Description

Technical Field

[0001] This invention belongs to the field of medical image processing technology, specifically, it relates to a method and device for detecting body cavity bleeding. Background Technology

[0002] In the field of trauma emergency care, timeliness is a core factor determining treatment effectiveness. For life-threatening injuries such as intrathoracic hemorrhage, Extended Forecasting with Ultrasound-Assisted Assessment (eFAST) technology enables rapid imaging diagnosis at the trauma site using portable ultrasound equipment, significantly shortening the response time from injury assessment to clinical intervention. This technology overcomes the dependence of traditional imaging examinations on equipment conditions and operating environments, providing real-time dynamic physiological parameter monitoring support for emergency teams in complex scenarios such as shipboard swaying, outdoor turbulence, or emergency room overload. Based on standardized operating procedures, medical teams can complete intrathoracic hemorrhage screening within 30 seconds, effectively ruling out life-threatening injuries such as tension pneumothorax and cardiac tamponade, and guiding treatment priority decisions through quantitative assessment of bleeding rate, achieving precise allocation of medical resources to critically ill patients.

[0003] However, in the above scenarios, the personnel involved in the rescue may not have professional ultrasound knowledge. Therefore, it is necessary to equip the examiners with a simple and effective auxiliary measurement method to help them quickly measure the pleural hemorrhage of the examinee and ensure that the measurement results are true and reliable.

[0004] In the existing technology, there are mainly the following two auxiliary measurement methods: (1) Manual measurement using manual measuring tools. This method provides medical staff with real-time measuring tools such as length, area, and volume. After discovering a suspicious bleeding area, the medical staff can measure it and finally output an examination report. This method can provide professional examination for injured persons in real time, but it is affected by the skill level of the medical staff and cannot be used in the absence of a professional ultrasound doctor.

[0005] (2) Automated detection based on deep learning. With the application of deep learning technology, target detection algorithms have been used in many medical imaging applications, providing assistance to medical staff. Some ultrasound equipment companies have launched products for detecting pleural hemorrhage, which can detect pleural hemorrhage in real time, providing assistance to medical staff. However, this method only considers the detection rate, and calculates the maximum diameter of the detection frame in the horizontal or vertical direction when calculating the amount of bleeding. It does not design the method from the perspective of actual needs and clinical practice, so the accuracy of the measurement is not high. Summary of the Invention

[0006] This invention provides a method for detecting intracavitary bleeding, which solves the technical problem of low accuracy in bleeding volume detection in the prior art.

[0007] To solve the above technical problems, the application adopts the following technical solutions to achieve the purpose: The body cavity bleeding detection method comprises the following steps: Step S1: the ultrasound images in different respiratory phases in a respiratory cycle are respectively subjected to the following steps S1-1 to S1-4 to obtain the physical value of the bleeding depth of the ultrasound images in different respiratory phases: Step S1-1, obtaining the ultrasound effective region of the ultrasound image, and obtaining the physical length and pixel length of the center line of the ultrasound effective region; Step S1-2, obtaining the bleeding region of the ultrasound image; Step S1-3, obtaining the overlapping region of the ultrasound effective region and the bleeding region, and obtaining the minimum inscribed rectangular frame of the overlapping region; Step S1-4, taking the length of the upper boundary of the minimum inscribed rectangular frame as the pixel value of the bleeding depth, and calculating the physical value of the bleeding depth based on the ratio of the physical length and the pixel length of the center line of the ultrasound effective region; Step S2, calculating the weighted average value of the physical value of the bleeding depth based on the physical value of the bleeding depth of the ultrasound images in different respiratory phases and the weight corresponding to the respiratory phase; Step S3, calculating the bleeding volume based on the weighted average value of the physical value of the bleeding depth.

[0008] In some embodiments of the application, the step S1-2 specifically comprises: inputting the ultrasound image into a pre-trained deep learning model, and the deep learning model outputs the bleeding region of the ultrasound image.

[0009] In some embodiments of the application, the respiratory cycle comprises three respiratory phases: end-expiratory phase, end-inspiratory phase and intermediate phase; the corresponding weights are different for different respiratory phases.

[0010] In some embodiments of the application, the step S3 specifically comprises: taking the product of the weighted average value of the physical value of the bleeding depth and a preset coefficient as the bleeding volume.

[0011] In some embodiments of the application, after step S3, the method further comprises: Step S4, judging the bleeding grade based on the calculated bleeding volume and outputting.

[0012] In some embodiments of the application, the step S4 specifically comprises: If the calculated bleeding volume is within a first set range, the bleeding grade is determined to be grade one bleeding; If the calculated bleeding volume is within a second set range, the bleeding grade is determined to be grade two bleeding; Any value in the first set range is less than any value in the second set range.

[0013] The body cavity bleeding detection device comprises: The bleeding depth physical value acquisition module is configured to: acquire an ultrasound effective region of the ultrasound image at different breathing phases in a breathing cycle, respectively, and obtain a physical length and a pixel length of a center line of the ultrasound effective region; acquire a bleeding region of the ultrasound image; acquire an overlapping region of the ultrasound effective region and the bleeding region, and acquire a minimum inscribed rectangular frame of the overlapping region; take a length of an upper boundary of the minimum inscribed rectangular frame as a pixel value of the bleeding depth; and calculate a physical value of the bleeding depth based on a ratio of the physical length to the pixel length of the center line of the ultrasound effective region. The weighted calculation module is configured to: calculate a weighted average value of the physical value of the bleeding depth based on the physical value of the bleeding depth of the ultrasound image at different breathing phases and a weight corresponding to the breathing phase. The bleeding amount calculation module is configured to: calculate the bleeding amount based on the weighted average value of the physical value of the bleeding depth.

[0014] In some embodiments of the present application, the bleeding region of the ultrasound image is specifically obtained by: inputting the ultrasound image into a pre-trained deep learning model, and outputting the bleeding region of the ultrasound image by the deep learning model.

[0015] In some embodiments of the present application, the body cavity bleeding monitoring device further comprises: The bleeding grade judgment module is configured to: judge the bleeding grade based on the calculated bleeding amount, and output the bleeding grade.

[0016] In some embodiments of the present application, the bleeding grade judgment module is specifically configured to: If the calculated bleeding amount is within a first set range, the bleeding grade is determined to be a first-grade bleeding. If the calculated bleeding amount is within a second set range, the bleeding grade is determined to be a second-grade bleeding. Any value in the first set range is less than any value in the second set range.

[0017] Compared with the prior art, the advantages and positive effects of the body cavity bleeding detection method and device are that: the body cavity bleeding detection method and device acquire the ultrasound effective area of the ultrasound image, and obtain the physical length and the pixel length of the center line of the ultrasound effective area; acquire the bleeding area of the ultrasound image; acquire the overlapping area of the ultrasound effective area and the bleeding area, and acquire the minimum inscribed rectangular frame of the overlapping area; take the length of the upper boundary of the minimum inscribed rectangular frame as the pixel value of the bleeding depth, calculate the physical value of the bleeding depth based on the ratio of the physical length and the pixel length of the center line of the ultrasound effective area; perform the above steps on the ultrasound image in different breathing phases in a breathing cycle respectively to obtain the physical value of the bleeding depth of the ultrasound image in different breathing phases; then calculate the weighted average value of the bleeding depth physical value based on the physical value of the bleeding depth of the ultrasound image in different breathing phases and the weight corresponding to the breathing phase; calculate the bleeding amount based on the weighted average value of the bleeding depth physical value; therefore, the body cavity bleeding detection method and device accurately calculate the bleeding amount based on the weighted average value of the bleeding depth physical value, and solve the technical problem of low bleeding amount detection accuracy in the prior art.

[0018] Other features and advantages of the present application will become more apparent from the following detailed description when read in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 is a flowchart of one embodiment of the body cavity bleeding detection method proposed by the present application; Figure 2 is a flowchart of one embodiment of step S1; Figure 3 is the ultrasound effective area of the acquired ultrasound image; Figure 4 is the bleeding area of the acquired ultrasound image; Figure 5 is the overlapping area of the ultrasound effective area and the bleeding area; Figure 6 is the minimum inscribed rectangular frame of the overlapping area; Figure 7 is a flowchart of another embodiment of the body cavity bleeding detection method proposed by the present application; Figure 8 is a flowchart of another embodiment of the body cavity bleeding detection method proposed by the present application; Figure 9 is a structural block diagram of one embodiment of the body cavity bleeding detection device proposed by the present application. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application will be further described in detail below in combination with the drawings and embodiments. Embodiment one,

[0021] The body cavity bleeding detection method of the embodiment mainly includes the following steps, see Figure 1 .

[0022] First, the acquired ultrasound image needs to be preprocessed. The ultrasound image is standardized (such as gray scale normalization, denoising, etc.), and the contrast of the bleeding area is enhanced.

[0023] Step S1: For the ultrasound image at different breathing phases in the breathing cycle, the following steps S1-1 ~ S1-4 are performed respectively to obtain the physical value of the bleeding depth of the ultrasound image at different breathing phases, see Figure 2 .

[0024] Step S1-1, obtain the ultrasound effective area of the ultrasound image, and obtain the physical length and pixel length of the center line of the ultrasound effective area.

[0025] In this step, the ultrasound effective area of the ultrasound image is extracted, and the key scale standard, that is, the pixel length of the center line of the effective area, is recorded as L i , the physical length of L i is a fixed value, recorded as L_r i , which is used for conversion of the physical value of the bleeding amount.

[0026] In this embodiment, the threshold method is used to obtain the ultrasound effective area of the ultrasound image. The pixel value in the ultrasound effective area is greater than 0, and the pixel value outside the ultrasound effective area is equal to 0, so the threshold method can be used to extract the ultrasound effective area and exclude the interference outside the area.

[0027] Figure 3 The sector area in

[0028] Step S1-2, obtain the bleeding area of the ultrasound image.

[0029] In some embodiments of the present application, obtaining the bleeding area of the ultrasound image specifically includes: Input the ultrasound image into a pre-trained deep learning model, and the deep learning model outputs the bleeding area of the ultrasound image.

[0030] For real-time ultrasound images, a pre-trained deep learning model (such as Yolo v11 model) is used to monitor the bleeding area of the image, and the boundary box of the bleeding area is marked.

[0031] Figure 4 The area in the rectangular box shown in

[0032] The pre-trained deep learning model is used to obtain the bleeding area of the ultrasound image, which is simple, convenient, fast and accurate.

[0033] In step S1-3, an overlapping area of the ultrasound effective area and the bleeding area is obtained, and a minimum inscribed rectangle frame of the overlapping area is obtained.

[0034] The overlapping area of the ultrasound effective area and the bleeding area is the intersection of the ultrasound effective area and the bleeding area, as shown in the following figure. Figure 5 Figure 6 The rectangular frame shown in the figure is the minimum inscribed rectangle frame of the overlapping area.

[0035] In step S1-4, the length of the upper boundary of the minimum inscribed rectangle frame is taken as the pixel value of the bleeding depth, and the physical value of the bleeding depth is calculated based on the ratio of the physical length to the pixel length of the center line of the ultrasound effective area.

[0036] The ratio of the physical length to the pixel length of the center line of the ultrasound effective area is equal to the ratio of the physical value of the bleeding depth to the pixel value of the bleeding depth. Therefore, according to the ratio of the physical length to the pixel length of the center line of the ultrasound effective area and the pixel value of the bleeding depth, the physical value of the bleeding depth can be calculated to be applicable to various models.

[0037] For example, d_r i = d i * L_r i / L i ; Wherein, d_r i , d i are the physical value and pixel value of the bleeding depth of the ultrasound image in the i th respiratory phase in the respiratory cycle, respectively. L_r i , L i are the physical length and pixel length of the center line of the ultrasound effective area of the ultrasound image in the i th respiratory phase in the respiratory cycle, respectively.

[0038] For each respiratory phase in the respiratory cycle, steps S1-1 to S1-4 are performed respectively to obtain the physical value of the bleeding depth of the ultrasound image in each respiratory phase in the respiratory cycle.

[0039] In step S2, the weighted average value of the physical value of the bleeding depth is calculated based on the physical value of the bleeding depth of the ultrasound image in different respiratory phases and the weight corresponding to the respiratory phase.

[0040] The weighted average value of the physical value of the bleeding depth is calculated by the following formula.

[0041]

[0042] ​wherein d com is a weighted average of the physical values of the bleeding depths, d_r i is a physical value of the bleeding depth of the ultrasound image in the i th respiratory phase in a respiratory cycle, w i is a weight corresponding to the i th respiratory phase in a respiratory cycle, and n is the number of respiratory phases in a respiratory cycle.

[0043] In some embodiments of the present application, a respiratory cycle includes three respiratory phases: an end-expiratory phase, an end-inspiratory phase, and an intermediate phase. Different respiratory phases correspond to different weights. In a respiratory cycle, the sum of the weights of all respiratory phases is 1.

[0044] That is, n = 3, d_r1 is a physical value of the bleeding depth of the ultrasound image in the first respiratory phase (such as the end-expiratory phase), and w1 is a weight corresponding to the first respiratory phase (such as the end-expiratory phase); d_r2 is a physical value of the bleeding depth of the ultrasound image in the second respiratory phase (such as the end-inspiratory phase), and w2 is a weight corresponding to the second respiratory phase (such as the end-inspiratory phase); d_r3 is a physical value of the bleeding depth of the ultrasound image in the third respiratory phase (such as the intermediate phase), and w3 is a weight corresponding to the third respiratory phase (such as the intermediate phase).

[0045] For example, w1 = 0.6, w2 = 0.3, and w3 = 0.1.

[0046] The respiratory parameters of the patient are obtained in real time by using a respiratory monitoring device. According to the real-time respiratory monitoring result, the respiratory phases are obtained. The bleeding depth is compensated according to the respiratory phases, and a compensation value, that is, a weighted average of the physical values of the bleeding depths, is obtained.

[0047]

[0048] By designing a respiratory cycle to include three different respiratory phases, the weighted average of the physical values of the bleeding depths can be accurately calculated according to the physical values of the bleeding depths in the three different respiratory phases and the corresponding weights, and the calculation is not complicated due to too many respiratory phases.

[0049] In step S3, the amount of bleeding is calculated based on the weighted average of the physical values of the bleeding depths.

[0050] In some embodiments of the present application, the weighted average d com of the physical values of the bleeding depths is multiplied by a preset coefficient a, and the product is taken as the amount of bleeding. That is, the amount of bleeding = d com × a.

[0051] Wherein, a is a preset clinical calibration coefficient. The unit of the amount of bleeding is ml, and the unit of the physical value of the bleeding depth is mm.

[0052] The weighted average value d of the physical value of the bleeding depth is obtained by com The product of the preset coefficient a is multiplied by the weighted average value d of the physical value of the bleeding depth as the amount of bleeding, so that the amount of bleeding can be simply, conveniently and accurately calculated.

[0053] The body cavity bleeding detection method of the embodiment obtains the ultrasound effective region of the ultrasound image, and obtains the physical length and the pixel length of the center line of the ultrasound effective region; obtains the bleeding region of the ultrasound image; obtains the overlapping region of the ultrasound effective region and the bleeding region, and obtains the minimum inscribed rectangular frame of the overlapping region; takes the length of the upper boundary of the minimum inscribed rectangular frame as the pixel value of the bleeding depth, calculates the physical value of the bleeding depth based on the ratio of the physical length to the pixel length of the center line of the ultrasound effective region; for the ultrasound images at different breathing phases in a breathing cycle, the above steps are performed respectively to obtain the physical value of the bleeding depth of the ultrasound images at different breathing phases; then, based on the physical value of the bleeding depth of the ultrasound images at different breathing phases and the weight corresponding to the breathing phase, the weighted average value of the physical value of the bleeding depth is calculated; the amount of bleeding is calculated based on the weighted average value of the physical value of the bleeding depth; therefore, the body cavity bleeding detection method of the embodiment calculates the amount of bleeding according to the weighted average value of the physical value of the bleeding depth, realizes accurate calculation of the amount of bleeding, and solves the technical problem of low accuracy of bleeding amount detection in the prior art.

[0054] In each breathing cycle, the ultrasound images at different breathing phases are obtained, steps S1-1 to S1-4 are performed respectively to obtain the physical value of the bleeding depth of the ultrasound images at different breathing phases, and then steps S2 to S3 are performed to obtain the amount of bleeding in the breathing cycle.

[0055] In a plurality of consecutive breathing cycles, steps S1 to S3 are performed respectively, the amount of bleeding in the plurality of consecutive breathing cycles can be obtained, and the maximum value of the amount of bleeding is saved.

[0056] Therefore, the maximum value of the amount of bleeding is obtained and saved by real-time analysis of the continuous ultrasound images.

[0057] In some embodiments of the present application, after step S3, the following steps are further included, as shown in Figure 7 .

[0058] Step S4, judging the bleeding grade based on the calculated amount of bleeding, and outputting.

[0059] Based on the calculated amount of bleeding, the bleeding grade is judged and outputted, so as to facilitate knowing the bleeding degree of the injured person.

[0060] In some embodiments of the present application, step S4 specifically includes: If the calculated amount of bleeding is within the first set range, the bleeding level is determined to be a first-grade bleeding (a small amount of bleeding) ; If the calculated amount of bleeding is within the second set range, the bleeding level is determined to be a second-grade bleeding (a medium amount of bleeding) ; wherein any value within the first set range is smaller than any value within the second set range.

[0061] According to the range to which the amount of bleeding belongs, the bleeding level is divided into two grades, which is simple and convenient to know the bleeding degree of the wounded.

[0062] For example, if the weighted average of the physical value of the bleeding depth < 60 mm, it is a small amount of bleeding; if the weighted average of the physical value of the bleeding depth ≥ 60 mm, it is a medium-large amount of bleeding.

[0063] That is, if the amount of bleeding < 60 x a, it is a small amount of bleeding (first-grade bleeding) ; if the amount of bleeding ≥ 60 x a, it is a medium-large amount of bleeding (second-grade bleeding).

[0064] Next, the steps of the body cavity bleeding detection method will be described in detail in combination with Figure 8 .

[0065] (1) Start collecting.

[0066] (2) Get the ultrasound image.

[0067] (3) Get the ultrasound effective area.

[0068] (4) Detect the bleeding area.

[0069] (5) Detect the maximum up-down diameter of the bleeding area. That is, the length of the upper boundary of the minimum inscribed rectangle frame, that is, the pixel value of the bleeding depth.

[0070] (6) Calculate the physical value of the up-down diameter. That is, the physical value of the bleeding depth.

[0071] (7) Real-time respiratory monitoring. That is, according to the respiratory phase for compensation, get the weighted average of the physical value of the bleeding depth d com .

[0072] (8) Calculate the amount of bleeding according to the formula. The amount of bleeding = d com x a.

[0073] (9) According to the amount of bleeding, the bleeding level is divided.

[0074] (10) Output the amount of bleeding and the bleeding level.

[0075] (11) End the examination.

[0076] The body cavity in the present application includes the internal cavity of the body such as the thoracic cavity and the abdominal cavity. The body cavity bleeding detection method of the present application is suitable for detecting the amount of bleeding in the internal cavity of the body such as the thoracic cavity and the abdominal cavity. For example, the acquired ultrasonic image is a longitudinal section image of the thoracic cavity, and based on the longitudinal section image of the thoracic cavity, steps S1-S4 are performed to realize the detection of the amount of thoracic cavity bleeding.

[0077] The body cavity bleeding detection method of the present application is a rapid and accurate thoracic cavity bleeding quantitative analysis method, which can enable non-ultrasound professionals in a trauma emergency scene to obtain accurate thoracic cavity bleeding amount and help the treatment of the injured personnel.

[0078] The present application provides a real-time thoracic cavity bleeding automatic measurement method for multiple models in a critical condition scene. The method is designed from a clinical perspective, the amount of bleeding is measured by automatic detection technology, and accurate bleeding amount value is obtained.

[0079] The body cavity bleeding detection method of the present application combines the critical condition scene and designs a thoracic cavity bleeding automatic measurement algorithm that meets the clinical needs. The longitudinal section image of the thoracic cavity is input, the bleeding area is automatically identified, the longitudinal maximum diameter (pixel value of bleeding depth) is calculated, the maximum diameter in the image is automatically converted into a physical value, and the amount of thoracic cavity bleeding is calculated to provide the user with accurate bleeding amount calculation value.

[0080] The body cavity bleeding detection method of the present application combines the critical condition scene and clinical knowledge to design a real-time thoracic cavity bleeding detection algorithm. The bleeding depth can be calculated in real time, and the amount of bleeding can be calculated. The present application considers the differences between multiple models, can effectively calculate the amount of bleeding, and give an effective detection report.

[0081] In the implementation of the present application, the injured personnel is examined in a sitting position, the abdominal convex array probe is used as the examination equipment, the detection depth of the probe is set to a fixed value, the medical staff scans from the posterior axillary line and the inferior angle of scapula from top to bottom, and the ultrasonic image is acquired; during the scanning process, the bleeding area is detected in real time, the physical value of the bleeding depth is obtained according to the conversion, the respiratory compensation value (weighted average value of the physical value of the bleeding depth) is calculated through real-time respiratory monitoring, the amount of bleeding is calculated, the maximum amount of bleeding is saved, after all the examinations are completed, the examination report is output according to the amount of bleeding and the set bleeding grade. Embodiment two,

[0082] Based on the design of the body cavity bleeding detection method of the above embodiment one, the present embodiment two provides a body cavity bleeding detection device, which includes a bleeding depth physical value acquisition module, a weighted calculation module, a bleeding amount calculation module, etc., as shown in Figure 9 .

[0083] the bleeding depth physical value acquisition module is configured to: acquire an ultrasound effective region of the ultrasound image in different breathing phases in a breathing cycle respectively, and obtain a physical length and a pixel length of a center line of the ultrasound effective region; acquire a bleeding region of the ultrasound image; acquire an overlapping region of the ultrasound effective region and the bleeding region, and acquire a minimum inscribed rectangle frame of the overlapping region; take a length of an upper boundary of the minimum inscribed rectangle frame as a pixel value of the bleeding depth, and calculate a physical value of the bleeding depth based on a ratio of the physical length to the pixel length of the center line of the ultrasound effective region. Thus, the bleeding depth physical value acquisition module obtains the physical value of the bleeding depth of the ultrasound image in different breathing phases in a breathing cycle.

[0084] the weighted calculation module is configured to: calculate a weighted average value of the bleeding depth physical value based on the physical value of the bleeding depth of the ultrasound image in different breathing phases and the weight corresponding to the breathing phase.

[0085] the bleeding amount calculation module is configured to: calculate the bleeding amount based on the weighted average value of the bleeding depth physical value.

[0086] In some embodiments of the present application, the bleeding region of the ultrasound image is acquired, specifically including: inputting the ultrasound image into a pre-trained deep learning model, and outputting the bleeding region of the ultrasound image by the deep learning model.

[0087] In some embodiments of the present application, the body cavity bleeding monitoring device further includes: the bleeding grade judgment module is configured to: judge the bleeding grade based on the calculated bleeding amount, and output.

[0088] In some embodiments of the present application, the bleeding grade judgment module is specifically configured to: if the calculated bleeding amount is within a first set range, the bleeding grade is determined to be first-grade bleeding; if the calculated bleeding amount is within a second set range, the bleeding grade is determined to be second-grade bleeding; wherein any value within the first set range is smaller than any value within the second set range.

[0089] The working process of the specific body cavity bleeding detection device has been described in detail in the above body cavity bleeding detection method, and will not be repeated here.

[0090] The body cavity bleeding detection device of the embodiment obtains the ultrasound effective region of the ultrasound image, and obtains the physical length and the pixel length of the center line of the ultrasound effective region; obtains the bleeding area of the ultrasound image; obtains the overlapping area of the ultrasound effective region and the bleeding area, and obtains the minimum inscribed rectangular frame of the overlapping area; takes the length of the upper boundary of the minimum inscribed rectangular frame as the pixel value of the bleeding depth, calculates the physical value of the bleeding depth based on the ratio of the physical length and the pixel length of the center line of the ultrasound effective region; the ultrasound image in different respiratory phases in the respiratory cycle is executed respectively The above steps obtain the physical value of the bleeding depth of the ultrasound image in different respiratory phases; then based on the physical value of the bleeding depth of the ultrasound image in different respiratory phases, and the weight corresponding to the respiratory phase, the weighted average value of the physical value of the bleeding depth is calculated; the bleeding volume is calculated based on the weighted average value of the physical value of the bleeding depth; therefore, the body cavity bleeding detection device of the embodiment calculates the bleeding volume according to the weighted average value of the physical value of the bleeding depth, realizes accurate calculation of the bleeding volume, and solves the technical problem of low accuracy of bleeding volume detection in the prior art.

[0091] The above embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, the technical solutions recorded in the foregoing embodiments can still be modified by those of ordinary skill in the art, or some technical features can be replaced by equivalent ones; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions claimed by the present application.

Claims

1. A method of detecting bleeding in a body cavity, characterized by: The method comprises the following steps: Step S1: for the ultrasound images in different respiratory phases in a respiratory cycle, the following steps S1-1 to S1-4 are performed respectively to obtain the physical value of the bleeding depth of the ultrasound images in different respiratory phases: Step S1-1: obtaining the ultrasound effective region of the ultrasound image, and obtaining the physical length and pixel length of the center line of the ultrasound effective region; Step S1-2: obtaining the bleeding area of the ultrasound image; Step S1-3: obtaining the overlapping region of the ultrasound effective region and the bleeding area, and obtaining the minimum inscribed rectangular frame of the overlapping region; Step S1-4: taking the length of the upper boundary of the minimum inscribed rectangular frame as the pixel value of the bleeding depth, and calculating the physical value of the bleeding depth based on the ratio of the physical length to the pixel length of the center line of the ultrasound effective region; Step S2: calculating the weighted average value of the physical value of the bleeding depth based on the physical value of the bleeding depth of the ultrasound images in different respiratory phases and the weight corresponding to the respiratory phase; Step S3: calculating the amount of bleeding based on the weighted average value of the physical value of the bleeding depth.

2. The method according to claim 1, wherein: The step S1-2 specifically comprises: inputting the ultrasound image into a pre-trained deep learning model, and the deep learning model outputs the bleeding area of the ultrasound image.

3. The method according to claim 1, wherein: The respiratory cycle comprises three respiratory phases: end-expiratory phase, end-inspiratory phase, and intermediate phase; and the weights corresponding to different respiratory phases are different.

4. The method according to claim 1, wherein: The step S3 specifically comprises: taking the product of the weighted average value of the physical value of the bleeding depth and a preset coefficient as the amount of bleeding.

5. The method according to claim 1, wherein: After step S3, the method further comprises: Step S4: judging the bleeding grade based on the calculated amount of bleeding and outputting.

6. The method according to claim 5, wherein: The step S4 specifically comprises: if the calculated amount of bleeding is within a first set range, determining that the bleeding grade is grade one bleeding; if the calculated amount of bleeding is within a second set range, determining that the bleeding grade is grade two bleeding; wherein any value within the first set range is smaller than any value within the second set range.

7. A body lumen bleed detection apparatus, characterized by: The method comprises the following steps: The bleeding depth physical value acquisition module is configured to: for the ultrasound images in different respiratory phases in a respiratory cycle, respectively obtaining the ultrasound effective region of the ultrasound image, and obtaining the physical length and pixel length of the center line of the ultrasound effective region; obtaining the bleeding area of the ultrasound image; obtaining the overlapping region of the ultrasound effective region and the bleeding area, and obtaining the minimum inscribed rectangular frame of the overlapping region; taking the length of the upper boundary of the minimum inscribed rectangular frame as the pixel value of the bleeding depth, and calculating the physical value of the bleeding depth based on the ratio of the physical length to the pixel length of the center line of the ultrasound effective region. a weighting calculation module configured to calculate a weighted average of the physical values of the blood depth based on the physical values of the blood depth in different respiratory phases and the weights corresponding to the respiratory phases; a blood volume calculation module configured to calculate the blood volume based on the weighted average of the physical values of the blood depth.

8. The body cavity hemorrhage detection device according to claim 7, characterized in that: the blood region in the ultrasound image is obtained by inputting the ultrasound image into a pre-trained deep learning model, and the deep learning model outputs the blood region in the ultrasound image.

9. The body cavity hemorrhage detection device according to claim 7, characterized in that: the body cavity hemorrhage monitoring device further comprises: a blood grade judgment module configured to judge the blood grade based on the calculated blood volume and output the blood grade.

10. The body cavity hemorrhage detection device according to claim 9, characterized in that: the blood grade judgment module is specifically configured to: if the calculated blood volume is within a first set range, determine that the blood grade is a first-grade hemorrhage; if the calculated blood volume is within a second set range, determine that the blood grade is a second-grade hemorrhage; wherein any value within the first set range is less than any value within the second set range.

Images