Thyroid post-operation hematoma early warning method and system based on multi-modal imaging, terminal and storage medium
By constructing a coordinate system using multimodal imaging technology and combining skin color compensation and blood flow velocity analysis, the problems of accuracy and timeliness in monitoring hematoma after thyroid surgery have been solved. This enables precise hematoma early warning and grading decision-making, reduces complications, and improves patient recovery.
Patent Information
- Application Number
- CN202511152719.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-18
- Publication Date
- 2025-12-09
AI Technical Summary
Current technologies for monitoring post-thyroidectomy hematoma rely on doctors' clinical experience and patients' self-observation, lacking multimodal information fusion and analysis, resulting in inaccurate and untimely hematoma early warning and grading decisions.
A multimodal imaging method is adopted, which constructs multiple coordinate systems and combines skin color compensation algorithm and blood flow velocity analysis to obtain information of multiple imaging modes. Time alignment processing is then performed to obtain hematoma early warning information.
It improves the accuracy and timeliness of postoperative hematoma monitoring after thyroid surgery, reduces complications, and enhances the quality of patient recovery.
Smart Images
Figure CN121096583A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of image detection technology, and in particular to a method, system, terminal, and computer-readable storage medium for early warning of postoperative hematoma in thyroid surgery based on multimodal imaging. Background Technology
[0002] Postoperative hematoma after thyroidectomy manifests as neck swelling, increased wound bleeding, and rapid breathing. Once the hematoma forms, it can compress the trachea, causing airway narrowing and obstruction, leading to difficulty breathing, and in severe cases, suffocation.
[0003] However, traditional hematoma monitoring methods mainly rely on doctors' clinical experience and patients' self-observation, which have the disadvantages of being highly subjective and not very accurate. While existing imaging technologies can assist doctors in monitoring hematomas, they often only provide a single imaging mode and lack the fusion and analysis of multimodal information, resulting in inaccurate and untimely early warning and grading decisions for hematomas.
[0004] Therefore, existing technologies still need to be improved and developed. Summary of the Invention
[0005] The main objective of this invention is to provide a method, system, terminal, and computer-readable storage medium for early warning of postoperative hematoma in thyroid surgery based on multimodal imaging. This invention aims to solve the problem that the existing technology for monitoring postoperative hematoma has a single imaging mode and lacks the fusion and analysis of multimodal information, resulting in inaccurate and untimely early warning and grading decisions for hematoma.
[0006] To achieve the above objectives, the present invention provides a method for early warning of postoperative hematoma after thyroid surgery based on multimodal imaging. The method includes the following steps:
[0007] Multiple post-thyroidectomy images are acquired, and multiple coordinate systems are constructed for each of the post-thyroidectomy images. Based on the post-thyroidectomy images in each coordinate system, the first swelling information of the target patient is defined.
[0008] The skin color compensation algorithm was used to analyze all the post-thyroidectomy images, and the second swelling information of the target patient was obtained by combining the timestamp of each coordinate system.
[0009] Obtain the blood flow velocity of the target patient at all the timestamps, and obtain the third swelling information of the target patient based on the blood flow velocity;
[0010] The first swelling information, the second swelling information, and the third swelling information are time-aligned to obtain hematoma warning information under each timestamp.
[0011] Optionally, the method for early warning of post-thyroid hematoma based on multimodal imaging, wherein acquiring multiple post-thyroid images, constructing multiple coordinate systems for each post-thyroid image, and defining first swelling information for the target patient based on the post-thyroid image in each coordinate system, further includes:
[0012] Multiple initial images of a target patient after thyroid surgery are acquired using a depth camera, and multiple image feature points are extracted from each initial image using a convolutional neural network.
[0013] Based on all the image feature points in each initial image, track the displacement field of the thyroid nodules in each initial image;
[0014] Based on the optical flow method, the adjacent displacements between each displacement field are estimated, and deformation compensation is performed on multiple initial images according to each adjacent displacement;
[0015] The adjacent displacements refer to the displacement difference between two adjacent initial images.
[0016] Optionally, the method for early warning of post-thyroid hematoma based on multimodal imaging, wherein acquiring multiple post-thyroid images, constructing multiple coordinate systems for each post-thyroid image, and defining first swelling information for the target patient based on the post-thyroid image in each coordinate system, specifically includes:
[0017] Acquire multiple postoperative thyroid images at different time stamps, and determine the location of the thyroid cartilage in the target patient based on the postoperative thyroid images at each time stamp;
[0018] The post-thyroidectomy images are obtained by deform compensation of the initial images, and the number of the initial images is one more than the number of post-thyroidectomy images.
[0019] Based on the location of the thyroid cartilage in each of the post-thyroidectomy images, a personalized anatomical coordinate system for the target patient at each of the timestamps is constructed.
[0020] Based on the personalized anatomical coordinate system in each of the post-thyroidectomy images, the degree of swelling between each of the post-thyroidectomy images is analyzed, and based on each degree of swelling, the first swelling information of the target patient is defined;
[0021] The first swelling information is used to represent the swelling process of the target patient across all timestamps.
[0022] Optionally, the method for early warning of post-thyroid hematoma based on multimodal imaging, wherein acquiring multiple post-thyroid images, constructing multiple coordinate systems for each post-thyroid image, and defining first swelling information for the target patient based on the post-thyroid image in each coordinate system, further includes:
[0023] Obtain ecchymosis information and diameter information from each of the post-thyroidectomy images, analyze each of the ecchymosis information and each of the diameter information, and obtain the ecchymosis growth rate and diameter growth rate of the target patient;
[0024] By analyzing the first swelling information, the ecchymosis growth rate, and the diameter growth rate, the incision tension information of the target patient is determined.
[0025] Optionally, the method for early warning of post-thyroid hematoma based on multimodal imaging, wherein the step of analyzing all the post-thyroid surgery images using a skin color compensation algorithm and combining the timestamps of each coordinate system to obtain the second swelling information of the target patient specifically includes:
[0026] Obtain a healthy skin color image of the target patient, extract the ultraviolet light band from the healthy skin color image, and extract a compensation color chart based on the ultraviolet light band;
[0027] The compensation color plate is fused with each of the post-thyroidectomy images to perform skin color compensation on all the post-thyroidectomy images, thereby obtaining the corresponding skin color image;
[0028] Analyze the skin color image corresponding to each timestamp to determine the color change information during the swelling process of the target patient, and obtain the second swelling information of the target patient based on the color change information.
[0029] Optionally, the method for early warning of post-thyroidectomy hematoma based on multimodal imaging, wherein obtaining the blood flow velocity of the target patient at all the timestamps and obtaining the third swelling information of the target patient based on the blood flow velocity specifically includes:
[0030] Obtain hemoglobin concentration distribution information and blood oxygen saturation corresponding to all the post-thyroidectomy images, and determine multiple blood flow velocities of the target patient between each time point based on each hemoglobin concentration distribution information and each blood oxygen saturation.
[0031] All blood flow velocities are evaluated according to preset criteria to obtain the third swelling information of the target patient.
[0032] Optionally, the post-thyroidectomy hematoma early warning method based on multimodal imaging, wherein performing time alignment processing on the first swelling information, the second swelling information, and the third swelling information to obtain hematoma early warning information for each timestamp specifically includes:
[0033] The swelling degree in the first swelling information, the color change information in the second swelling information, and the blood flow velocity in the third swelling information are time-aligned to obtain the swelling degree, color change information, and blood flow velocity at each timestamp.
[0034] Based on the degree of swelling, color change information, blood flow velocity, and incision tension information at each time stamp, the hematoma warning information for the target patient at each time stamp is determined.
[0035] Furthermore, to achieve the above objectives, the present invention also provides a post-thyroidectomy hematoma early warning system based on multimodal imaging, wherein the post-thyroidectomy hematoma early warning system based on multimodal imaging includes:
[0036] The first information analysis module is used to acquire multiple post-thyroidectomy images, construct multiple coordinate systems for each post-thyroidectomy image, and define the first swelling information of the target patient based on the post-thyroidectomy image in each coordinate system.
[0037] The second information analysis module is used to analyze all the post-thyroidectomy images using a skin color compensation algorithm and combine the timestamps of each coordinate system to obtain the second swelling information of the target patient.
[0038] The third information analysis module is used to obtain the blood flow velocity of the target patient at all the timestamps, and to obtain the third swelling information of the target patient based on the blood flow velocity;
[0039] The information integration module is used to perform time alignment processing on the first swelling information, the second swelling information, and the third swelling information to obtain hematoma warning information under each timestamp.
[0040] Furthermore, to achieve the above objectives, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a multimodal imaging-based postoperative hematoma early warning program stored in the memory and executable on the processor, wherein when the multimodal imaging-based postoperative hematoma early warning program is executed by the processor, it implements the steps of the multimodal imaging-based postoperative hematoma early warning method described above.
[0041] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a post-thyroid hematoma early warning program based on multimodal imaging, and the post-thyroid hematoma early warning program based on multimodal imaging, when executed by a processor, implements the steps of the post-thyroid hematoma early warning method based on multimodal imaging as described above.
[0042] In this invention, multiple post-thyroidectomy images are acquired. For each image, multiple coordinate systems are constructed, and a first swelling information for the target patient is defined based on the image in each coordinate system. A skin color compensation algorithm is used to analyze all the images, and combined with the timestamp of each coordinate system, a second swelling information for the target patient is obtained. The blood flow velocity of the target patient at all timestamps is acquired, and a third swelling information is obtained based on the blood flow velocity. The first, second, and third swelling information are time-aligned to obtain hematoma warning information for each timestamp. This invention combines multiple imaging modes, such as ultrasound and optical imaging, for real-time monitoring and intelligent early warning of post-thyroidectomy hematomas. By fusing and analyzing multimodal information, the size, location, and morphology of the hematoma can be accurately determined, thereby enabling hematoma grading decisions. This provides doctors with precise auxiliary diagnostic information, significantly improving the accuracy and timeliness of post-thyroidectomy hematoma monitoring, reducing complications, and improving the quality of patient recovery. Attached Figure Description
[0043] Figure 1 This is a flowchart of a preferred embodiment of the post-thyroidectomy hematoma early warning method based on multimodal imaging of the present invention;
[0044] Figure 2 This is a structural diagram of a preferred embodiment of the post-thyroidectomy hematoma early warning system based on multimodal imaging of the present invention;
[0045] Figure 3 This is a structural diagram of a preferred embodiment of the terminal of the present invention. Detailed Implementation
[0046] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
[0047] Hematomas can compress the recurrent laryngeal nerve and superior laryngeal nerve, exacerbating nerve damage symptoms such as hoarseness, decreased voice pitch, or choking when drinking. Local compression affects blood circulation, creating a vicious cycle of hematoma-ischemia-rebleed, expanding the hematoma's area. Prolonged compression can cause ischemia and edema of the neck's soft tissues, and even necrosis of the skin and subcutaneous tissue. Furthermore, hematomas increase the risk of incision infection or neck abscess.
[0048] Traditional hematoma monitoring methods mainly rely on doctors' clinical experience and patients' self-observation, which have the disadvantages of being highly subjective and not very accurate. While existing imaging technologies can assist doctors in monitoring hematomas, they often only provide a single imaging mode and lack the fusion and analysis of multimodal information, resulting in inaccurate and untimely early warning and grading decisions for hematomas.
[0049] The preferred embodiment of the present invention describes a method for early warning of post-thyroid hematoma based on multimodal imaging, such as... Figure 1 As shown, the post-thyroidectomy hematoma early warning method based on multimodal imaging includes the following steps:
[0050] Step S10: Obtain multiple post-thyroidectomy images, construct multiple coordinate systems for each post-thyroidectomy image, and define the first swelling information of the target patient based on the post-thyroidectomy image in each coordinate system.
[0051] For images captured by a depth camera, preprocessing is required to extract feature points from the images to compensate for the deformation of the hematoma, thereby improving the accuracy of hematoma monitoring. Specifically, multiple initial images of the target patient after thyroid surgery are acquired by the depth camera, and multiple image feature points are extracted from each initial image using a convolutional neural network. Based on all the image feature points in each initial image, the displacement field of the thyroid nodule in each initial image is tracked. The adjacent displacements between each displacement field are estimated based on optical flow, and deformation compensation is performed on the multiple initial images based on each adjacent displacement. Here, the adjacent displacement represents the displacement difference between two adjacent initial images.
[0052] For each initial image, multiple image feature points are extracted (e.g., for a target patient's entire incision site after thyroid surgery, multiple image feature points are extracted). For the same image feature point in two adjacent initial images, the displacement field of the thyroid nodule is tracked, and the adjacent displacements are estimated based on optical flow. Then, deformation compensation is performed to obtain multiple post-thyroid surgery images. This process can be implemented using the following code:
[0053] "def elastic_compensation(img_sequence):
[0054] Feature point extraction and matching
[0055] sift = cv2.SIFT_create()
[0056] kp1,des1=sift.detectAndCompute(img_sequence[0],None)
[0057] kp2,des2=sift.detectAndCompute(img_sequence[1],None)
[0058] Optical flow field calculation
[0059] flow=cv2.calcOpticalFlowFarneback(
[0060] prev = img_sequence[0],
[0061] next = img_sequence[1],
[0062] flow = None,
[0063] pyr_scale = 0.5,
[0064] levels=3,
[0065] winsize=15,
[0066] iterations=3,
[0067] poly_n = 5,
[0068] poly_sigma = 1.2,
[0069] flags=0 )
[0071] Deformation compensation
[0072] h,w = flow.shape[:2]
[0073] map_x=np.tile(np.arange(w),(h,1))
[0074] map_y=np.swapaxes(np.tile(np.arange(h),(w,1)),0,1)
[0075] map_xy=(flow+np.dstack((map_x,map_y))).astype('float32')
[0076] return cv2.remap(img_sequence[1],map_xy,None,cv2.INTER_LINEAR)".
[0077] The displacement field and deformation compensation process is achieved through two adjacent initial images. Therefore, the final post-thyroidectomy image will be obtained by discarding the first initial image based on all the initial images.
[0078] Specifically, multiple post-thyroidectomy images at different time stamps are acquired, and the location of the thyroid cartilage in the target patient is determined based on the post-thyroidectomy image at each time stamp. The post-thyroidectomy images are obtained by deforming the initial images, and the number of initial images is one more than the number of post-thyroidectomy images. A personalized anatomical coordinate system for the target patient at each time stamp is constructed based on the location of the thyroid cartilage in each post-thyroidectomy image. The degree of swelling between each post-thyroidectomy image is analyzed based on the personalized anatomical coordinate system, and a first swelling information for the target patient is defined based on each degree of swelling. This first swelling information represents the swelling process of the target patient across all time stamps.
[0079] In this process, each acquired post-thyroidectomy image is marked with a corresponding timestamp, thus obtaining a post-thyroidectomy image that changes over time. For each post-thyroidectomy image, a personalized anatomical coordinate system is constructed based on the location of the thyroid cartilage (bone landmarks such as the sternal notch) of the target patient to locate the post-thyroidectomy hematoma. Then, by using the located hematoma position, the degree of swelling between different post-thyroidectomy images can be determined. That is, by comparing the previous post-thyroidectomy image with the next post-thyroidectomy image, it can be determined whether the hematoma has increased in size between these two adjacent timestamps, thus obtaining the degree of swelling of the target patient's hematoma between these two timestamps.
[0080] Furthermore, information on ecchymosis and diameter is obtained from each post-thyroidectomy image, and each piece of information on ecchymosis and diameter is analyzed to obtain the ecchymosis growth rate and diameter growth rate of the target patient; the first swelling information, the ecchymosis growth rate and the diameter growth rate are analyzed to determine the incision tension information of the target patient.
[0081] In addition to monitoring the degree of swelling of the hematoma, it also acquires real-time information on the ecchymosis of the target patient (this information can be determined by the area of local color abnormalities, such as...). Figure 2 The system obtains the incision tension information of the target patient (as shown in the figure) and diameter information (specifically, the circumference growth rate as it changes with the timestamp). Then, based on this swelling-related information, the system obtains the incision tension information of the target patient. If the incision tension index exceeds the safety threshold, the system directly issues an early warning alarm to prompt caregivers to inspect the target patient's wound. Combining multiple swelling information to make a preliminary judgment improves the accuracy of monitoring hematoma after thyroid surgery.
[0082] Step S20: Analyze all the post-thyroidectomy images using a skin color compensation algorithm and combine them with the timestamp of each coordinate system to obtain the second swelling information of the target patient.
[0083] Specifically, a healthy skin color image of the target patient is acquired, the ultraviolet light band in the healthy skin color image is extracted, and a compensation color chart is extracted based on the ultraviolet light band; the compensation color chart is fused with each of the post-thyroidectomy images to perform skin color compensation on all the post-thyroidectomy images to obtain the corresponding skin color image; the skin color image corresponding to each timestamp is analyzed to determine the color change information during the swelling process of the target patient, and the second swelling information of the target patient is obtained based on the color change information.
[0084] The process involves first acquiring a color space image (e.g., a 2048×1536 resolution color space image) of the target patient's healthy skin area using a dermoscope, and simultaneously acquiring an ultraviolet (365 nm) wavelength image to enhance pigment contrast. Then, a skin color compensation algorithm is used to generate a compensated skin color, which is then fused with the mechanical energy of the post-thyroidectomy image to achieve skin color compensation. The skin color compensation algorithm can be implemented in the following ways:
[0085] "def skin_compensation(img_lab):
[0086] Extract from healthy skin areas (non-lesion areas)
[0087] healthy_mask=cv2.inRange(img_lab,
[0088] (70,-10,-10),
[0089] (90,10,10))#L*a*b*threshold range
[0090] Calculate the statistical characteristics of healthy skin color
[0091] mean,std=cv2.meanStdDev(img_lab,mask=healthy_mask)
[0092] Generate a compensation color palette
[0093] compensation_map=np.zeros_like(img_lab)
[0094] compensation_map[:,:,0]=mean[0]#Preserve the luminance channel
[0095] compensation_map[:,:,1:]=gaussian_filter(img_lab[:,:,1:],sigma=5)
[0096] "return compensation_map".
[0097] In particular, skin color compensation is applied to each post-thyroidectomy image to avoid misjudging the color of the hematoma due to skin color differences among target patients, thereby better monitoring the color of the hematoma and further improving the accuracy of monitoring the ecchymosis of the hematoma.
[0098] Step S30: Obtain the blood flow velocity of the target patient at all the timestamps, and obtain the third swelling information of the target patient based on the blood flow velocity.
[0099] Hematoma is mainly caused by postoperative bleeding (such as vascular injury, incomplete hemostasis or abnormal coagulation function), which leads to the accumulation of blood under the skin or in the surgical area. Therefore, the degree of hematoma in the target patient after thyroid surgery is judged by another dimension of blood flow velocity.
[0100] Specifically, the hemoglobin concentration distribution information and blood oxygen saturation corresponding to all the post-thyroidectomy images are obtained. Based on each of the hemoglobin concentration distribution information and each of the blood oxygen saturation, multiple blood flow velocities of the target patient between each time stamp are determined. All the blood flow velocities are evaluated according to preset standards to obtain the third swelling information of the target patient.
[0101] Edema is often caused by non-bleeding factors such as surgical stimulation (e.g., reactive edema of the fat layer caused by flap separation or the use of energy devices), lymphatic extravasation, inflammatory reactions, or fluid retention (e.g., excessive sodium intake). By examining the hemoglobin concentration distribution and blood oxygen saturation in the wound area using ultrasound, a hematoma will show a fluid-filled dark area with blood clot echoes, while edema is mostly diffuse hypoechoic or subcutaneous fluid accumulation without obvious blood clots. This difference can help distinguish between edema and hematoma. Furthermore, blood flow velocity can better identify changes in hematoma size.
[0102] Furthermore, edema and hematoma can be distinguished by skin color and texture. If, based on post-thyroidectomy images, the post-operative wound shows a hard, fixed area with purplish-red or ecchymotic local skin, it is judged as a hematoma. If it feels soft and movable, the skin is mostly normal or light red, without obvious ecchymosis, it is judged as edema. By judging from multiple dimensions such as appearance, texture, hemoglobin concentration distribution information, and blood oxygen saturation, hematoma and edema can be accurately distinguished, improving the accuracy of hematoma judgment and avoiding the waste of resources caused by misjudgment.
[0103] Step S40: Perform time alignment processing on the first swelling information, the second swelling information, and the third swelling information to obtain hematoma warning information under each timestamp.
[0104] Each swelling information is composed of information from different timestamps. Therefore, for each timestamp, different swelling information is time-aligned to obtain the overall swelling information under different dimensional combinations at each timestamp. This enables accurate judgment of the size, location, and shape of the hematoma.
[0105] Specifically, all swelling degrees in the first swelling information, all color change information in the second swelling information, and all blood flow velocities in the third swelling information are time-aligned to obtain the swelling degree, color change information, and blood flow velocity at each timestamp; based on the swelling degree, color change information, blood flow velocity, and incision tension information at each timestamp, the hematoma warning information for the target patient at each timestamp is determined.
[0106] After obtaining the first, second, and third swelling information and incision tension information, a weighted fusion process can be performed based on the specific circumstances of the target patient. For example, the coagulation function of target patients of different ages is different. If the coagulation function is strong, the weight of the third swelling information will be increased; if the coagulation function is weak, the weight of the third swelling information will be decreased. This process can improve the personalized monitoring of hematoma in target patients, thereby improving the monitoring accuracy for different users.
[0107] The monitoring frequency for the target patient can be set by caregivers or medical staff. For example, under normal circumstances, monitoring can be set for 24 hours, with testing every half hour for the first 6 hours, and the monitoring frequency gradually decreasing thereafter, from once every 1 hour to once every 3 hours.
[0108] The overall swelling information for each time stamp is assessed. If the overall swelling information at a particular time stamp indicates that the hematoma in the target patient is severe, an early warning message will be issued to alert caregivers to treat the patient's wound. At the same time, the target patient will be monitored at a consistent time stamp frequency, for example, every half hour. If no abnormalities are found, the target patient will be monitored within 24 hours as usual.
[0109] The system detects the overall swelling information at each time point and determines whether the overall swelling information at the current time point is abnormal. If an abnormality is found, the target patient's wound is treated immediately. In subsequent time points, if the overall swelling information becomes more severe, different measures will be taken accordingly, thereby achieving graded decision-making for the target patient's wound and improving the target patient's recovery efficiency.
[0110] Furthermore, for non-monitored areas in the image, blurring can be performed using human feature blurring algorithms (such as...). Figure 2 (as shown), thereby protecting patient privacy information.
[0111] This invention combines multiple imaging modes, such as ultrasound imaging and optical imaging, to monitor and provide intelligent early warning of hematomas after thyroid surgery in real time. By fusing and analyzing multimodal information, it can accurately determine the size, location, and shape of the hematoma, thereby enabling hematoma grading decisions. This provides doctors with precise auxiliary diagnostic information, significantly improving the accuracy and timeliness of monitoring hematomas after thyroid surgery, reducing the occurrence of complications, and improving the quality of patient recovery.
[0112] Furthermore, such as Figure 2 As shown, based on the above-mentioned method for early warning of postoperative hematoma after thyroid surgery based on multimodal imaging, the present invention also provides a system for early warning of postoperative hematoma after thyroid surgery based on multimodal imaging, wherein the system for early warning of postoperative hematoma after thyroid surgery based on multimodal imaging includes:
[0113] The first information analysis module 51 is used to acquire multiple post-thyroidectomy images, construct multiple coordinate systems for each post-thyroidectomy image, and define the first swelling information of the target patient based on the post-thyroidectomy image in each coordinate system.
[0114] The second information analysis module 52 is used to analyze all the post-thyroidectomy images using a skin color compensation algorithm and combine the timestamps of each coordinate system to obtain the second swelling information of the target patient.
[0115] The third information analysis module 53 is used to obtain the blood flow velocity of the target patient at all the timestamps, and to obtain the third swelling information of the target patient based on the blood flow velocity;
[0116] The information integration module 54 is used to perform time alignment processing on the first swelling information, the second swelling information and the third swelling information to obtain hematoma warning information under each timestamp.
[0117] Furthermore, such as Figure 3 As shown, based on the above-mentioned method and system for early warning of post-thyroid hematoma based on multimodal imaging, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 3 Only some of the terminal components are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.
[0118] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory. In other embodiments, the memory 20 may be an external storage device of the terminal, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc. Further, the memory 20 may include both internal and external storage devices. The memory 20 is used to store application software and various types of data installed on the terminal, such as program code installed on the terminal. The memory 20 may also be used to temporarily store data that has been output or will be output. In one embodiment, the memory 20 stores a post-thyroidectomy hematoma early warning program 40 based on multimodal imaging. This post-thyroidectomy hematoma early warning program 40 based on multimodal imaging can be executed by the processor 10, thereby implementing the post-thyroidectomy hematoma early warning method based on multimodal imaging in this application.
[0119] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in the memory 20 or process data, such as executing the post-thyroid hematoma early warning method based on multimodal imaging.
[0120] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface. The components of the terminal communicate with each other via a system bus.
[0121] In one embodiment, when the processor 10 executes the multimodal imaging-based post-thyroid hematoma early warning program 40 in the memory 20, the following steps are performed:
[0122] Multiple post-thyroidectomy images are acquired, and multiple coordinate systems are constructed for each of the post-thyroidectomy images. Based on the post-thyroidectomy images in each coordinate system, the first swelling information of the target patient is defined.
[0123] The skin color compensation algorithm was used to analyze all the post-thyroidectomy images, and the second swelling information of the target patient was obtained by combining the timestamp of each coordinate system.
[0124] Obtain the blood flow velocity of the target patient at all the timestamps, and obtain the third swelling information of the target patient based on the blood flow velocity;
[0125] The first swelling information, the second swelling information, and the third swelling information are time-aligned to obtain hematoma warning information under each timestamp.
[0126] The process of acquiring multiple post-thyroidectomy images, constructing multiple coordinate systems for each image, and defining the first swelling information of the target patient based on the post-thyroidectomy image in each coordinate system, further includes:
[0127] Multiple initial images of a target patient after thyroid surgery are acquired using a depth camera, and multiple image feature points are extracted from each initial image using a convolutional neural network.
[0128] Based on all the image feature points in each initial image, track the displacement field of the thyroid nodules in each initial image;
[0129] Based on the optical flow method, the adjacent displacements between each displacement field are estimated, and deformation compensation is performed on multiple initial images according to each adjacent displacement;
[0130] The adjacent displacements refer to the displacement difference between two adjacent initial images.
[0131] The process of acquiring multiple post-thyroidectomy images, constructing multiple coordinate systems for each image, and defining the first swelling information of the target patient based on the post-thyroidectomy image in each coordinate system specifically includes:
[0132] Acquire multiple postoperative thyroid images at different time stamps, and determine the location of the thyroid cartilage in the target patient based on the postoperative thyroid images at each time stamp;
[0133] The post-thyroidectomy images are obtained by deform compensation of the initial images, and the number of the initial images is one more than the number of post-thyroidectomy images.
[0134] Based on the location of the thyroid cartilage in each of the post-thyroidectomy images, a personalized anatomical coordinate system for the target patient at each of the timestamps is constructed.
[0135] Based on the personalized anatomical coordinate system in each of the post-thyroidectomy images, the degree of swelling between each of the post-thyroidectomy images is analyzed, and based on each degree of swelling, the first swelling information of the target patient is defined;
[0136] The first swelling information is used to represent the swelling process of the target patient across all timestamps.
[0137] The process includes acquiring multiple post-thyroidectomy images, constructing multiple coordinate systems for each image, and defining the first swelling information of the target patient based on the post-thyroidectomy image in each coordinate system. The process further includes:
[0138] Obtain ecchymosis information and diameter information from each of the post-thyroidectomy images, analyze each of the ecchymosis information and each of the diameter information, and obtain the ecchymosis growth rate and diameter growth rate of the target patient;
[0139] By analyzing the first swelling information, the ecchymosis growth rate, and the diameter growth rate, the incision tension information of the target patient is determined.
[0140] Specifically, the analysis of all the post-thyroidectomy images using a skin color compensation algorithm, combined with the timestamp of each coordinate system, to obtain the second swelling information of the target patient includes:
[0141] Obtain a healthy skin color image of the target patient, extract the ultraviolet light band from the healthy skin color image, and extract a compensation color chart based on the ultraviolet light band;
[0142] The compensation color plate is fused with each of the post-thyroidectomy images to perform skin color compensation on all the post-thyroidectomy images, thereby obtaining the corresponding skin color image;
[0143] Analyze the skin color image corresponding to each timestamp to determine the color change information during the swelling process of the target patient, and obtain the second swelling information of the target patient based on the color change information.
[0144] Specifically, obtaining the blood flow velocity of the target patient at all the specified time stamps, and obtaining the third swelling information of the target patient based on the blood flow velocity, includes:
[0145] Obtain hemoglobin concentration distribution information and blood oxygen saturation corresponding to all the post-thyroidectomy images, and determine multiple blood flow velocities of the target patient between each time point based on each hemoglobin concentration distribution information and each blood oxygen saturation.
[0146] All blood flow velocities are evaluated according to preset criteria to obtain the third swelling information of the target patient.
[0147] Specifically, the step of performing time alignment processing on the first swelling information, the second swelling information, and the third swelling information to obtain hematoma warning information under each timestamp includes:
[0148] The swelling degree in the first swelling information, the color change information in the second swelling information, and the blood flow velocity in the third swelling information are time-aligned to obtain the swelling degree, color change information, and blood flow velocity at each timestamp.
[0149] Based on the degree of swelling, color change information, blood flow velocity, and incision tension information at each time stamp, the hematoma warning information for the target patient at each time stamp is determined.
[0150] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a post-thyroid hematoma early warning program based on multimodal imaging, and the post-thyroid hematoma early warning program based on multimodal imaging implements the steps of the post-thyroid hematoma early warning method based on multimodal imaging as described above when executed by a processor.
[0151] In summary, this invention provides a method and related equipment for early warning of post-thyroid hematoma based on multimodal imaging. The method includes: acquiring multiple post-thyroid images; constructing multiple coordinate systems for each post-thyroid image; defining first swelling information for a target patient based on the post-thyroid images in each coordinate system; analyzing all the post-thyroid images using a skin color compensation algorithm and combining the timestamps of each coordinate system to obtain second swelling information for the target patient; acquiring the blood flow velocity of the target patient at all the timestamps and obtaining third swelling information based on the blood flow velocity; and performing time alignment processing on the first, second, and third swelling information to obtain hematoma early warning information for each timestamp. This invention combines multiple imaging modes such as ultrasound imaging and optical imaging to perform real-time monitoring and intelligent early warning of post-thyroid hematoma. By fusing and analyzing multimodal information, it can accurately determine the size, location, and morphology of the hematoma, thereby achieving hematoma grading decisions, providing doctors with accurate auxiliary diagnostic information, significantly improving the accuracy and timeliness of post-thyroid hematoma monitoring, reducing the occurrence of complications, and improving the quality of patient recovery.
[0152] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal that includes that element.
[0153] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.). The program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The computer-readable storage medium can be a memory, magnetic disk, optical disk, etc.
[0154] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.
Claims
1. A method for early warning of post-thyroidectomy hematoma based on multimodal imaging, characterized in that, The postoperative hematoma early warning method based on multimodal imaging includes: Multiple post-thyroidectomy images of the target patient are acquired. For each of the post-thyroidectomy images, multiple coordinate systems are constructed, and the first swelling information of the target patient is defined based on the post-thyroidectomy images in each coordinate system. The skin color compensation algorithm was used to analyze all the post-thyroidectomy images, and the second swelling information of the target patient was obtained by combining the timestamp of each coordinate system. Obtain the blood flow velocity of the target patient at all the timestamps, and obtain the third swelling information of the target patient based on the blood flow velocity; The first swelling information, the second swelling information, and the third swelling information are time-aligned to obtain hematoma warning information under each timestamp.
2. The method for early warning of post-thyroidectomy hematoma based on multimodal imaging according to claim 1, characterized in that, The process of acquiring multiple post-thyroidectomy images of the target patient, constructing multiple coordinate systems for each of the post-thyroidectomy images, and defining the target patient's first swelling information based on the post-thyroidectomy images in each coordinate system, further includes: Multiple initial images of a target patient after thyroid surgery are acquired using a depth camera, and multiple image feature points are extracted from each initial image using a convolutional neural network. Based on all the image feature points in each initial image, track the displacement field of the thyroid nodules in each initial image; Based on the optical flow method, the adjacent displacements between each displacement field are estimated, and deformation compensation is performed on multiple initial images according to each adjacent displacement; The adjacent displacements refer to the displacement difference between two adjacent initial images.
3. The method for early warning of postoperative hematoma after thyroid surgery based on multimodal imaging according to claim 2, characterized in that, The process involves acquiring multiple post-thyroidectomy images, constructing multiple coordinate systems for each image, and defining the first swelling information for the target patient based on the post-thyroidectomy image in each coordinate system. Specifically, this includes: Acquire multiple postoperative thyroid images at different time stamps, and determine the location of the thyroid cartilage in the target patient based on the postoperative thyroid images at each time stamp; The post-thyroidectomy images are obtained by deform compensation of the initial images, and the number of the initial images is one more than the number of post-thyroidectomy images. Based on the location of the thyroid cartilage in each of the post-thyroidectomy images, a personalized anatomical coordinate system for the target patient at each of the timestamps is constructed. Based on the personalized anatomical coordinate system in each of the post-thyroidectomy images, the degree of swelling between each of the post-thyroidectomy images is analyzed, and based on each degree of swelling, the first swelling information of the target patient is defined; The first swelling information is used to represent the swelling process of the target patient across all timestamps.
4. The method for early warning of postoperative hematoma after thyroid surgery based on multimodal imaging according to claim 3, characterized in that, The process involves acquiring multiple post-thyroidectomy images, constructing multiple coordinate systems for each image, and defining the first swelling information for the target patient based on the post-thyroidectomy image in each coordinate system. The process further includes: Obtain ecchymosis information and diameter information from each of the post-thyroidectomy images, analyze each of the ecchymosis information and each of the diameter information, and obtain the ecchymosis growth rate and diameter growth rate of the target patient; By analyzing the first swelling information, the ecchymosis growth rate, and the diameter growth rate, the incision tension information of the target patient is determined.
5. The method for early warning of post-thyroid hematoma based on multimodal imaging according to claim 1, characterized in that, The process of analyzing all the post-thyroidectomy images using a skin color compensation algorithm and combining the timestamps of each coordinate system to obtain the second swelling information of the target patient specifically includes: Obtain a healthy skin color image of the target patient, extract the ultraviolet light band from the healthy skin color image, and extract a compensation color chart based on the ultraviolet light band; The compensation color plate is fused with each of the post-thyroidectomy images to perform skin color compensation on all the post-thyroidectomy images, thereby obtaining the corresponding skin color image; Analyze the skin color image corresponding to each timestamp to determine the color change information during the swelling process of the target patient, and obtain the second swelling information of the target patient based on the color change information.
6. The method for early warning of post-thyroidectomy hematoma based on multimodal imaging according to claim 1, characterized in that, The step of obtaining the blood flow velocity of the target patient at all the time stamps, and obtaining the third swelling information of the target patient based on the blood flow velocity, specifically includes: Obtain hemoglobin concentration distribution information and blood oxygen saturation corresponding to all the post-thyroidectomy images, and determine multiple blood flow velocities of the target patient between each time point based on each hemoglobin concentration distribution information and each blood oxygen saturation. All blood flow velocities are evaluated according to preset criteria to obtain the third swelling information of the target patient.
7. The method for early warning of post-thyroid hematoma based on multimodal imaging according to claim 4, characterized in that, The step of performing time alignment processing on the first swelling information, the second swelling information, and the third swelling information to obtain hematoma warning information under each timestamp specifically includes: The swelling degree in the first swelling information, the color change information in the second swelling information, and the blood flow velocity in the third swelling information are time-aligned to obtain the swelling degree, color change information, and blood flow velocity at each timestamp. Based on the degree of swelling, color change information, blood flow velocity, and incision tension information at each time stamp, the hematoma warning information for the target patient at each time stamp is determined.
8. A post-thyroidectomy hematoma early warning system based on multimodal imaging, characterized in that, The postoperative hematoma early warning system based on multimodal imaging is applied to the postoperative hematoma early warning method based on multimodal imaging as described in any one of claims 1-7, wherein the postoperative hematoma early warning system based on multimodal imaging comprises: The first information analysis module is used to acquire multiple post-thyroidectomy images, construct multiple coordinate systems for each post-thyroidectomy image, and define the first swelling information of the target patient based on the post-thyroidectomy image in each coordinate system. The second information analysis module is used to analyze all the post-thyroidectomy images using a skin color compensation algorithm and combine the timestamps of each coordinate system to obtain the second swelling information of the target patient. The third information analysis module is used to obtain the blood flow velocity of the target patient at all the timestamps, and to obtain the third swelling information of the target patient based on the blood flow velocity; The information integration module is used to perform time alignment processing on the first swelling information, the second swelling information, and the third swelling information to obtain hematoma warning information under each timestamp.
9. A terminal, characterized in that, The terminal includes: a memory, a processor, and a multimodal imaging-based postoperative hematoma warning program for thyroid surgery, stored in the memory and executable on the processor. When executed by the processor, the multimodal imaging-based postoperative hematoma warning program for thyroid surgery implements the steps of the multimodal imaging-based postoperative hematoma warning method for thyroid surgery as described in any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a post-thyroid hematoma early warning program based on multimodal imaging, which, when executed by a processor, implements the steps of the post-thyroid hematoma early warning method based on multimodal imaging as described in any one of claims 1-7.