Thyroid nodule recognition device, method, storage medium and electronic equipment

Abstract

This application provides a thyroid nodule identification device, method, storage medium, and electronic device. The thyroid nodule identification device includes: an ultrasound video acquisition module for acquiring thyroid ultrasound video; a bounding box acquisition module for acquiring nodule bounding boxes of multiple frames in the thyroid ultrasound video, the nodule bounding boxes being used to identify the location and extent of the thyroid nodules; and an identification result acquisition module for acquiring keyframes and / or multiple nodules in the thyroid ultrasound video based on the nodule bounding boxes of the multiple frames. The thyroid nodule identification device can automatically identify keyframes and / or multiple nodules in a thyroid ultrasound video.

Description

Technical Field

[0001] This application belongs to the field of medical image processing technology, and relates to a nodule recognition device, and particularly to a thyroid nodule recognition device, method, storage medium and electronic device. Background Technology

[0002] Thyroid nodules are lumps or growths within the thyroid gland and are a common clinical condition. Most thyroid nodules are benign, but some can be malignant. Currently, examination of thyroid nodules typically includes ultrasound and blood tests to measure the levels of thyroid hormones and markers. Treatment for thyroid nodules generally depends on their size, type, growth pattern, and the results of diagnostic tests.

[0003] In the ultrasound diagnosis of thyroid nodules, medical personnel need to determine the number of nodules in the video and the most representative video frame (keyframe) based on the target detection results of thyroid nodules in each frame of the ultrasound video. This method is inefficient and prone to errors due to subjective mistakes by medical personnel. Summary of the Invention

[0004] This application provides a thyroid nodule identification device, method, storage medium, and electronic device to solve the aforementioned problems existing in the prior art.

[0005] In a first aspect, embodiments of this application provide a thyroid nodule identification device. The thyroid nodule identification device includes: an ultrasound video acquisition module for acquiring thyroid ultrasound video; a bounding box acquisition module for acquiring nodule bounding boxes of multiple frames in the thyroid ultrasound video, the nodule bounding boxes being used to identify the location and extent of the thyroid nodules; and an identification result acquisition module for acquiring keyframes and / or multiple nodules of the thyroid ultrasound video based on the nodule bounding boxes of the multiple frames.

[0006] In one implementation of the first aspect, the bounding box acquisition module uses a trained neural network model to process the thyroid ultrasound video to obtain the nodule bounding boxes of the multi-frame images.

[0007] In one implementation of the first aspect, the recognition result acquisition module includes: a similarity acquisition unit, used to acquire the similarity between the nodule bounding box of the first image and the nodule bounding box of the second image, wherein the first image and the second image are two frames of images of the thyroid ultrasound nodule; and a nodule recognition unit, used to determine whether the nodules in the first image and the second image are the same nodule or multiple nodules based on the similarity between the nodule bounding box of the first image and the nodule bounding box of the second image.

[0008] In one implementation of the first aspect, the similarity between the nodule bounding boxes of the first image and the nodule bounding boxes of the second image includes positional similarity, size similarity, and / or temporal similarity.

[0009] In one implementation of the first aspect, the second image is a first frame image that includes nodule bounding boxes preceding the first image.

[0010] In one implementation of the first aspect, the recognition result acquisition module includes a keyframe recognition unit, used to acquire keyframes of thyroid nodules based on the nodule bounding boxes of the multi-frame images.

[0011] Secondly, embodiments of this application provide a method for identifying thyroid nodules. The method includes: acquiring a thyroid ultrasound video; acquiring nodule bounding boxes from multiple frames of images in the thyroid ultrasound video, the nodule bounding boxes being used to identify the location and extent of the thyroid nodules; and acquiring keyframes and / or multiple nodules from the thyroid ultrasound video based on the nodule bounding boxes from the multiple frames of images.

[0012] In one implementation of the second aspect, obtaining multiple nodules in the thyroid ultrasound video based on the nodule bounding boxes of the multiple frames of images includes: obtaining the similarity between the nodule bounding boxes of the first image and the nodule bounding boxes of the second image, wherein the first image and the second image are two frames of images of the thyroid ultrasound nodules; and determining whether the nodules in the first image and the second image are the same nodule or multiple nodules based on the similarity between the nodule bounding boxes of the first image and the nodule bounding boxes of the second image.

[0013] Thirdly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the thyroid nodule identification method described in any of the second aspects of this application.

[0014] Fourthly, embodiments of this application provide an electronic device. The electronic device includes: a memory storing a computer program; and a processor communicatively connected to the memory, which, when the computer program is invoked, executes the thyroid nodule identification method according to any one of the second aspects of this application.

[0015] The thyroid nodule identification device provided in this application embodiment can automatically detect key frames and multiple nodules in thyroid ultrasound video. This process requires almost no human intervention, is highly efficient, and avoids errors caused by subjective mistakes of medical personnel. Attached Figure Description

[0016] Figure 1 The diagram shown is a structural schematic of an electronic device according to an embodiment of this application.

[0017] Figure 2 The diagram shown is a structural schematic of a thyroid nodule identification device according to an embodiment of this application.

[0018] Figure 3A This is shown as a training method for a neural network model in one embodiment of this application.

[0019] Figure 3B The diagram shown is an example of a nodule boundary frame in one embodiment of this application.

[0020] Figure 4 The diagram shown is a structural schematic of the identification result acquisition module in one embodiment of this application.

[0021] Figure 5 The flowchart shown is a method for identifying thyroid nodules in one embodiment of this application.

[0022] Figure 6 The diagram shown is a structural schematic of an electronic device according to an embodiment of this application.

[0023] Component designation explanation

[0024] 1. Electronic equipment

[0025] 11. Ultrasonic video acquisition device

[0026] 111 Ultrasonic probe

[0027] 12 General Purpose Processors

[0028] 121 Central Processing Unit

[0029] 122 Neural Network Processor

[0030] 13 Monitors

[0031] 14. Memory

[0032] 2. Thyroid nodule identification device

[0033] 21. Ultrasound Video Acquisition Module

[0034] 22 Bounding Box Acquisition Module

[0035] 23 Recognition Result Acquisition Module

[0036] 231 Similarity Acquisition Unit

[0037] 232 Nodule Identification Unit

[0038] 233 Keyframe Recognition Unit

[0039] 600 electronic devices

[0040] 610 Memory

[0041] 620 processor

[0042] 630 monitor

[0043] Steps S31 to S34

[0044] Steps S51 to S53 Detailed Implementation

[0045] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0046] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. Therefore, the drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0047] Thyroid nodules are lumps or hyperplasia within the thyroid gland, and are a common clinical condition. Thyroid nodules can be diagnosed using ultrasound. During ultrasound diagnosis of thyroid nodules, medical personnel need to determine the number of nodules in the video and the most representative video frame (keyframe) based on the target detection results of the thyroid nodules in each frame of the ultrasound video. This method is relatively inefficient and prone to errors due to subjective mistakes by medical personnel.

[0048] To address at least the aforementioned issues, this application provides a thyroid nodule identification device. This device includes an ultrasound video acquisition module, a bounding box acquisition module, and a recognition result acquisition module. It can automatically detect keyframes and multiple nodules in thyroid ultrasound videos. The detection process requires virtually no manual intervention, is highly efficient, and avoids errors caused by subjective mistakes by medical personnel.

[0049] The thyroid ultrasound diagnostic device provided in this application can be installed in an electronic device. The principle and implementation of the thyroid nodule identification device according to this application will be described in detail below with reference to the accompanying drawings, so that those skilled in the art can understand the thyroid nodule identification device provided in this application without creative effort.

[0050] Figure 1The diagram shown is a structural schematic of an electronic device 1 used in the thyroid nodule identification device provided in this application embodiment. Figure 1 As shown, the electronic device 1 includes an ultrasonic video acquisition device 11, at least one general-purpose processor 12, a display 13, and a memory 14.

[0051] The ultrasound video acquisition device 11 may include an ultrasound probe 111 for performing ultrasound scanning on the thyroid region of a patient to obtain thyroid ultrasound video. The ultrasound video acquisition device 11 is communicatively connected to a general-purpose processor 12, transmitting the acquired thyroid ultrasound video to the general-purpose processor 12. This transmission can be real-time or non-real-time, and this embodiment does not limit this.

[0052] The general-purpose processor 12 can be any type of device capable of processing electronic instructions. In this embodiment, the electronic device 1 may include one or more general-purpose processors 12, such as one or both of a central processing unit (CPU) 121 and a neural-network processing unit (NPU) 122. Furthermore, it may include one or more of a graphics processing unit (GPU), microprocessor, microcontroller, main processor, controller, and ASIC (Application Specific Integrated Circuit). The general-purpose processor 12 is configured to execute various types of digital storage instructions, such as software or firmware programs stored in memory 14, enabling the electronic device 1 to provide a variety of services.

[0053] The main function of the central processing unit 121 is to parse computer instructions and process data in computer software, realize the overall control of electronic device 1, and control and allocate all hardware resources of electronic device 1 (such as storage resources, communication resources, I / O interfaces, etc.).

[0054] Neural network processors are a general term for new types of processors based on neural network algorithms and acceleration. They are specifically designed for artificial intelligence to accelerate neural network operations and solve the problem of low efficiency of traditional chips in neural network operations.

[0055] It should be noted that the name of the neural network processor does not constitute a limitation of this application. For example, in other application scenarios, the neural network processor can also be modified or replaced by other processors with similar functions, such as a tensor processing unit (TPU), a deep learning processing unit (DPU), and so on.

[0056] The display 13 may specifically include a display screen (display panel). In some implementations, the display panel may be configured using a liquid crystal display (LCD), an organic light-emitting diode (OLED), or other similar forms. The display device may also be a touch panel (touchscreen, touch screen), which may include a display screen and a touch-sensitive surface. When the touch-sensitive surface detects a touch operation on or near it, it transmits the information to the central processing unit 121 to determine the type of touch event. Subsequently, the central processing unit 121 provides corresponding visual output on the display device based on the type of touch event.

[0057] Memory 14 may include volatile memory, such as random access memory (RAM) or cache. Memory 14 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state drive (SSD). Memory 14 can be used to store data such as images acquired by the ultrasonic video acquisition device 11. Memory 14 can also be used to store program instructions for the processor to call and execute corresponding algorithms.

[0058] Figure 2 The diagram shown is a structural schematic of the thyroid nodule identification device 2 provided in an embodiment of this application. Figure 2 As shown, the thyroid nodule identification device 2 provided in this application embodiment includes an ultrasound video acquisition module 21, a bounding box acquisition module 22, and an identification result acquisition module 23.

[0059] The ultrasound video acquisition module 21 is used to acquire ultrasound video of the thyroid gland. The thyroid ultrasound video is obtained through an ultrasound video acquisition device. An ultrasound video acquisition device is used in the medical field to capture and record ultrasound signals. The captured and recorded ultrasound signals can be used to generate medical images, helping doctors diagnose and monitor a patient's thyroid condition. An ultrasound video acquisition device typically includes an ultrasound transducer, an amplifier, and a data acquisition system. The ultrasound transducer emits high-frequency sound waves and monitors the sound waves reflected back from the patient's body. The data acquisition system captures and records these signals and converts them into images.

[0060] In some implementations, the ultrasound video acquisition module includes an ultrasound video acquisition device, which acquires ultrasound video of the patient's thyroid gland. In other implementations, the ultrasound video acquisition device is not included in the ultrasound video acquisition module. In this case, the ultrasound video acquisition module is communicatively connected to the ultrasound video acquisition device, which acquires the ultrasound video of the patient's thyroid gland and transmits it to the ultrasound video acquisition module.

[0061] The bounding box acquisition module 22 is connected to the ultrasound video acquisition module 21 and is used to acquire the nodule bounding boxes of multiple frames in the thyroid ultrasound video. The nodule bounding boxes are used to identify the location and extent of the thyroid nodules. The shape of the nodule bounding box can be, for example, rectangular, but this embodiment is not limited thereto.

[0062] The recognition result acquisition module 23 is connected to the bounding box acquisition module 22 and is used to acquire keyframes and / or multiple nodules in the thyroid ultrasound video based on the nodule bounding boxes of multiple frames of images. Specifically, the location and extent of thyroid nodules in each frame of images can be obtained based on the nodule bounding boxes of multiple frames of images, thereby determining whether the thyroid ultrasound video contains multiple nodules and the keyframes therein.

[0063] As can be seen from the above description, the thyroid nodule identification device 2 provided in this application embodiment can automatically detect key frames and multiple nodules in thyroid ultrasound video. This process basically does not require human intervention, is highly efficient, and will not result in errors caused by subjective mistakes of medical personnel.

[0064] In one embodiment of this application, the bounding box acquisition module uses a trained neural network model to process thyroid ultrasound video to obtain nodule bounding boxes for multiple frames. For example, the bounding box acquisition module can use the neural network model to process each frame of the thyroid ultrasound video to obtain nodule bounding boxes for each frame, but this application is not limited thereto.

[0065] Figure 3A This is a flowchart illustrating the training of a neural network model in an embodiment of this application. For example... Figure 3A As shown, the training process includes the following steps S31 to S34.

[0066] S31, a neural network model is constructed to segment thyroid ultrasound images to obtain bounding boxes for thyroid nodules. Specifically, a neural network model is an algorithm in the field of artificial intelligence that can be used to identify patterns and predict results. In this embodiment, the neural network model is based on the working mechanism of the human brain's nervous system and consists of an input layer, a hidden layer, and an output layer. Each layer has multiple nodes, and each node is a small computational unit. This neural network model learns to identify relevant features of thyroid nodules from input data and make predictions by learning from a large amount of data. In this embodiment, the YOLO deep learning model can be used to construct this neural network model, but this application is not limited to this.

[0067] S32, Obtain training data, which can be ultrasound images with nodule bounding boxes labeled. Figure 3B The image shown is an example of an ultrasound image with nodule bounding boxes labeled in an embodiment of this application.

[0068] S33, train the neural network model using training data. For example, in this embodiment, the neural network model can be trained using methods such as backpropagation algorithm, gradient descent algorithm, and transfer learning. In this embodiment, the above training methods can be used individually or in combination.

[0069] S34. Test the trained neural network model. For example, in this embodiment, holdout validation, cross validation, and / or automated evaluation metrics can be used to test the neural network model, thereby evaluating its performance and allowing for adjustments and optimizations based on the evaluation results.

[0070] Figure 4 The diagram shown illustrates a structural schematic of the recognition result acquisition module 23 in one embodiment of this application, where N is a positive integer. For example... Figure 4 As shown, in the embodiments of this application, the identification result acquisition module 23 may include a similarity acquisition unit 231 and a nodule identification unit 232.

[0071] The similarity acquisition unit 231 is used to acquire the similarity between the nodule bounding box of the first image and the nodule bounding box of the second image, wherein the first image and the second image are two frames of images in the thyroid ultrasound nodule.

[0072] In some embodiments, the second image is a first frame image that contains the nodule bounding box preceding the first image. Specifically, the first image is the nth frame image in a thyroid ultrasound video, where n>1. Tracing back from the first image, the frame image containing the nodule bounding box and closest to the first image is identified as the second image.

[0073] In some embodiments, the similarity between the nodule bounding boxes of the first image and the nodule bounding boxes of the second image includes positional similarity, size similarity, and / or temporal similarity. Positional similarity between nodule bounding boxes can be, for example, the degree of proximity of the two nodule bounding boxes in the ultrasound image, which can be represented by the distance between the center points of the two nodule bounding boxes or the intersection over union (IoU). Size similarity between nodule bounding boxes can be the similarity of the two nodule bounding boxes in both width and height dimensions, for example, by the absolute value of the difference in area between the two nodule bounding boxes. Temporal similarity of nodule bounding boxes refers to the similarity of the frame sequence in the thyroid ultrasound video, for example, by the difference in frame numbers between the frames containing the two nodule bounding boxes.

[0074] The nodule identification unit 232 is connected to the similarity acquisition unit 231 and is used to determine whether the nodules in the first image and the second image are the same nodule or multiple nodules based on the similarity between the nodule bounding boxes of the first image and the second image. Specifically, the higher the similarity between the nodule bounding box b1 of the first image and the nodule bounding box b2 of the second image, the more likely the thyroid nodules corresponding to b1 and b2 are the same nodule. Conversely, the lower the similarity between b1 and b2, the more likely b1 and b2 are different nodules.

[0075] Please continue reading. Figure 4 In one embodiment of this application, the recognition result acquisition module 23 may further include a keyframe recognition unit 233. The keyframe recognition unit 233 is connected to the similarity acquisition unit 231 and is used to acquire keyframes of thyroid nodules based on the nodule bounding boxes of multiple frames of images. Specifically, in this embodiment, keyframes can be extracted based on the nodule bounding boxes of the same thyroid nodule.

[0076] Optionally, the keyframe recognition unit 233 may select the middle frame of the image frame corresponding to all nodule bounding boxes as the keyframe.

[0077] Optionally, the keyframe recognition unit 233 may select the image frame corresponding to the nodal bounding box with the largest area among all nodal bounding boxes as the keyframe.

[0078] Optionally, the keyframe recognition unit 233 can employ a trained neural network model to obtain the Thyroid Imaging Reporting and Data System (TIRADS) scores for multiple frames in the thyroid ultrasound video, and use the TIRADS scores to assist in obtaining keyframes. The neural network model used to obtain the TIRADS scores and the neural network model used to obtain the nodule bounding boxes can be the same model or different models. Specifically, the keyframe recognition unit 233 can filter images based on the frequency or score of the TIRADS scores for multiple frames in the thyroid ultrasound video. For example, it can select the image with the highest score from multiple frames as the filtering result. If the filtering result is unique, the selected frame is the keyframe. If the filtering result is not unique, the keyframe recognition unit 233 can select the middle frame of all image frames corresponding to nodule bounding boxes as the keyframe, or select the image frame corresponding to the nodule bounding box with the largest area among all nodule bounding boxes as the keyframe. This method of obtaining keyframes has higher accuracy.

[0079] It should be noted that the above-described methods for obtaining keyframes are only feasible in some embodiments, but this application is not limited thereto.

[0080] Based on the above description of the thyroid nodule identification device, this application also provides a thyroid nodule identification method. Figure 5 The flowchart shown is a representation of the thyroid nodule identification method in this application. Figure 5 As shown, the thyroid nodule identification method provided in this application includes the following steps S51 to S53.

[0081] S51, acquire thyroid ultrasound video.

[0082] S52, Obtain the nodule bounding boxes from multiple frames of images in the thyroid ultrasound video. The nodule bounding boxes are used to identify the location and extent of the thyroid nodules.

[0083] S53, obtain keyframes and / or multiple nodules from the nodule bounding boxes of multi-frame images from the thyroid ultrasound video.

[0084] Optionally, in step S52, the trained neural network model can be used to process the thyroid ultrasound video to obtain nodule bounding boxes of multiple frames.

[0085] Optionally, obtaining multiple nodules from a thyroid ultrasound video based on nodule bounding boxes of multiple frames includes: obtaining the similarity between nodule bounding boxes in a first image and nodule bounding boxes in a second image, wherein the first image and the second image are two frames of images from a thyroid ultrasound nodule; and determining whether the nodules in the first image and the second image are the same nodule or multiple nodules based on the similarity between nodule bounding boxes in the first image and the second image.

[0086] Optionally, the similarity between the nodule bounding boxes of the first image and the nodule bounding boxes of the second image includes positional similarity, size similarity, and / or temporal similarity.

[0087] Optionally, the second image is a first frame image that contains nodule bounding boxes preceding the first image.

[0088] Optionally, in step S53, keyframes of the thyroid nodule can be obtained from the nodule bounding boxes of multiple frames of images.

[0089] It should be noted that the thyroid nodule identification method provided in this application embodiment is different from... Figure 2 The ultrasound video acquisition module 21, the bounding box acquisition module 22, and the recognition result acquisition module 23 in the thyroid nodule recognition device 2 shown correspond one-to-one, and will not be described in detail here.

[0090] This application also provides a computer-readable storage medium having a computer program stored thereon, which is implemented when executed by a processor. Figure 5 The method for identifying thyroid nodules is shown.

[0091] In the embodiments of this application, any combination of one or more storage media can be used. The storage medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, RAM, ROM, erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0092] This application also provides an electronic device. Figure 6 The diagram shown is a structural schematic of an electronic device 600 according to an embodiment of this application. Figure 6As shown, in this embodiment of the application, the electronic device 600 includes a memory 610 and a processor 620.

[0093] The memory 610 is used to store computer programs; preferably, the memory 610 includes various media that can store program code, such as ROM, RAM, magnetic disk, USB flash drive, memory card or optical disk.

[0094] The processor 620 is connected to the memory 610 and is used to execute the computer program stored in the memory 610 so that the electronic device 600 performs the thyroid nodule identification method.

[0095] Optionally, the processor 620 may be a general-purpose processor, including a central processing unit, a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0096] Optionally, the electronic device 600 in this embodiment may further include a display 630. The display 630 is communicatively connected to the memory 610 and the processor 620, and is used to display the relevant GUI interactive interface of the thyroid nodule identification method.

[0097] In the several embodiments provided in this application, it should be understood that the disclosed apparatus or method can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or modules or units may be electrical, mechanical, or other forms.

[0098] The modules / units described as separate components may or may not be physically separate. The components shown as modules / units may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules / units can be selected to achieve the objectives of the embodiments of this application, depending on actual needs. For example, the functional modules / units in the various embodiments of this application may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0099] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0100] In summary, the thyroid nodule identification device provided in this application can automatically detect keyframes and multiple nodules in thyroid ultrasound videos. This examination process requires virtually no manual intervention, is highly efficient, and avoids errors caused by subjective mistakes by medical personnel. Furthermore, the thyroid nodule identification device provided in this application can assist medical personnel in determining the number of thyroid nodules and keyframes in ultrasound videos, thereby improving diagnostic accuracy. Moreover, the TIRADS scoring method can be selected in the application embodiments to assist in obtaining keyframes, further improving the accuracy of the obtained keyframes. Therefore, this application effectively overcomes the various shortcomings of the prior art and has high industrial applicability.

[0101] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A thyroid nodule identification device, characterized in that, include: Ultrasound video acquisition module, used to acquire thyroid ultrasound video; The bounding box acquisition module is used to acquire the nodule bounding boxes of multiple frames in the thyroid ultrasound video, and the nodule bounding boxes are used to identify the location and extent of the thyroid nodules. The recognition result acquisition module is used to acquire key frames and multiple nodules from the thyroid ultrasound video based on the nodule bounding boxes of the multi-frame images; The recognition result acquisition module includes a similarity acquisition unit and a nodule recognition unit. The similarity acquisition unit is used to acquire the similarity between the nodule bounding box of the first image and the nodule bounding box of the second image, wherein the first image and the second image are two frames in the thyroid ultrasound video. The nodule recognition unit is used to determine whether the nodules in the first image and the second image are the same nodule or multiple nodules based on the similarity between the nodule bounding box of the first image and the nodule bounding box of the second image. The similarity between the nodule bounding box of the first image and the nodule bounding box of the second image includes positional similarity, size similarity, and temporal similarity, wherein the temporal similarity is represented by the difference between the frame number of the frame containing the nodule bounding box of the first image and the frame containing the nodule bounding box of the second image. The recognition result acquisition module further includes a keyframe recognition unit, which is used to: use a trained neural network model to acquire thyroid imaging reports and data system scores for multiple frames of images in the thyroid ultrasound video, and filter the multiple frames of images according to the scoring frequency or score. If the filtering result is a unique value, the selected frame image is the keyframe. If the filtering result is not a unique value, the middle frame of all image frames corresponding to nodule bounding boxes is selected as the keyframe, or the image frame corresponding to the nodule bounding box with the largest area among all nodule bounding boxes is selected as the keyframe.

2. The thyroid nodule identification device according to claim 1, characterized in that, The bounding box acquisition module uses a trained neural network model to process the thyroid ultrasound video to obtain the nodule bounding boxes of the multi-frame images.

3. The thyroid nodule identification device according to claim 1, characterized in that, The second image is a first frame image that contains nodule bounding boxes preceding the first image.

4. A method for identifying thyroid nodules, characterized in that, include: Obtain thyroid ultrasound video; Obtain nodule bounding boxes from multiple frames of images in the thyroid ultrasound video; the nodule bounding boxes are used to identify the location and extent of the thyroid nodules. Keyframes and multiple nodules in the thyroid ultrasound video are obtained based on the nodule bounding boxes of the multi-frame images; The process of obtaining multiple nodules in the thyroid ultrasound video based on the nodule bounding boxes of the multiple frames of images includes: obtaining the similarity between the nodule bounding boxes of the first image and the nodule bounding boxes of the second image, wherein the first image and the second image are two frames in the thyroid ultrasound video; determining whether the nodules in the first image and the second image are the same nodule or multiple nodules based on the similarity between the nodule bounding boxes of the first image and the nodule bounding boxes of the second image; the similarity between the nodule bounding boxes of the first image and the nodule bounding boxes of the second image includes positional similarity, size similarity and temporal similarity, wherein the temporal similarity is represented by the difference between the frame number of the frame containing the nodule bounding box of the first image and the frame containing the nodule bounding box of the second image; Obtaining the keyframe includes: using a trained neural network model to obtain thyroid imaging reports and data system scores for multiple frames of images in the thyroid ultrasound video, and filtering the multiple frames of images according to the scoring frequency or score. If the filtering result is a unique value, the selected frame image is the keyframe. If the filtering result is not a unique value, the middle frame of all image frames corresponding to nodule bounding boxes is selected as the keyframe, or the image frame corresponding to the nodule bounding box with the largest area among all nodule bounding boxes is selected as the keyframe.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the thyroid nodule identification method of claim 4.

6. An electronic device, characterized in that, The electronic device includes: A memory that stores a computer program; The processor, which is communicatively connected to the memory, executes the thyroid nodule identification method of claim 4 when calling the computer program.

Images