Ultrasonic device, radiotherapy equipment including the same and working method thereof

Abstract

The present invention discloses an ultrasonic device, a radiotherapy device including the ultrasonic device, and a working method thereof. The ultrasonic device includes: an ultrasonic probe and a coupling agent layer. The ultrasonic probe includes: an internal center element and an epoxy resin sealing layer coated on the outside of the internal center. The internal center element includes: a piezoelectric element array layer, an acoustic backing layer and an acoustic matching layer arranged on opposite sides of the piezoelectric element array layer. The piezoelectric element array layer has a built-in integrated circuit for collecting ultrasonic images. The acoustic backing layer is used to eliminate the resonance effect. The acoustic matching layer is used to increase the acoustic propagation effect to the skin. The coupling agent layer is a hydrogel elastomer. The coupling agent layer is used to attach the ultrasonic probe to the human skin. The ultrasonic device includes a three-layer internal center element and an epoxy resin sealing layer. It has low cost, high reliability, and improves the accuracy of radiotherapy. The coupling agent layer of the hydrogel elastomer can firmly adhere the ultrasonic probe to the human skin to maintain robust and lasting comfortable ultrasonic real-time imaging.

Description

Technical Field

[0001] The present invention belongs to the technical field of tumor radiotherapy based on image guidance, and in particular relates to an ultrasonic device, radiotherapy equipment including the ultrasonic device and a working method thereof. Background Art

[0002] Precision radiotherapy for tumors has become one of the important means of cancer treatment, and its application is becoming more and more extensive. The precise positioning of the tumor target area is the key to its therapeutic effect. However, the physiological movement of the thoracic organs, such as heartbeat, breathing, and vascular pulsation, will cause the tumor target area to move and deform, especially the thoracic and abdominal tumors are more significantly affected. The inaccuracy of the target area can not only cause the target area to be missed, but also may cause high-energy rays to move outside the target area or even to the dangerous organ area, thereby causing serious complications or sequelae. Therefore, by tracking the real-time movement state of the respiratory organs, respiratory movement compensation and management in radiotherapy is currently the most practical and promising technical route for precision radiotherapy for tumors. Among them, many medical imaging methods based on ray imaging have the disadvantages of poor real-time tumor imaging, additional doses of irradiation for patients, and invasive implantation of markers in the body, so their practicality is limited. The solution of external structured light body surface scanning has the following problems: visible light is difficult to penetrate the body membrane to obtain data, and the prediction algorithm is difficult to adapt to the serious nonlinearity of the internal and external respiratory movement of the thoracic and abdominal bodies, and the sample points are relatively few, which is difficult to meet the needs of both prediction accuracy and real-time performance in clinical radiotherapy.

[0003] Clinical-grade real-time continuous imaging of various internal organs in the body remains a key task and challenge in the field of oncology radiation. Zero-radiation ultrasound imaging enables clinicians to evaluate tissue and organ function and make diagnoses during radiotherapy. Existing ultrasound devices have the following problems:

[0004] First, traditional ultrasound imaging usually relies on installing bulky ultrasound probes on the skin through robotic systems or mechanical fixation devices, such as straps and tapes, which hinders patient mobility and the convenience and comfort of wearing, and makes it difficult to achieve stable imaging for a long time.

[0005] Second, in order to achieve good acoustic coupling between the ultrasound transducer (probe) and the patient's body, conventional ultrasound imaging usually applies a coupling agent (i.e., a fluid-type ultrasonic conductive glue) between the ultrasound transducer and the patient's body. During the application of such an ultrasound transducer to the patient's body, a large pressure is applied to expel bubbles from the coupling agent in order to establish the required acoustic contact.

[0006] In the typical clinical setting, ultrasound examinations are of relatively short duration (e.g., 1–5 minutes), meaning that the coupling agent does not have time to dry out or otherwise degrade, but in radiation therapy, typical tumor delivery times are 8–12 minutes, which may also make it difficult to obtain stable tumor images.

[0007] Third, ultrasound is only applicable to limited parts of the body, such as muscles, because the ultrasound probe is thick and requires high pressure. If a traditional ultrasound probe is mechanically installed on the neck to image the carotid artery, jugular vein, and vagus nerve, it may cause suffocation. In addition, applying high pressure can easily cause deformation of the skin, and the conduction of force may cause the tumor boundary to deform and be reflected in the ultrasound image, resulting in inaccurate prediction of tumor boundaries, tracking of blood vessels, etc., thus affecting the accuracy of radiotherapy. Summary of the invention

[0008] In order to solve the above technical problems, the present invention proposes an ultrasonic device, a radiotherapy device including the ultrasonic device and a working method thereof.

[0009] In order to achieve the above object, the technical solution of the present invention is as follows:

[0010] In one aspect, the present invention discloses an ultrasonic device, comprising:

[0011] The ultrasonic probe comprises: an internal central element and an epoxy resin sealing layer covering the outer side of the internal central element, the internal central element comprises: a piezoelectric element array layer, an acoustic backing layer and an acoustic matching layer arranged on opposite sides of the piezoelectric element array layer, the piezoelectric element array layer has a built-in integrated circuit for collecting ultrasonic images, the acoustic backing layer is used to eliminate resonance effects, and the acoustic matching layer is used to increase acoustic propagation effects to the skin;

[0012] The coupling agent layer is a hydrogel elastomer and is used to attach the ultrasound probe to human skin.

[0013] Based on the above technical solution, the following improvements can be made:

[0014] As a preferred solution, the hydrogel elastomer includes: a hydrogel and an elastic film wrapped around the outside of the hydrogel.

[0015] As a preferred solution, the hydrogel comprises: a chitosan-polyacrylamide interpenetrating polymer network and water.

[0016] As a preferred solution, the thickness of the elastomeric film is less than 40 mm.

[0017] In another aspect, the present invention further discloses a radiotherapy device, comprising:

[0018] Any of the above-mentioned ultrasonic devices, the ultrasonic device is attached to human skin, and the ultrasonic device is used to collect ultrasonic images;

[0019] A host computer, the host computer is electrically connected to the ultrasonic probe of the ultrasonic device, and the host computer is used to process the ultrasonic image collected by the ultrasonic device and extract key information;

[0020] The accelerator is electrically connected to the host computer and is used for radiotherapy.

[0021] As a preferred solution, the host computer has a built-in SA-UNet neural network model, which includes an encoder, a spatial attention module and a decoder connected in sequence. The encoder is used to obtain image features of ultrasound images, the spatial attention module is used to further obtain fine-grained features of ultrasound images, and the decoder is used to reconstruct features and output results.

[0022] In addition, on the other hand, the present invention further discloses a working method of a radiotherapy device, which is used to implement the working of any of the above-mentioned radiotherapy devices, and specifically comprises the following steps:

[0023] Step 1: The ultrasound device collects ultrasound images;

[0024] Step 2: The ultrasound image collected by the ultrasound device is sent to the host computer, which processes it and extracts key information;

[0025] Step 3: The accelerator controls the grating movement and performs radiotherapy based on the key information extracted by the host computer.

[0026] As a preferred solution, the host computer has a built-in SA-UNet neural network model. The training of the SA-UNet neural network model includes the following steps:

[0027] S1: Annotate the collected ultrasound images;

[0028] S2: The original ultrasound images and the annotated ultrasound images are matched one by one, and divided into training set, test set and validation set according to a certain ratio;

[0029] S3: Set the training parameters of the SA-UNet neural network model, input the training set into the SA-UNet neural network model for forward propagation, and obtain the label result;

[0030] S4: Compare the label result obtained in S3 with the true label to obtain the classification loss function and Dice evaluation coefficient;

[0031] S5: Based on the classification loss function and Dice evaluation coefficient obtained in S4, the SA-UNet neural network model is back-propagated to optimize the parameters of the SA-UNet neural network model;

[0032] S6: Repeat S3-S5 until the training results of the SA-UNet neural network model meet expectations.

[0033] As a preferred solution, in S3, after data enhancement processing is performed on the training set, it is input into the SA-UNet neural network model for forward propagation.

[0034] As a preferred solution, the SA-UNet neural network model includes an encoder, a spatial attention module, and a decoder connected in sequence. The operation of the SA-UNet neural network model includes the following steps:

[0035] A1: The encoder obtains the image features of the ultrasound image; A2: The image features obtained by the encoder are transmitted to the spatial attention module and input into the global maximum pooling and global average pooling layers respectively;

[0036] A3: Perform feature concatenation on the output feature maps obtained from the two branches;

[0037] A4: The concatenated feature maps are subjected to 7×7 convolution, batch normalization layer and Sigmoid activation function to obtain the intermediate feature map;

[0038] A5: Multiply the feature map obtained by the encoder and the intermediate feature map pixel by pixel to obtain the final spatial attention map;

[0039] A6: Pass the spatial attention map as input feature to the decoder;

[0040] A7: The decoder reconstructs features and outputs the results.

[0041] The present invention provides an ultrasonic device, a radiotherapy device including the ultrasonic device and a working method thereof, wherein the ultrasonic device includes: an ultrasonic probe and a coupling agent layer, and has the following beneficial effects:

[0042] First, the ultrasound probe includes: an internal center element and an epoxy resin sealing layer wrapped around the outside of the internal center. The internal center element includes: a piezoelectric element array layer, an acoustic backing layer and an acoustic matching layer arranged on opposite sides of the piezoelectric element array layer, and the outermost layer is an epoxy resin sealing layer, which has high stability and reliability in long-term applications. At the same time, the three-layer internal center element reduces the cost and weight of the ultrasound probe, so that when the probe covers the human skin, on the one hand, it reduces the pressure caused by the ultrasound probe on the patient and reduces the discomfort, and on the other hand, it makes the skin deformation caused by pressure less severe, which improves the accuracy of radiotherapy to a certain extent.

[0043] Second, the coupling agent layer is a hydrogel elastomer, which is soft, tough, resistant to dehydration and has good bioadhesion. It can adhere firmly to human skin, and can maintain a stable position with slight body movements during radiotherapy, and the reliability of the ultrasound probe is high. The coupling agent layer can effectively transmit sound waves, protect the ultrasound probe from skin deformation, and maintain robust and lasting comfortable ultrasound real-time imaging.

[0044] Third, the hydrogel elastomer includes: a hydrogel and an elastic film wrapped around the outside of the hydrogel, and the elastic film can effectively prevent the hydrogel from dehydrating.

[0045] Fourth, the ultrasound probe is connected to the host computer to obtain real-time and continuous ultrasound images. The host computer processes the ultrasound images, extracts the location information of the tumor, predicts the breathing curve, etc., and feeds back to the accelerator to control the movement of the grating and achieve precise radiotherapy of the tumor.

[0046] Fifth, the present invention adopts the SA-UNet neural network model to combine the original information such as texture and spatial position with the image semantic information, and the overall network is lightweight and convenient. Moreover, a spatial attention module is added between the encoder and the decoder to improve the segmentation effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0048] Figure 1 A schematic diagram of the structure of an ultrasonic device provided in an embodiment of the present invention.

[0049] Figure 2 An exploded diagram of an ultrasonic probe provided in an embodiment of the present invention.

[0050] Figure 3 A block diagram of a radiotherapy device provided in an embodiment of the present invention.

[0051] Figure 4 A schematic diagram of the structure of the SA-UNet neural network model provided in an embodiment of the present invention.

[0052] Figure 5 A schematic diagram of the structure of a spatial attention module provided in an embodiment of the present invention.

[0053] Figure 6 A flowchart of a working method of a radiotherapy device provided in an embodiment of the present invention.

[0054] Figure 7 A flowchart of the SA-UNet neural network model training provided in an embodiment of the present invention.

[0055] Wherein: 1 ultrasonic probe, 11 piezoelectric element array layer, 12 acoustic backing layer, 13 acoustic matching layer, 14 epoxy resin sealing layer, 2 coupling agent layer, 3 human skin, 4 host computer, 5 accelerator. DETAILED DESCRIPTION

[0056] The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0057] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] Using ordinal numbers “first,” “second,” “third,” etc. to describe common objects merely indicates that different instances of similar objects are involved and is not intended to imply that the objects so described must have a given order in time, space, order, or in any other manner.

[0059] In addition, the expression of “comprising” an element is an “open” expression, which merely means that corresponding components or steps exist, and should not be interpreted as excluding additional components or steps.

[0060] In order to achieve the purpose of the present invention, some embodiments of an ultrasonic device, a radiotherapy device including the ultrasonic device, and a working method thereof are as follows: Figure 1 and 2 As shown, the ultrasonic device includes: an ultrasonic probe 1 and a coupling agent layer 2.

[0061] The ultrasonic probe 1 includes: an internal central element and an epoxy resin sealing layer 14 coated on the outside of the internal center, the internal central element includes: a piezoelectric element array layer 11, an acoustic backing layer 12 and an acoustic matching layer 13 arranged on opposite sides of the piezoelectric element array layer, the piezoelectric element array layer 11 has a built-in integrated circuit for collecting ultrasonic images, the acoustic backing layer 12 is used to eliminate resonance effects, and the acoustic matching layer 13 is used to increase the acoustic propagation effect to the skin.

[0062] The coupling agent layer 2 is a hydrogel elastomer and is used to attach the ultrasound probe to human skin 3 .

[0063] Furthermore, the ultrasonic probe 1 is a thin and hard patch probe, which can realize high-density elements (eg, 500 elements per square centimeter).

[0064] The center frequency of the piezoelectric element array layer is 3, 7 or 10 MHz. The piezoelectric element array layer of the present invention has a built-in integrated circuit, which has faster and better signal transmission performance than the traditional internal center element with top circuit and bottom circuit, and can better collect ultrasound images.

[0065] The present invention discloses an ultrasonic device for tumor radiotherapy that can continuously obtain the dynamics of organs in a patient's body in real time. The ultrasonic probe 1 includes: an internal central element with a three-layer structure and an epoxy resin sealing layer 14 coated on the outer side of the internal central element. The internal central element with the three-layer structure has a simple structure and does not affect the detection performance of the piezoelectric element array layer, which can reduce costs. In addition, when the probe is attached to the patient's skin surface during radiotherapy, it is more lightweight. On the one hand, it can reduce the pressure of the ultrasonic probe on the patient, and on the other hand, it makes the skin deformation caused by the pressure less serious, thereby improving the accuracy of treatment to a certain extent.

[0066] In addition, the coupling agent layer 2 is a hydrogel elastomer, which is soft, tough, resistant to dehydration and has good bioadhesion, and can adhere firmly to the skin. During radiotherapy, the slight movement of the body can still maintain a stable position, and the coupling agent layer can ensure the high reliability of the probe in the present invention. The coupling agent effectively transmits sound waves, protects the probe from skin deformation, and maintains robust and long-lasting (48 hours) comfortable ultrasound real-time imaging.

[0067] In order to further optimize the implementation effect of the present invention, in some other embodiments, the remaining characteristic technologies are the same, except that the hydrogel elastomer includes: a hydrogel and an elastomer film wrapped around the outside of the hydrogel.

[0068] Furthermore, the hydrogel comprises: a chitosan-polyacrylamide interpenetrating polymer network and water.

[0069] In a specific embodiment, the hydrogel is composed of chitosan-polyacrylamide interpenetrating polymer network (10%) and water (90%). The elastic film is made of polyurethane material, and the elastic film can prevent the hydrogel from dehydrating.

[0070] Further, the thickness of the elastomeric film is less than 40 mm.

[0071] On the other hand, Figure 3 As shown, an embodiment of the present invention further discloses a radiotherapy device, comprising:

[0072] The ultrasonic device disclosed in any of the above embodiments is attached to human skin and is used to collect ultrasonic images;

[0073] A host computer 4, the host computer 4 is electrically connected to the ultrasonic probe 1 of the ultrasonic device through a data line 6, and the host computer 4 is used to process the ultrasonic image collected by the ultrasonic device and extract key information;

[0074] The accelerator 5 is electrically connected to the host computer 4 and is used for performing radiotherapy.

[0075] The present invention discloses a radiotherapy device, wherein the ultrasonic device can be linked with a radiotherapy accelerator. The ultrasonic probe of the present invention is adhered to human skin through a coupling agent layer, and is connected to a host computer through a data line, so as to obtain real-time continuous ultrasonic images of organs. The host computer processes the collected ultrasonic images, extracts key information (key information may be, but is not limited to, tumor location information, predicted respiratory curve, etc.), and feeds it back to the accelerator to control the movement of the grating, thereby achieving accurate radiotherapy of the tumor.

[0076] Furthermore, the host computer has a built-in SA-UNet neural network model, such as Figure 4 As shown in FIG, the SA-UNet neural network model includes an encoder, a spatial attention module, and a decoder connected in sequence. The encoder is used to obtain the image features of the ultrasound image, and the spatial attention module is used to further obtain the fine-grained features of the ultrasound image, such as Figure 5 As shown, the decoder is used to reconstruct features and output the results.

[0077] The present invention adopts the SA-UNet neural network model, which consists of two parts, encoding and decoding, and combines the original information such as texture and spatial position with the image semantic information. The overall network is lightweight and convenient.

[0078] Since blood vessel segmentation is a delicate task, and its structure is different from the integrity of other organs, a spatial attention module is added between the encoder and decoder. The model assigns weights to different spatial positions through learning, so that the model can be biased towards learning the characteristics of blood vessels, thereby improving the segmentation effect.

[0079] In addition, on the other hand, the embodiment of the present invention also discloses a working method of a radiotherapy device, which is used to implement the operation of the radiotherapy device disclosed in any of the above embodiments, such as Figure 6 As shown, the specific steps include:

[0080] Step 1: The ultrasound device collects ultrasound images;

[0081] Step 2: The ultrasound image collected by the ultrasound device is sent to the host computer, which processes it and extracts key information;

[0082] Step 3: The accelerator controls the grating movement and performs radiotherapy based on the key information extracted by the host computer.

[0083] Furthermore, the host computer has a built-in SA-UNet neural network model, such as Figure 7 As shown, the training of the SA-UNet neural network model includes the following steps:

[0084] S1: Perform ultrasound imaging of organs in different parts and annotate the collected ultrasound images manually;

[0085] S2: The original ultrasound images and the annotated ultrasound images are matched one by one, and divided into training set, test set and validation set in a ratio of 6:3:1. The training set is used to optimize the neural network model, the validation set is used to intuitively evaluate the performance of the current model during training, and the test set is used to test the generalization performance of the neural network model after training.

[0086] S3: Set parameters such as the number of training rounds, initial learning rate, batch size, etc., input the training set into the SA-UNet neural network model for forward propagation, and obtain the label result;

[0087] S4: Compare the label result obtained in S3 with the true label to obtain the classification loss function and Dice evaluation coefficient;

[0088] S5: Based on the classification loss function and Dice evaluation coefficient obtained in S4, the SA-UNet neural network model is back-propagated to optimize the parameters of the SA-UNet neural network model;

[0089] S6: Repeat S3-S5 until the training results of the SA-UNet neural network model meet expectations.

[0090] In some specific embodiments:

[0091] The classification loss function is

[0092] Where: p(y i ) is the category y i The probability of Indicates that the prediction is class y i probability;

[0093] The Dice evaluation coefficient is used to measure the effect of the model on image segmentation. The Dice evaluation coefficient is

[0094] Among them, X is the ultrasound image with the real label, and Y is the ultrasound image of the prediction result of the SA-UNet neural network model.

[0095] Furthermore, in S3, after data enhancement processing is performed on the training set, it is input into the SA-UNet neural network model for forward propagation.

[0096] Data enhancement processing includes but is not limited to: processing the training set data by random rotation, cropping, etc., so that the data sent to the neural network model in each round is different, thereby increasing the generalization ability of the model.

[0097] The trained neural network model is deployed on the host computer to extract key information (such as blood vessels) from the collected ultrasound images, and the tumor location is tracked and predicted based on the extracted results.

[0098] Furthermore, the SA-UNet neural network model includes an encoder, a spatial attention module, and a decoder connected in sequence. The operation of the SA-UNet neural network model includes the following steps:

[0099] A1: The encoder obtains the image features of the ultrasound image; A2: The image features obtained by the encoder are transmitted to the spatial attention module and input into the global maximum pooling and global average pooling layers respectively;

[0100] A3: Perform feature concatenation on the output feature maps obtained from the two branches;

[0101] A4: The concatenated feature maps are subjected to 7×7 convolution, batch normalization layer and Sigmoid activation function to obtain the intermediate feature map;

[0102] A5: Multiply the feature map obtained by the encoder and the intermediate feature map pixel by pixel to obtain the final spatial attention map;

[0103] A6: Pass the spatial attention map as input feature to the decoder;

[0104] A7: The decoder reconstructs features and outputs the results.

[0105] In summary, the encoder obtains the image features of the ultrasound image and then transmits the obtained image features to the spatial attention module. The spatial attention module further obtains the fine-grained features in the ultrasound image, focuses on more important regional information, and suppresses irrelevant information such as the background. Finally, the spatial attention module outputs the image features and transmits them to the decoder. The decoder reconstructs the features and restores the image to the original image size and resolution, and outputs the final result.

[0106] The spatial attention module assigns different weights to different spatial positions of the collected ultrasound images. By using the gradient descent algorithm and back-propagation mechanism, the SA-UNet neural network automatically updates the model weights in continuous iterative training, thereby learning image features and focusing on important areas in the ultrasound image.

[0107] The present invention provides an ultrasonic device, a radiotherapy device including the ultrasonic device and a working method thereof, wherein the ultrasonic device includes: an ultrasonic probe and a coupling agent layer, and has the following beneficial effects:

[0108] First, the ultrasound probe includes: an internal center element and an epoxy resin sealing layer wrapped around the outside of the internal center. The internal center element includes: a piezoelectric element array layer, an acoustic backing layer and an acoustic matching layer arranged on opposite sides of the piezoelectric element array layer, and the outermost layer is an epoxy resin sealing layer, which has high stability and reliability in long-term applications. At the same time, the three-layer internal center element reduces the cost and weight of the ultrasound probe, so that when the probe covers the human skin, on the one hand, it reduces the pressure caused by the ultrasound probe on the patient and reduces the discomfort, and on the other hand, it makes the skin deformation caused by pressure less severe, which improves the accuracy of radiotherapy to a certain extent.

[0109] Second, the coupling agent layer is a hydrogel elastomer, which is soft, tough, resistant to dehydration and has good bioadhesion. It can adhere firmly to human skin, and can maintain a stable position with slight body movements during radiotherapy, and the reliability of the ultrasound probe is high. The coupling agent layer can effectively transmit sound waves, protect the ultrasound probe from skin deformation, and maintain robust and lasting comfortable ultrasound real-time imaging.

[0110] Third, the hydrogel elastomer includes: a hydrogel and an elastic film wrapped around the outside of the hydrogel, and the elastic film can effectively prevent the hydrogel from dehydrating.

[0111] Fourth, the ultrasound probe is connected to the host computer to obtain real-time and continuous ultrasound images. The host computer processes the ultrasound images, extracts the location information of the tumor, predicts the breathing curve, etc., and feeds back to the accelerator to control the movement of the grating and achieve precise radiotherapy of the tumor.

[0112] Fifth, the present invention adopts the SA-UNet neural network model to combine the original information such as texture and spatial position with the image semantic information, and the overall network is lightweight and convenient. Moreover, a spatial attention module is added between the encoder and the decoder to improve the segmentation effect.

[0113] In the description of the present invention, it is necessary to understand that the terms "coaxial", "bottom", "one end", "top", "middle", "the other end", "upper", "one side", "top", "inside", "front", "center", "both ends" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0114] In the present invention, unless otherwise clearly stipulated and limited, the terms such as "installation", "setting", "connection", "fixation" and "screw-on" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral one; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be the internal connection of two elements or the interaction relationship between two elements. Unless otherwise clearly defined, ordinary technicians in this field can understand the specific meanings of the above terms in the present invention according to the specific circumstances.

[0115] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for illustrating the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which shall fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

[0116] The control method of the present invention is to control by manually starting and closing the switch. The wiring diagram of the power element and the provision of power supply are common knowledge in the field, and the present invention is mainly used to protect mechanical devices, so the present invention will no longer explain the control method and wiring layout in detail.

Claims

1. Radiotherapy equipment, characterized in that include: An ultrasonic device, the ultrasonic device is attached to human skin, the ultrasonic device is used to collect ultrasonic images, the ultrasonic device comprises: an ultrasonic probe and a coupling agent layer, the ultrasonic probe comprises: an internal center element and an epoxy resin sealing layer coated on the outside of the internal center, the internal center element comprises: a piezoelectric element array layer, an acoustic backing layer and an acoustic matching layer arranged on opposite sides of the piezoelectric element array layer; the piezoelectric element array layer has a built-in integrated circuit for collecting ultrasonic images, the acoustic backing layer is used to eliminate resonance effects, and the acoustic matching layer is used to increase acoustic propagation effects to the skin; the coupling agent layer is a hydrogel elastomer, and the coupling agent layer is used to attach the ultrasonic probe to human skin; A host computer, the host computer is electrically connected to the ultrasonic probe of the ultrasonic device, and the host computer is used to process the ultrasonic image collected by the ultrasonic device to extract key information; An accelerator, the accelerator is electrically connected to the host computer, and the accelerator is used for radiotherapy; The host computer has a built-in SA-UNet neural network model, which includes an encoder, a spatial attention module and a decoder connected in sequence, the encoder is used to obtain image features of ultrasound images, the spatial attention module is used to further obtain fine-grained features of ultrasound images, and the decoder is used to perform feature reconstruction and output results; The training of the SA-UNet neural network model includes the following steps: S1: Annotate the collected ultrasound images; S2: The original ultrasound images and the annotated ultrasound images are matched one by one, and divided into training set, test set and validation set according to a certain ratio; S3: Setting training parameters of the SA-UNet neural network model, inputting the training set into the SA-UNet neural network model for forward propagation, and obtaining label results; S4: Compare the label result obtained in S3 with the true label to obtain the classification loss function and Dice evaluation coefficient; S5: Based on the classification loss function and Dice evaluation coefficient obtained in S4, back-propagation is performed on the SA-UNet neural network model to optimize the parameters of the SA-UNet neural network model; S6: Repeat S3-S5 until the training result of the SA-UNet neural network model reaches the expected result.

2. The radiotherapy device according to claim 1, characterized in that The radiotherapy device is implemented using the following method, which specifically includes the following steps: Step 1: The ultrasound device collects ultrasound images; Step 2: The ultrasound image collected by the ultrasound device is sent to the host computer, which processes it and extracts key information; Step 3: The accelerator controls the grating movement and performs radiotherapy based on the key information extracted by the host computer.

3. The radiotherapy device according to claim 1, characterized in that In S3, after data enhancement processing is performed on the training set, it is input into the SA-UNet neural network model for forward propagation.

4. The radiotherapy device according to claim 3, characterized in that The SA-UNet neural network model includes an encoder, a spatial attention module, and a decoder connected in sequence. The operation of the SA-UNet neural network model includes the following steps: A1: The encoder obtains the image features of the ultrasound image; A2: The image features obtained by the encoder are transferred to the spatial attention module and input into the global maximum pooling and global average pooling layers respectively; A3: Perform feature concatenation on the output feature maps obtained from the two branches; A4: The concatenated feature maps are subjected to 7×7 convolution, batch normalization layer and Sigmoid activation function to obtain the intermediate feature map; A5: Multiply the feature map obtained by the encoder and the intermediate feature map pixel by pixel to obtain the final spatial attention map; A6: Pass the spatial attention map as input feature to the decoder; A7: The decoder reconstructs features and outputs the results.

5. The radiotherapy device according to claim 1, characterized in that The hydrogel elastomer comprises: a hydrogel and an elastic film wrapped around the outside of the hydrogel.

6. The radiotherapy device according to claim 5, characterized in that The hydrogel comprises: a chitosan-polyacrylamide interpenetrating polymer network and water.

7. The radiotherapy device according to claim 5, characterized in that The thickness of the elastomeric film is less than 40 mm.

Images