Focused ultrasound treatment device based on real-time visual guidance and control method thereof

Abstract

This invention relates to a focused ultrasound therapy device and its control method based on real-time visualization guidance. The system includes a treatment head device comprising a housing with an oval hole on its wall. A concave emission source is installed inside the housing, and an imaging linear array is integrated inside the concave emission source. The concave emission source is connected to the working end of a first actuator and to the working end of a second actuator. The main unit includes a main control module, a pulse generator, and an image reconstruction module. By configuring the first actuator, the second actuator, the imaging linear array, and the image reconstruction module, focus detection and offset correction can be automatically completed based on a deep learning model, thereby improving treatment accuracy and safety, reducing human error, and lowering the risk of tissue damage. It is particularly suitable for precise and safe focused ultrasound therapy on superficial targets such as the facial skin and fascia layer.

Description

Technical Field

[0001] This invention relates to the field of ultrasound therapy equipment technology, and in particular to a focused ultrasound therapy device and its control method based on real-time visualization guidance. Background Technology

[0002] High-Intensity Focused Ultrasound (HIFU) technology can focus ultrasound energy at a specific depth under the skin to produce a thermal coagulation effect, thereby lifting and tightening the fascia layer.

[0003] In existing technologies, focused ultrasound (FUS) devices used for skin lifting and tightening generally employ concave treatment sources with fixed curvature or interchangeable treatment heads at different depths as the transmitter. Operators estimate the subcutaneous target area based on experience and surface markers to implement treatment. However, because traditional FUS devices generally lack a visual channel from the body surface to the target area, the entire treatment process is essentially "blind." Operators cannot observe in real time whether the treatment focus is accurately acting on the target fascia layer. Furthermore, operators cannot visually avoid non-target high-risk structures such as bone contours and large blood vessels, posing potential safety hazards of burns or nerve damage. In addition, traditional FUS devices lack an active control mechanism that dynamically adjusts the focus position based on real-time tissue imaging. If the patient moves slightly or there are individual differences in local tissue characteristics, the preset focus depth will deviate from the actual target area, leading to unstable treatment results or even complete misses. Summary of the Invention

[0004] Therefore, it is necessary to address the problems of treatment target deviation, uneven energy deposition, and safety hazards caused by the lack of effective target area visualization and focus adaptive control methods in the operation of existing focused ultrasound equipment. This paper proposes a focused ultrasound treatment device and its control method based on real-time visualization guidance.

[0005] The technical solution adopted in this invention is as follows: A focused ultrasound therapy device based on real-time visualization guidance includes: A treatment head device includes a housing with an oval hole on its wall. A concave emission source is installed inside the housing. Ultrasonic pulses generated by the concave emission source are emitted through the oval hole. An imaging linear array is integrated inside the concave emission source. The concave emission source is connected to the working end of a first actuator, which drives the concave emission source to move linearly along the long axis of the oval hole. The concave emission source is also connected to the working end of a second actuator, which drives the concave emission source to move linearly along the acoustic axis. The host includes a main control module, a pulse generator, and an image reconstruction module. The main control module generates a global synchronization clock, the pulse generator drives the concave emission source to generate ultrasonic pulses, and the image reconstruction module generates tomographic images of superficial tissues in real time based on the image information transmitted back from the imaging linear array. The image reconstruction module is equipped with a deep learning model, which synchronously performs tissue hierarchical segmentation on the tomographic image to generate a real-time hierarchical segmentation image. The deep learning model performs focused feature bright spot detection on the real-time hierarchical segmentation image and sends corresponding control signals to the first and second actuators based on the detection results to adjust the spatial position of the concave emission source.

[0006] As a further improvement to the above technical solution: The concave emission source is fixed on the mounting base, which is fitted onto the outer circular surface of the guide rod. The mounting base and the guide rod are slidably fitted together. The mounting base is connected to the working end of the first actuator. The first actuator drives the mounting base to slide along the axial direction of the guide rod, thereby causing the concave emission source to move linearly. The guide rod is slidably installed inside the housing, and a connecting seat is fitted on the outer circular surface of the guide rod. The connecting seat is connected to the working end of the second actuator. The second actuator drives the connecting seat to move along the height direction of the housing, thereby driving the mounting seat to move along the height direction of the housing through the guide rod, and thus driving the concave surface emitter to make linear motion.

[0007] The structure of the first actuator is as follows: it includes a first motor, the output end of the first motor is connected to a first driving gear, the first driving gear is meshed with a driven gear fixed to the outer circle of the lead screw, the lead screw and the sleeve are rotatably installed, and the end of the sleeve is fixedly connected to the mounting base; The first motor drives the first drive gear to rotate, which in turn drives the lead screw to rotate through the driven gear, and then drives the mounting base to move linearly along the axial direction of the guide rod through the sleeve rod.

[0008] The structure of the second actuator is as follows: it includes a second motor, the output end of the second motor is connected to a second drive gear, and the second drive gear is meshed with a rack fixed on the connecting seat; The second motor drives the second drive gear to rotate, and under the action of the rack, it drives the connecting seat to move linearly along the height direction of the outer shell.

[0009] The concave emission source adopts a boat-shaped structure or a bowl-shaped structure.

[0010] A control method for the above-mentioned focused ultrasound therapy device based on real-time visualization guidance includes the following steps: S1. Scanning and positioning: The ultrasound treatment system is started and operates in imaging mode. The operator places the treatment head device against the surface of the skin to be treated and moves it to scan through the imaging linear array. The imaging linear array transmits the image signal of the skin to be treated to the image reconstruction module. S2. Target area confirmation: The image reconstruction module generates tomographic images of superficial tissues in real time. At the same time, the deep learning model segments the fascia layer on the tomographic image in real time and outlines it with colored contour lines to obtain a real-time hierarchical segmentation image. The real-time hierarchical segmentation image is presented by the host computer, and the operator selects the treatment target area based on the real-time hierarchical segmentation image. S3. Autofocus: The operator aligns the concave emission source with the treatment target area and then manually triggers the focus command. The ultrasound treatment system automatically executes the focus process to perform individualized autofocus, so that the measured focal position of the real-time ultrasound pulse emitted by the concave emission source coincides with the center of the fascia layer contour on the real-time hierarchical segmentation image delineated by the deep learning model. S4. Start treatment. The operator manually triggers the treatment command. Subsequently, the ultrasound treatment system works in treatment mode. The pulse generator drives the concave emission source to emit treatment ultrasound pulses. The deep learning model captures the instantaneous bright spot generated by the treatment ultrasound pulse on the real-time hierarchical segmentation image and calculates the centroid coordinates of the instantaneous bright spot. The centroid coordinates are used as the measured focal position. S5. Closed-loop monitoring: In treatment mode, the deep learning model detects in real time whether the measured focal position of the therapeutic ultrasound pulse emitted by the concave emission source coincides with the center of the fascia layer contour on the real-time hierarchical segmentation image. If the deep learning model detects an offset between the measured focal position and the center of the fascia layer contour on the real-time hierarchical segmentation image, it sends corresponding control signals to the first and second actuators through the image reconstruction module, thereby adjusting the spatial position of the concave emission source so that the measured focal position moves toward the corresponding center of the fascia layer contour, ensuring that the focal point of the therapeutic ultrasound pulse is always within the treatment target area.

[0011] As a further improvement to the above technical solution: In S3, the focusing process includes the following steps: The pulse generator drives the concave emission source to emit a set of test ultrasonic pulses; The deep learning model captures the instantaneous bright spot generated by the test ultrasonic pulse on the real-time hierarchical segmentation image, calculates the centroid coordinates of the instantaneous bright spot, and uses the centroid coordinates as the measured focal position. If there is an offset between the measured focal position and the center of the fascia layer contour on the real-time hierarchical segmentation image, the deep learning model calculates the offset vector between the two and sends corresponding control signals to the first and second actuators through the image reconstruction module, thereby adjusting the spatial position of the concave emission source so that the measured focal position moves toward the center of the corresponding fascia layer contour until the offset vectors of the two converge within the preset alignment confirmation threshold range.

[0012] The imaging linear array is equipped with isolation circuitry and a preamplifier.

[0013] During the emission of ultrasonic pulses by the concave emission source, the input of the preamplifier is switched to a protected high-impedance state through the isolation circuit, thereby keeping the imaging linear array in an isolated and protected state. During the imaging linear array scan, the concave emission source is silenced, the isolation circuit is restored, and the imaging linear array performs the scan in cooperation with the preamplifier.

[0014] In treatment mode, the ultrasound therapy system is equipped with a safety protection mechanism that continuously monitors abnormal signals outside the treatment target area through a deep learning model; If the measured focal position deviates from the treatment target area and exceeds the effective compensation range of the ultrasound treatment system for multiple consecutive cycles, or if the measured focal position falls into a preset high-risk position, the concave emission source will automatically stop emitting treatment ultrasound pulses, and the main unit will issue an audible and visual alarm to remind the operator to manually adjust the position of the treatment head device, or to remind the operator to manually trigger the focusing command and re-execute the focusing process.

[0015] The beneficial effects of this invention are as follows: This invention features a compact and reasonable structure and is easy to operate. By setting up a first actuator, a second actuator, an imaging linear array, and an image reconstruction module, it can automatically complete focus detection and offset correction based on a deep learning model, thereby improving treatment accuracy and safety, reducing human error, and lowering the risk of tissue damage. It is especially suitable for precise and safe focused ultrasound treatment of superficial targets such as the facial skin fascia layer.

[0016] The present invention also has the following advantages: (1) The control method of the present invention establishes an image-physical space mapping by alternating timing of “treatment-imaging” and individualized calibration of low-energy test pulses to achieve automatic focusing; during treatment, the real-time deviation between the measured focus and the target area is monitored in a closed loop and automatically corrected, and with the safety protection mechanism, the focus is always accurately locked on the fascia layer target area, effectively preventing accidental injury and realizing intelligent visual adaptive focusing treatment.

[0017] (2) By setting a focusing process before treatment, the ultrasound treatment system can establish an accurate spatial mapping based on the real tissue acoustic characteristics of the current patient, thus completely solving the problem of focus calibration deviation caused by differences in individual tissues.

[0018] (2) In this invention, by setting up an imaging linear array, an image reconstruction module, a first actuator and a second actuator, the focal position of the ultrasonic pulse is adjustable, thereby ensuring that the measured focal position of the ultrasonic pulse always follows the center of the treatment target area, which significantly improves the accuracy, individualized adaptability and consistency of the treatment.

[0019] (3) By setting up a closed-loop monitoring process, the present invention can realize continuous visual monitoring and offset detection during the treatment process, so that the ultrasound treatment system can automatically adjust and relock when the measured focal position deviates from the treatment target area, effectively overcoming the influence of dynamic factors such as patient movement.

[0020] (4) By setting up a safety protection mechanism, the risk of adverse reactions such as burns can be significantly reduced, and the dependence on the operator's experience and skills can be effectively reduced. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the treatment head device in this invention.

[0022] Figure 2 for Figure 1 The main view.

[0023] Figure 3 for Figure 2 A sectional view of section AA in the middle.

[0024] Figure 4 This is a schematic diagram of the installation structure of the first actuator, the second actuator, and the concave emission source in this invention.

[0025] Figure 5 for Figure 4 The main view.

[0026] Figure 6 This is a schematic diagram of the concave emission source in the present invention. Figure 1 .

[0027] Figure 7 This is a schematic diagram of the concave emission source in the present invention. Figure 2 .

[0028] Figure 8 This is a schematic diagram of the concave emission source in the present invention. Figure 3 .

[0029] Figure 9 This is a timing diagram for the present invention.

[0030] Figure 10(a) shows the results before ultrasound treatment using the present invention. Figure 1 .

[0031] Figure 10(b) shows the results of ultrasound therapy performed using the present invention. Figure 1 .

[0032] Figure 11(a) shows the results before ultrasound treatment using the present invention. Figure 2 .

[0033] Figure 11(b) shows the results of ultrasound therapy performed using the present invention. Figure 2 .

[0034] Figure 12(a) shows the results before ultrasound treatment using the present invention. Figure 3 .

[0035] Figure 12(b) shows the results of ultrasound therapy performed using the present invention. Figure 3 .

[0036] Figure 13(a) shows the results before ultrasound treatment using the present invention. Figure 4 .

[0037] Figure 13(b) shows the results of ultrasound therapy performed using the present invention. Figure 4 .

[0038] The components include: 1. Concave emission source; 2. Housing; 3. Oval hole; 4. Guide groove; 5. First actuator; 6. Second actuator; 7. Mounting base; 8. Connecting base; 9. Guide rod; 501. First motor; 502. First driving gear; 503. Driven gear; 504. Lead screw; 505. Sleeve rod; 601. Second motor; 602. Second drive gear; 603. Rack; 604. Adapter. Detailed Implementation

[0039] The specific embodiments of the present invention will now be described with reference to the accompanying drawings.

[0040] like Figures 1-8As shown, the focused ultrasound therapy device based on real-time visualization guidance in this embodiment includes: a treatment head device, which includes a housing 2. An oval hole 3 is formed on the wall of the housing 2. A concave emission source 1 is installed inside the housing 2. Ultrasonic pulses generated by the concave emission source 1 are emitted through the oval hole 3. An imaging linear array is integrated inside the concave emission source 1. The concave emission source 1 is connected to the working end of a first actuator 5. The first actuator 5 drives the concave emission source 1 to move linearly along the long axis of the oval hole 3. The concave emission source 1 is also connected to the working end of a second actuator 6. The second actuator 6 drives the concave emission source 1 to move linearly along the acoustic axis. The treatment head device of this embodiment includes a concave emission source 1, a housing 2, a first actuator 5, and a second actuator 6. Concave emission source 1 is an ultrasonic transducer, which is based on a spherical concave piezoelectric material component, such as... Figure 4 , Figures 6-7 As shown, the concave emission source 1 adopts a boat-shaped structure, or, as... Figure 8 As shown, the concave emission source 1 adopts a bowl-shaped structure, which can achieve focusing of ultrasonic pulses; the concave emission source 1 is equipped with an imaging array composed of a linear array of charge-coupled devices (CCDs), which can achieve rapid and accurate scanning of the skin to be treated; in this embodiment, the front end face of the individual CCD constituting the imaging array is adapted to the curvature trend of the spherical concave piezoelectric material of the concave emission source 1, and the outer wall surface of the imaging array and the inner wall surface of the concave emission source 1 are filled with a sound insulation and decoupling material layer, which can block mechanical vibration and acoustic crosstalk; The concave emission source 1 is fixed on the mounting base 7, which is fitted onto the outer circular surface of the guide rod 9. The mounting base 7 and the guide rod 9 are slidably fitted together. The mounting base 7 is connected to the working end of the first actuator 5. The first actuator 5 drives the mounting base 7 to slide along the axial direction of the guide rod 9, thereby causing the concave emission source 1 to move linearly. The guide rod 9 is slidably installed inside the housing 2. A connecting seat 8 is fitted onto the outer circular surface of the guide rod 9. The connecting seat 8 is connected to the working end of the second actuator 6. The second actuator 6 drives the connecting seat 8 to move along the height direction of the housing 2, thereby causing the mounting base 7 to move along the height direction of the housing 2 through the guide rod 9, and thus causing the concave emission source 1 to move linearly.

[0041] In this embodiment, the mounting base 7, the connecting base 8, and the guide rod 9 are all arranged inside the outer shell 2. The outer shell 2 has guide grooves 4 symmetrically arranged on its inner wall surface. The guide grooves 4 are installed in conjunction with the guide rod 9 to guide the movement of the guide rod 9. In addition, an oval hole 3 is provided on the wall surface of the outer shell 2 to allow the ultrasonic pulse emitted by the concave surface emission source 1 to be emitted. The structure of the first actuator 5 is as follows: it includes a first motor 501, the output end of the first motor 501 is connected to a first driving gear 502, the first driving gear 502 is meshed with a driven gear 503 fixed to the outer surface of the lead screw 504, the lead screw 504 is rotatably mounted with a sleeve 505, and the end of the sleeve 505 is fixedly connected to the mounting base 7; the first motor 501 drives the first driving gear 502 to rotate, thereby driving the lead screw 504 to rotate through the driven gear 503, and then driving the mounting base 7 to move linearly along the axial direction of the guide rod 9 through the sleeve 505; by setting the first actuator 5, the concave surface emission source 1 can be driven to move linearly along the long axis of the oblong hole 3, thereby adjusting the focal position of the ultrasonic pulse; The structure of the second actuator 6 is as follows: it includes a second motor 601, the output end of which is connected to a second drive gear 602, and the second drive gear 602 is meshed with a rack 603 fixed on the connecting seat 8; the second motor 601 drives the second drive gear 602 to rotate, and under the action of the rack 603, it drives the connecting seat 8 to move linearly along the height direction of the outer shell 2; in this embodiment, the output end of the second motor 601 is connected to a first adapter shaft, and the first adapter shaft is connected to the second adapter shaft through an adapter 604, and the end of the second adapter shaft is fixed with the second drive gear 602; by setting the second actuator 6, the concave surface emission source 1 can be driven to move linearly along the acoustic axis direction, thereby adjusting the focal position of the ultrasonic pulse; in this embodiment, the acoustic axis direction refers to the acoustic axis direction of the imaging array, which is perpendicular to the radial end face of the oval hole 3; The host unit includes a main control module, a pulse generator, and an image reconstruction module. The main control module generates a global synchronization clock, the pulse generator drives the concave emission source 1 to generate ultrasonic pulses, and the image reconstruction module generates real-time tomographic images of superficial tissues based on image information transmitted from the imaging linear array. The image reconstruction module contains a deep learning model that synchronously performs tissue layer segmentation on the tomographic images to generate real-time layered segmented images. The deep learning model also performs focused feature bright spot detection on the real-time layered segmented images and sends corresponding control signals to the first actuator 5 and the second actuator 6 based on the detection results, thereby adjusting the spatial position of the concave emission source 1. The main control module is used to generate a global synchronization clock, such as Figure 9 As shown, this is used to strictly define the alternating rhythm of the treatment phase, the test pulse phase, and the imaging phase; The pulse generator is used to drive the concave emission source 1 to generate ultrasonic pulses; The image reconstruction module consists of a large-scale FPGA and a GPU. The FPGA is responsible for beamforming the multi-channel parallel echo data of the imaging linear array to generate high-frame-rate tomographic images of shallow tissues in real time. The GPU is equipped with a trained deep learning model that performs tissue layer segmentation and focus feature bright spot detection on the tomographic images simultaneously, and outputs control signals to drive the first actuator 5 and the second actuator 6 to adjust the spatial position of the concave emission source 1. In addition, the main unit is equipped with a display screen, a focus button, and a treatment button. The display screen is used to show real-time hierarchical segmentation images and focused feature bright spots. The focus button and treatment button are used to switch the working mode of the ultrasound therapy system.

[0042] The ultrasound therapy system of this embodiment, by setting up an imaging linear array, an image reconstruction module, a first actuator 5 and a second actuator 6, can integrate real-time imaging, deep learning model analysis and spatial position control into a closed loop, enabling the focus of the ultrasound pulse to be precisely locked on the fascia layer target area, realizing a leap from "blind shooting" to "intelligent visual adaptive focusing", significantly improving the accuracy and safety of treatment.

[0043] This embodiment also provides a control method for a focused ultrasound therapy device based on real-time visualization guidance, including the following steps: S1. Scanning and positioning, the ultrasound treatment system is started and operates in imaging mode. The operator places the treatment head device against the surface of the skin to be treated and moves it to scan through the imaging array. The imaging array transmits the image signal of the skin to be treated to the image reconstruction module. In the ultrasound therapy system of this embodiment, the imaging linear array is equipped with an isolation circuit and a preamplifier, both of which are installed inside the main unit. During the imaging linear array scan, the concave emission source 1 is silenced, the isolation circuit is restored, and the imaging linear array performs the scan in cooperation with the preamplifier.

[0044] S2. Target area confirmation: The image reconstruction module generates tomographic images of superficial tissues in real time. At the same time, the deep learning model segments the fascia layer on the tomographic image in real time and outlines it with colored contour lines to obtain a real-time hierarchical segmentation image. The real-time hierarchical segmentation image is presented by the host computer, and the operator selects the treatment target area based on the real-time hierarchical segmentation image. By setting up an imaging linear array and an image reconstruction module, an alternating "treatment-imaging" time sequence is established to acquire real-time images of the treatment target area. The fascia layer is then segmented and focusing features are identified through a deep learning model, thereby enabling visualization of the operation.

[0045] S3. Autofocus: The operator aligns the concave emission source 1 with the treatment target area. Then, the operator manually presses the focus button on the main unit to manually trigger the focus command. The ultrasound treatment system automatically executes the focus process to perform individualized autofocus, so that the measured focus position of the real-time ultrasound pulse emitted by the concave emission source 1 coincides with the center of the fascia layer contour on the real-time hierarchical segmentation image delineated by the deep learning model. Specifically, the focusing process includes the following steps: The pulse generator drives the concave emission source 1 to emit a set of test ultrasonic pulses. In this embodiment, the test ultrasonic pulses are low-energy pulses, and their energy levels are strictly controlled within a safe range that will not cause irreversible tissue damage. Meanwhile, during the transmission of test ultrasonic pulses by the concave emission source 1, the input of the preamplifier is switched to a protected high-impedance state through the isolation circuit, thereby keeping the imaging linear array in an isolated and protected state. The deep learning model captures the instantaneous bright spot generated by the test ultrasonic pulse on the real-time hierarchical segmentation image, calculates the centroid coordinates of the instantaneous bright spot, and uses the centroid coordinates as the measured focal position. If the measured focal position is offset from the center of the fascia layer contour on the real-time hierarchical segmentation image, the deep learning model calculates the offset vector between the two and sends corresponding control signals to the first actuator 5 and the second actuator 6 through the image reconstruction module, thereby adjusting the spatial position of the concave emission source 1 so that the measured focal position moves toward the center of the corresponding fascia layer contour until the offset vectors of the two converge within the preset alignment confirmation threshold range. By setting the focusing process, the focal position of the ultrasonic pulse emitted by the concave emission source 1 can be calibrated, thereby establishing the image-physical space mapping relationship of the current individual. This facilitates the deep learning model to automatically adjust the ultrasonic pulse emitted by the concave emission source 1 during the treatment process, so as to achieve autonomous focusing and ensure that the focus is locked on the fascia layer of the superficial tissue and the regional center of the fascia layer, i.e. the treatment target area, thereby achieving adaptive focusing.

[0046] S4. Start treatment. The operator manually triggers the treatment command. Subsequently, the ultrasound treatment system works in treatment mode. The pulse generator drives the concave emission source 1 to emit treatment ultrasound pulses. The deep learning model captures the instantaneous bright spot generated by the treatment ultrasound pulse on the real-time hierarchical segmentation image and calculates the centroid coordinates of the instantaneous bright spot. The centroid coordinates are used as the measured focal position. In this embodiment, during the emission of therapeutic ultrasound pulses from the concave emission source 1, the input terminal of the preamplifier is switched to a protected high-impedance state by the isolation circuit, thereby keeping the imaging linear array in an isolated and protected state. During the treatment process, such as Figure 9As shown, the ultrasound therapy system operates in a cyclical alternation of treatment ultrasound pulse emission, isolation circuit protection, imaging linear array scanning and image reconstruction module imaging, and deep learning model analysis. In each imaging frame, the deep learning model continuously compares the positional relationship between the measured focal point and the treatment target area to perform closed-loop monitoring. In addition, during the treatment mode, the ultrasound treatment system is equipped with a safety protection mechanism that continuously monitors abnormal signals outside the treatment target area through a deep learning model; If the measured focal position is detected to deviate from the treatment target area and exceed the effective compensation range of the ultrasound treatment system for multiple consecutive cycles, or if the measured focal position falls in a preset high-risk position (such as the periosteum), the concave emission source 1 will automatically stop emitting treatment ultrasound pulses, and the host will issue an audible and visual alarm to remind the operator to manually adjust the position of the treatment head device, or to remind the operator to manually trigger the focusing command and re-execute the focusing process.

[0047] S5. Closed-loop monitoring: In treatment mode, the deep learning model detects in real time whether the measured focal position of the treatment ultrasound pulse emitted by the concave emission source 1 coincides with the center of the fascia layer contour on the real-time hierarchical segmentation image. If the deep learning model detects an offset between the measured focal position and the center of the fascia layer contour on the real-time hierarchical segmentation image, it sends corresponding control signals to the first actuator 5 and the second actuator 6 through the image reconstruction module, thereby adjusting the spatial position of the concave emission source 1 so that the measured focal position moves toward the corresponding center of the fascia layer contour, ensuring that the focal point of the therapeutic ultrasound pulse is always within the therapeutic target area.

[0048] Based on the above control method, after treating multiple skin samples with the ultrasound treatment system of this embodiment, the treatment results are shown in Figures 10(a)-13(b). After treatment, each layer of superficial tissue (epidermis, dermis, SMAS fascia layer) formed a thermal coagulation lesion with clear boundaries and uniform size. The coagulation area accurately corresponds to the target area without deviation or excessive damage. The deep periosteum remained intact and unaffected, confirming that the system achieved a safe and efficient treatment effect of precise layered focusing, controllable thermal coagulation of the target area, and effective protection of non-target tissues.

[0049] The above description is an explanation of the present invention and not a limitation thereof. The scope of the present invention is defined by the claims. Within the scope of protection of the present invention, any form of modification may be made.

Claims

1. A focused ultrasound therapy device based on real-time visualization guidance, characterized in that: include: The treatment head device includes a housing (2), with an oval hole (3) on the wall of the housing (2). A concave emission source (1) is installed inside the housing (2). The ultrasonic pulse generated by the concave emission source (1) is emitted through the oval hole (3). An imaging linear array is integrated inside the concave emission source (1). The concave emission source (1) is connected to the working end of the first actuator (5). The first actuator (5) drives the concave emission source (1) to move linearly along the long axis of the oval hole (3). The concave emission source (1) is connected to the working end of the second actuator (6). The second actuator (6) drives the concave emission source (1) to move linearly along the acoustic axis. The host includes a main control module, a pulse generator and an image reconstruction module. The main control module is used to generate a global synchronization clock. The pulse generator is used to drive the concave emission source (1) to generate ultrasonic pulses. The image reconstruction module is used to generate tomographic images of superficial tissues in real time based on the image information transmitted back by the imaging linear array. The image reconstruction module is equipped with a deep learning model. The deep learning model performs tissue hierarchical segmentation on the tomographic image in a synchronous manner to generate a real-time hierarchical segmentation image. The deep learning model performs focused feature bright spot detection on the real-time hierarchical segmentation image. Based on the detection results, the corresponding control signals are sent to the first actuator (5) and the second actuator (6) respectively to adjust the spatial position of the concave emission source (1).

2. The focused ultrasound therapy device based on real-time visualization guidance as described in claim 1, characterized in that: The concave emission source (1) is fixed on the mounting base (7). The mounting base (7) is fitted on the outer circular surface of the guide rod (9). The mounting base (7) and the guide rod (9) are slidably fitted. The mounting base (7) is connected to the working end of the first actuator (5). The first actuator (5) drives the mounting base (7) to slide along the axial direction of the guide rod (9), thereby driving the concave emission source (1) to make linear motion. The guide rod (9) is slidably installed inside the outer shell (2). A connecting seat (8) is fitted on the outer circular surface of the guide rod (9). The connecting seat (8) is connected to the working end of the second actuator (6). The second actuator (6) drives the connecting seat (8) to move along the height direction of the outer shell (2), thereby driving the mounting seat (7) to move along the height direction of the outer shell (2) through the guide rod (9), and thus driving the concave surface emission source (1) to make linear motion.

3. The focused ultrasound therapy device based on real-time visualization guidance as described in claim 2, characterized in that: The structure of the first actuator (5) is as follows: it includes a first motor (501), the output end of the first motor (501) is connected to a first driving gear (502), the first driving gear (502) meshes with a driven gear (503) fixed on the outer surface of the lead screw (504), the lead screw (504) is rotatably mounted with a sleeve (505), and the end of the sleeve (505) is fixedly connected to the mounting base (7); The first motor (501) drives the first drive gear (502) to rotate, thereby driving the lead screw (504) to rotate through the driven gear (503), and then driving the mounting base (7) to move linearly along the axial direction of the guide rod (9) through the sleeve rod (505).

4. The focused ultrasound therapy device based on real-time visualization guidance as described in claim 2, characterized in that: The structure of the second actuator (6) is as follows: it includes a second motor (601), the output end of the second motor (601) is connected to a second drive gear (602), and the second drive gear (602) is meshed with a rack (603) fixed on the connecting seat (8); The second motor (601) drives the second drive gear (602) to rotate, and under the action of the rack (603), it drives the connecting seat (8) to move linearly along the height direction of the outer shell (2).

5. The focused ultrasound therapy device based on real-time visualization guidance as described in claim 1, characterized in that: The concave emission source (1) adopts a boat-shaped structure or a bowl-shaped structure.

6. A control method for a focused ultrasound therapy device based on real-time visualization guidance as described in claim 1, characterized in that: Includes the following steps: S1. Scanning and positioning: The ultrasound treatment system is started and operates in imaging mode. The operator places the treatment head device against the surface of the skin to be treated and moves it to scan through the imaging linear array. The imaging linear array transmits the image signal of the skin to be treated to the image reconstruction module. S2. Target area confirmation: The image reconstruction module generates tomographic images of superficial tissues in real time. At the same time, the deep learning model segments the fascia layer on the tomographic image in real time and outlines it with colored contour lines to obtain a real-time hierarchical segmentation image. The real-time hierarchical segmentation image is presented by the host computer, and the operator selects the treatment target area based on the real-time hierarchical segmentation image. S3. Automatic focusing: The operator aligns the concave emission source (1) with the treatment target area, and then manually triggers the focusing command. The ultrasound treatment system automatically executes the focusing process to perform individualized automatic focusing, so that the measured focus position of the real-time ultrasonic pulse emitted by the concave emission source (1) coincides with the center of the fascia layer contour on the real-time hierarchical segmentation image outlined by the deep learning model. S4. Start treatment. The operator manually triggers the treatment command. Then the ultrasound treatment system works in treatment mode. The pulse generator drives the concave emission source (1) to emit treatment ultrasound pulses. The deep learning model captures the instantaneous bright spot generated by the treatment ultrasound pulse on the real-time hierarchical segmentation image and calculates the centroid coordinates of the instantaneous bright spot. The centroid coordinates are used as the measured focal position. S5. Closed-loop monitoring: In the treatment mode, the deep learning model detects in real time whether the measured focal position of the treatment ultrasound pulse emitted by the concave emission source (1) coincides with the center of the fascia layer contour on the real-time hierarchical segmentation image; If the deep learning model detects an offset between the measured focal position and the center of the fascia layer contour on the real-time hierarchical segmentation image, it sends corresponding control signals to the first actuator (5) and the second actuator (6) respectively through the image reconstruction module, thereby adjusting the spatial position of the concave emission source (1) so that the measured focal position moves toward the center of the corresponding fascia layer contour, so that the focal point of the therapeutic ultrasound pulse is always located within the therapeutic target area.

7. The control method for a focused ultrasound therapy device based on real-time visualization guidance as described in claim 6, characterized in that: In S3, the focusing process includes the following steps: The pulse generator drives the concave emission source (1) to emit a set of test ultrasonic pulses; The deep learning model captures the instantaneous bright spot generated by the test ultrasonic pulse on the real-time hierarchical segmentation image, calculates the centroid coordinates of the instantaneous bright spot, and uses the centroid coordinates as the measured focal position. If the measured focal position is offset from the center of the fascia layer contour on the real-time hierarchical segmentation image, the deep learning model calculates the offset vector between the two and sends corresponding control signals to the first actuator (5) and the second actuator (6) respectively through the image reconstruction module, thereby adjusting the spatial position of the concave emission source (1) so that the measured focal position moves toward the center of the corresponding fascia layer contour until the offset vectors of the two converge within the preset alignment confirmation threshold range.

8. The control method for a focused ultrasound therapy device based on real-time visualization guidance as described in claim 6, characterized in that: The imaging linear array is equipped with isolation circuitry and a preamplifier.

9. The control method for a focused ultrasound therapy device based on real-time visualization guidance as described in claim 8, characterized in that: During the emission of ultrasonic pulses from the concave emission source (1), the input terminal of the preamplifier is switched to a protected high-impedance state through the isolation circuit, thereby putting the imaging linear array in an isolated and protected state. During the scanning of the imaging linear array, the concave emission source (1) is silent, the isolation circuit is restored, and the imaging linear array performs the scanning with the cooperation of the preamplifier.

10. The control method for a focused ultrasound therapy device based on real-time visualization guidance as described in claim 6, characterized in that: In treatment mode, the ultrasound therapy system is equipped with a safety protection mechanism that continuously monitors abnormal signals outside the treatment target area through a deep learning model; If the measured focal position is detected to deviate from the treatment target area and exceed the effective compensation range of the ultrasound treatment system for multiple consecutive cycles, or if the measured focal position falls in the preset high-risk position, the concave emission source (1) will automatically stop emitting treatment ultrasound pulses, and the host will issue an audible and visual alarm to remind the operator to manually adjust the position of the treatment head device, or remind the operator to manually trigger the focusing command and re-execute the focusing process.

Images