Method, apparatus and electronic device for constructing biological tissue identification model

Abstract

The present invention provides a method, apparatus and electronic device for constructing a biological tissue identification model, which relates to the field of medical technology. The present invention uses a miniature UHF ultrasonic probe to obtain an image of a biological tissue to be identified, and then inputs the image of the biological tissue to be identified into a biological tissue identification model, thereby achieving tissue identification under a spinal endoscope and improving the efficiency of tissue identification under a spinal endoscope. [Solution] The method includes the steps of: decomposing training tissue sample information to obtain features of the training tissue sample image; constructing a mapping relationship between the features of the training tissue sample image and the training sample tissue type; training a tissue identification model using a plurality of the mapping relationships to obtain initial parameters of the tissue identification model; and inputting supporting tissue sample information into the tissue identification model in which the initial parameters have been placed, to fine-tune the parameters to obtain an optimized tissue identification model.

Description

[Technical field]

[0001] The present invention relates to the field of medical technology, and in particular to a method, an apparatus and an electronic device for constructing a biological tissue identification model. [Background technology]

[0002] Spine-related diseases, which have always been one of the major risk factors threatening the health of the public, mainly include lumbar disc herniation and lumbar spinal canal stenosis.

[0003] Minimally invasive spine surgery is less risky than traditional open surgery, but it depends somewhat on the experience of the chief surgeon, has a narrow field of view, and is limited by the lack of "hand touch" under a microscope. Currently, ultrasound probes (e.g., liver) used in existing surgeries are too large to pass through minimally invasive access to the spine. There are examples of small ultrasound probes being used in gastrointestinal endoscopes and intravascular catheters, but these small ultrasound probes perform lateral or circular scanning, and can only identify specific tissue structures such as the gastrointestinal wall and blood vessel wall, and are used only to determine the nature and extent of the lesion because of their narrow scanning range and difficulty in morphological reconstruction.

[0004] Therefore, we propose a method, an apparatus and an electronic device for constructing a biological tissue identification model. Summary of the Invention [Problem to be solved by the invention]

[0005] The present invention aims to provide a method, apparatus and electronic device for constructing a biological tissue identification model to achieve tissue identification under spinal endoscopy and to improve the efficiency of tissue identification under spinal endoscopy. [Means for solving the problem]

[0006] The method for constructing a biological tissue identification model according to the present invention comprises the steps of: obtaining training tissue sample information including a training tissue sample image and a training sample tissue type, and supporting tissue sample information; decomposing the training tissue sample information to obtain features of the training tissue sample images; building a mapping relationship between features of the training tissue sample images and the training sample tissue types; training a tissue identification model using a plurality of the mapping relationships to obtain initial parameters of the tissue identification model; and inputting the supporting tissue sample information into the tissue-identification model in which the initial parameters have been populated to fine-tune the parameters to obtain an optimized tissue-identification model.

[0007] Preferably, the training sample tissue types include bone, muscle, fat and blood vessels.

[0008] Preferably, the supporting tissue sample information includes a supporting tissue sample image and a supporting sample tissue type, the supporting sample tissue type including a nerve, a nucleus pulposus, an annulus fibrosus, and a spinal cord.

[0009] Preferably, said step of decomposing said training tissue sample information to obtain features of said training tissue sample images comprises: performing wavelet decomposition on the training tissue sample information using a one-dimensional Haar wavelet transform to obtain signal features of the training tissue sample image;

[0010] Preferably, the step of inputting the supporting tissue sample information into the tissue identification model in which the initial parameters have been set, and fine-tuning the parameters to obtain the optimized tissue identification model, comprises: decomposing the supporting tissue sample information to obtain features of the supporting tissue sample image; establishing a mapping relationship between features of the supporting tissue sample image and the supporting sample tissue type; optimizing a tissue identification model using a plurality of the mapping relationships to obtain optimized parameters of the tissue identification model; determining an optimized tissue identification model based on the optimization parameters.

[0011] Preferably, after the step of determining the optimized tissue identification model based on the optimization parameters, the construction method according to the present invention further comprises: obtaining an image of the biological tissue to be identified using a miniature UHF ultrasound probe; and inputting an image of the tissue to be identified into the optimized tissue identification model to obtain a tissue type.

[0012] The apparatus for constructing a biological tissue identification model according to the present invention comprises: a collection module used to obtain training tissue sample information including a training tissue sample image and a training sample tissue type, and supporting tissue sample information; a decomposition module used for decomposing the training tissue sample information to obtain features of the training tissue sample image; a construction module used for constructing a mapping relationship between the features of the training tissue sample images and the training sample tissue types; a training module adapted to train a tissue identification model using a plurality of the mapping relationships to obtain initial parameters of the tissue identification model; and a fine-tuning module that is used to input the supporting tissue sample information into the tissue identification model in which the initial parameters have been configured, and fine-tune the parameters to obtain an optimized tissue identification model.

[0013] Preferably, the training sample tissue types include bone, muscle, fat and blood vessels.

[0014] Preferably, the supporting tissue sample information includes a supporting tissue sample image and a supporting sample tissue type, the supporting sample tissue type including a nerve, a nucleus pulposus, an annulus fibrosus, and a spinal cord.

[0015] Preferably, the decomposition module comprises: The function includes performing wavelet decomposition on the training tissue sample information using one-dimensional Haar wavelet transform to obtain signal features of the training tissue sample image.

[0016] Preferably, the fine-tuning module comprises: decomposing the supporting tissue sample information to obtain features of the supporting tissue sample image; establishing a mapping relationship between the features of the supporting biological tissue sample image and the supporting sample tissue type; a function of optimizing a tissue identification model using a plurality of the mapping relationships to obtain an optimization parameter of the tissue identification model; and determining the optimized tissue identification model based on the optimization parameters.

[0017] Preferably, after the function of determining the optimized tissue identification model based on the optimization parameters, the construction device according to the present invention further comprises: A function to obtain images of biological tissue to be identified using an ultra-small UHF ultrasonic probe; A function of inputting an image of the biological tissue to be identified into the optimized biological tissue identification model to obtain a biological tissue type is added.

[0018] The present invention further provides an electronic device.

[0019] The electronic device includes a processor and and a memory for storing executable instructions in a computer, the executable instructions, when executed, causing the processor to perform any one of the aspects of the method.

[0020] The present invention further provides a computer-readable storage medium storing one or more programs, the computer-readable storage medium performing any one of the aspects of the above-described methods when a processor executes the one or more programs. Effect of the Invention

[0021] The present invention relates to training a tissue identification model by few-shot learning, and after completing the training of the tissue identification model using a large amount of data that is easy to collect, the tissue identification model is optimized using a small amount of data that is difficult to collect.Then, based on the optimized tissue identification model, the tissue is scanned under the spinal endoscope in the conventional endoscopic surgical path using an ultra-small UHF ultrasonic probe, and tissue identification is achieved under the spinal endoscope, thereby improving the tissue identification efficiency under the spinal endoscope. [Brief description of the drawings]

[0022] In order to more clearly explain the technical solutions in the embodiments of the present invention, we will now briefly describe the drawings necessary for the description of the embodiments. It is obvious that the drawings in the following description are only a part of the embodiments of the present invention, and other related drawings can also be obtained based on these drawings under the premise that no creative labor is required for those skilled in the art.

[0023] [Figure 1] FIG. 1 is a principle schematic diagram of the tissue identification model construction method provided in the embodiment of the present invention.

[0024] [Diagram 2] FIG. 2 is a principle schematic diagram of the tissue identification model construction device provided in the embodiment of the present invention.

[0025] [Diagram 3] FIG. 3 is a schematic diagram of an electronic device provided in an embodiment of the present invention.

[0026] [Figure 4] FIG. 4 is a principle schematic diagram of a computer-readable medium provided in an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0027] The following description discloses the present invention to enable those skilled in the art to realize the invention. The preferred embodiments in the following description are merely illustrative, and those skilled in the art may devise other obvious variations. The basic principles of the present invention defined in the following description may be applied to other embodiments, improved solutions, modified solutions, equivalent solutions and other technical solutions without departing from the spirit and scope of the present invention.

[0028] We will more fully describe exemplary embodiments of the present invention in combination with Figures 1 to 4. However, the exemplary embodiments may be implemented in various forms and should not be construed as being limited to the embodiments described herein. On the contrary, these exemplary embodiments are provided to make the present invention more complete and complete, and to more conveniently convey the concept of the present invention to those skilled in the art. In the drawings, the same drawing labels represent the same or similar elements, parts or portions, and therefore repeated description thereof will be omitted.

[0029] As long as it is in accordance with the technical concept of the present invention, the features, structures, characteristics or other details described in a particular embodiment may be combined in any suitable manner with one or more other embodiments.

[0030] In the description of the specific embodiments, the present invention will be described in detail with respect to features, structures, characteristics, or other details for a person skilled in the art to fully understand. However, the person skilled in the art may implement the technical solution of the present invention without the specific features, structures, characteristics, or other details.

[0031] The flowcharts shown in the drawings are merely examples and do not necessarily include all contents and operations / steps, or need to be performed in the order described. For example, some operations / steps can be separated, and some operations / steps can be merged or partially merged, so that the actual execution order may change according to actual circumstances.

[0032] The block diagrams shown in the drawings are merely functional entities which do not necessarily correspond to physically separate entities, i.e. which functional entities are realized in the form of software, one or more hardware modules or integrated circuits, different network and / or processor and / or microcontroller devices.

[0033] The term "and / or" includes any and all combinations of one or more of the associated listed items.

[0034] 1 is a principle schematic diagram of the tissue identification model construction method provided in the embodiment of the present invention, which includes the following steps:

[0035] S110: Obtain training tissue sample information, including a training tissue sample image and a training sample tissue type, and supporting tissue sample information.

[0036] Preferably, the training sample tissue types include bone, muscle, fat and blood vessels.

[0037] Preferably, the supporting tissue sample information includes a supporting tissue sample image and a supporting sample tissue type, the supporting sample tissue type including a nerve, a nucleus pulposus, an annulus fibrosus, and a spinal cord.

[0038] In a specific embodiment of the present invention, few-shot learning (FSL) is a type of machine learning method, and all existing machine learning and deep learning tasks rely on large amounts of annotation data for training. However, this is not the case for human learning. Humans can learn well on new problems with just a few examples by incorporating knowledge learned in the past. Few-shot learning corresponds to such a process, and is expected to incorporate prior knowledge and require only a few samples for new problems, just like humans. Existing application fields for few-shot learning focus on images, and have been widely applied to problems such as image classification and reinforcement learning, but there are still many problems in determining biological tissue types, and mature models are lacking.

[0039] A large number of cross-sectional images of bones, muscles, fat, blood vessels, etc. are collected and their types are labeled, and a small number of cross-sectional images of nerves, annulus fibrosus, spinal cord, etc. are collected and their types are labeled. Since bones, muscles, fat, blood vessels, etc. are relatively on the surface and therefore can be easily obtained, and since the shapes and sizes of bones, muscles, fat, blood vessels, etc. are easy to obtain, the large amount of data on bones, muscles, fat, blood vessels, etc. serves as training biological tissue sample information. Since data on nerves, nucleus pulposus, annulus fibrosus, spinal cord, etc. are located in the human spinal cord channel, are difficult to obtain, are difficult to distinguish, and the amount of data available is small, they serve as supporting biological tissue sample information.

[0040] S120: Decompose the training tissue sample information to obtain features of the training tissue sample images.

[0041] Preferably, said step of decomposing said training tissue sample information to obtain features of said training tissue sample images comprises: performing wavelet decomposition on the training tissue sample information using a one-dimensional Haar wavelet transform to obtain signal features of the training tissue sample image;

[0042] In a specific embodiment of the present invention, a one-dimensional Haar wavelet transform is used to perform wavelet decomposition on the ultrasound scan signals of the training tissue samples to obtain signal features of each training tissue sample. The images of the training tissue samples include a range of tissues at a scanning depth.

[0043] S130: Establish a mapping relationship between the features of the training tissue sample images and the tissue types of the training samples.

[0044] S140: Training a tissue identification model using the plurality of mapping relationships to obtain initial parameters of the tissue identification model.

[0045] S150: Input the supporting tissue sample information into the tissue identification model in which the initial parameters have been set, and fine-tune the parameters to obtain an optimized tissue identification model.

[0046] Preferably, the step of inputting the supporting tissue sample information into the tissue identification model in which the initial parameters have been set, and fine-tuning the parameters to obtain the optimized tissue identification model, comprises: decomposing the supporting tissue sample information to obtain features of the supporting tissue sample image; establishing a mapping relationship between features of the supporting tissue sample image and the supporting sample tissue type; optimizing a tissue identification model using a plurality of the mapping relationships to obtain optimized parameters of the tissue identification model; determining an optimized tissue identification model based on the optimization parameters.

[0047] In a specific embodiment of the present invention, a one-dimensional Haar wavelet transform is used to perform wavelet decomposition on the ultrasound scan signals of the supporting tissue samples to obtain signal characteristics of each supporting tissue sample. The images of the supporting tissue samples include the range of tissues at the scan depth. For example, the signal characteristics of the nucleus pulposus and vertebrae include homogeneity and low acoustic impedance, the signal characteristics of the nerve include a punctate surface and low acoustic impedance, and the signal characteristics of the annulus fibrosus include a high acoustic impedance.

[0048] Preferably, after the step of determining the optimized tissue identification model based on the optimization parameters, the construction method according to the present invention further comprises: obtaining an image of the biological tissue to be identified using a miniature UHF ultrasound probe; and inputting an image of the tissue to be identified into the optimized tissue identification model to obtain a tissue type.

[0049] In a specific embodiment of the present invention, the diameter range of the ultra-miniature UHF ultrasound probe is within 6 mm, preferably 4 mm; the orientation angle of the ultra-miniature UHF ultrasound probe faces forward and deflects laterally from 0 to 60 degrees, which is convenient for adapting to the use of spinal endoscopes, realizing ultrasonic scanning of soft tissues along minimally invasive access to minimize unnecessary damage to the human body during surgery.

[0050] The ultra-miniature UHF ultrasonic probe is used to scan tissue under the spinal endoscope in the surgical path of a conventional endoscope. The ultra-miniature UHF ultrasonic probe connected to an imaging device through a conductor is used to excite an electric pulse signal and collect the electric pulse signal converted from the ultrasonic echo, and the transmission and reception of phased array ultrasonic waves is realized by changing ultrasonic parameters, and then the contents detected by the ultrasonic waves are imaged through the imaging device.

[0051] The present invention relates to training a tissue identification model by few-shot learning, and after completing the training of the tissue identification model using a large amount of data that is easy to collect, the tissue identification model is optimized using a small amount of data that is difficult to collect.Then, based on the optimized tissue identification model, the tissue is scanned under the spinal endoscope in the conventional endoscopic surgical path using an ultra-small UHF ultrasonic probe, and tissue identification is achieved under the spinal endoscope, thereby improving the tissue identification efficiency under the spinal endoscope.

[0052] FIG. 2 is a schematic diagram of the principle of the tissue identification model construction device provided in the embodiment of the present invention. an acquisition module 10 used to obtain training tissue sample information including training tissue sample images and training sample tissue types, and supporting tissue sample information; a decomposition module 20 used to decompose the training tissue sample information to obtain features of the training tissue sample image; a construction module 30 used for constructing a mapping relationship between the features of the training tissue sample images and the training sample tissue types; a training module 40 adapted to train a tissue identification model using a plurality of said mapping relationships to obtain initial parameters of said tissue identification model; and a fine-tuning module 50 which is used to input the supporting tissue sample information into the tissue identification model in which the initial parameters have been set, to fine-tune the parameters to obtain an optimized tissue identification model.

[0053] Preferably, the training sample tissue types include bone, muscle, fat and blood vessels.

[0054] Preferably, the supporting tissue sample information includes a supporting tissue sample image and a supporting sample tissue type, the supporting sample tissue type including a nerve, a nucleus pulposus, an annulus fibrosus, and a spinal cord.

[0055] Preferably, the decomposition module comprises: The function includes performing wavelet decomposition on the training tissue sample information using one-dimensional Haar wavelet transform to obtain signal features of the training tissue sample image.

[0056] Preferably, the fine-tuning module comprises: decomposing the supporting tissue sample information to obtain features of the supporting tissue sample image; establishing a mapping relationship between the features of the supporting biological tissue sample image and the supporting sample tissue type; a function of optimizing a tissue identification model using a plurality of the mapping relationships to obtain an optimization parameter of the tissue identification model; and determining the optimized tissue identification model based on the optimization parameters.

[0057] Preferably, after the function of determining the optimized tissue identification model based on the optimization parameters, the construction device according to the present invention further comprises: A function to obtain images of biological tissue to be identified using an ultra-small UHF ultrasonic probe; A function of inputting an image of the biological tissue to be identified into the optimized biological tissue identification model to obtain a biological tissue type is added.

[0058] Since the functions of the apparatus of the embodiment of the present invention are described in the above method embodiment, in the description of this embodiment, please refer to the relevant descriptions in the above embodiment for non-detailed parts, and they will not be repeated here.

[0059] Based on a similar inventive concept, an embodiment of the present invention further provides an electronic device.

[0060] We next describe an embodiment of an electronic device of the present invention, which is regarded as a concrete embodiment for the above-mentioned method and apparatus embodiments of the present invention. Details described in the embodiment of the electronic device of the present invention are regarded as supplementary to the above-mentioned method or apparatus embodiments, and details not disclosed in the embodiment of the electronic device of the present invention can be realized with reference to the above-mentioned method or apparatus embodiments.

[0061] Fig. 3 is a schematic diagram of an electronic device provided in an embodiment of the present invention. We will now describe an electronic device 300 according to an embodiment of the present invention with reference to Fig. 3. The electronic device 300 shown in Fig. 3 is merely an example and should not bring any limitation to the function and application scope of the embodiment of the present invention.

[0062] 3, the electronic device 300 is represented in the form of a general-purpose computing device. The components of the electronic device 300 include, but are not limited to, at least one processing unit 310, at least one memory unit 320, a bus 330 connecting various system components (including the memory unit 320 and the processing unit 310), a display unit 340, etc.

[0063] Here, the storage unit is for storing program code, which, when executed by the processing unit 310, causes the processing unit 310 to perform the steps of the methods described herein above and various exemplary embodiments of the present invention. For example, the processing unit 310 can perform the steps illustrated in FIG.

[0064] The storage unit 320 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 3201 and / or a cache storage unit 3202 , and may further include a read only memory unit (ROM) 3203 .

[0065] The storage unit 320 may further comprise a program / utilities 3204 having a set of (at least one) program modules 3205. Such program modules 3205 may comprise, but are not limited to, an operating system, one or more applications, other program modules and program data, each or any combination of which in this example comprises a network environment.

[0066] Bus 330 may represent one or more of several bus structures, including a storage unit bus, or a storage unit controller, a bus for peripherals, an accelerated graphics port, a processing unit, or a local bus employing any of a variety of bus structures.

[0067] The electronic device 300 can communicate with one or more external devices 400 (e.g., a keyboard, a pointing device, a Bluetooth device, etc.), one or more devices that allow a user to interact with the electronic device 300, and / or any device (e.g., a router, a modem, etc.) that allows the electronic device 300 to communicate with one or more other computing devices. This communication can occur via an input / output (I / O) interface 350. Additionally, the electronic device 300 can communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, e.g., the Internet, etc.) via a network adapter 360. The network adapter 360 can communicate with other modules of the electronic device 300 via a bus 330. Although not shown in FIG. 3, it should be understood that the electronic device 300 can be used in combination with other hardware and / or software modules, including, but not limited to, microcode, device drives, redundant processing units, external disk drive arrays, RAID systems, tape drives, data backup storage systems, etc.

[0068] The description of the above embodiments makes it easy for those skilled in the art to understand that the exemplary embodiments described in the present invention can be realized through software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of the present invention can be embodied in the form of a software product. The software product is stored in a computer-readable storage medium (such as a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or a network, and includes some instructions that enable a computing device (such as a personal computer, a server, a network device, etc.) to execute the above method according to the present invention. When the computer program is executed by a data processing device, the computer-readable medium can realize the above method of the present invention shown in FIG. 1.

[0069] FIG. 4 is a principle schematic diagram of a computer-readable medium provided in an embodiment of the present invention.

[0070] A computer program implementing the method shown in FIG. 1 can be stored on one or more computer readable media. The computer readable medium can be either a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or apparatus, or any combination thereof. More specific examples of readable storage media (non-exhaustive) include an electrical connection with one or more conductors, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0071] The computer readable storage medium comprises a propagated data signal, either in baseband or as part of a carrier wave, that carries readable program code. The propagated data signal may take many forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The readable storage medium may take any readable medium other than a readable storage medium. The readable medium may transmit, propagate, or transmit a program for use by or in connection with an instruction execution system, device, or apparatus device. The program code contained in the readable storage medium may be transmitted via any suitable medium, including but not limited to wireless, wired, cable, RF, etc., or any suitable combination thereof.

[0072] A programming language or any combination of one or more programming languages ​​encodes program code for carrying out the operations of the present invention. Programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as traditional procedural programming languages, e.g., "C" or similar programming languages. The program code may be executed entirely on the user's computing device, partially on the user's device, as a separate software package, partially on both the user's device and a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or connected to an external computing device (e.g., connected via the Internet via an Internet service provider).

[0073] In summary, the present invention may be realized in the form of hardware, software modules running on one or more processors, or a combination thereof. Those skilled in the art should understand that in practice, some or all of the functions of some or all of the components in the embodiments of the present invention may be realized using a general-purpose data processing device such as a microprocessor or a digital signal processor (DSP). The present invention may be realized as a device or apparatus program (e.g., computer program and computer program product) for performing some or all of the methods described herein. Such programs for the implementation of the present invention may be stored on a computer-readable medium or may comprise the form of one or more signals. Such signals may be downloaded from an Internet website, provided from a carrier signal, or provided in other forms.

[0074] In the above specific embodiment, the objectives, technical solutions and beneficial effects of the present invention are further described in detail, and it should be understood that the present invention is not inherently related to any specific computer, virtual device or electronic device, and various general-purpose devices can also realize the present invention. The above content is only a specific embodiment of the present invention, and is not intended to limit the present invention, and any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0075] Each embodiment in this specification is described step by step, and the same and similar parts between each embodiment may be referred to each other. Each embodiment will be described by focusing on the differences from other embodiments.

[0076] The above contents are merely examples of the present application and are not intended to limit the present application. Those skilled in the art may make various modifications and variations to the present application. Modifications, equivalent replacements, improvements, etc. made within the scope of the spirit and principles of the present application shall be included in the scope of the claims of the present application.

[0077] (Additional Note) (Appendix 1) obtaining training tissue sample information including a training tissue sample image and a training sample tissue type, and supporting tissue sample information; decomposing the training tissue sample information to obtain features of the training tissue sample images; building a mapping relationship between features of the training tissue sample images and the training sample tissue types; training a tissue identification model using a plurality of the mapping relationships to obtain initial parameters of the tissue identification model; inputting the supporting tissue sample information into the tissue identification model in which the initial parameters have been set, and fine-tuning the parameters to obtain an optimized tissue identification model; A method for constructing a biological tissue identification model.

[0078] (Appendix 2) the training sample tissue types include bone, muscle, fat, and blood vessels; 2. A method for constructing a biological tissue identification model according to claim 1.

[0079] (Appendix 3) The supporting biological tissue sample information includes a supporting biological tissue sample image and a supporting sample tissue type, and the supporting sample tissue type includes a nerve, a nucleus pulposus, an annulus fibrosus, and a spinal cord. 3. A method for constructing a biological tissue identification model according to claim 2.

[0080] (Appendix 4) said step of decomposing said training tissue sample information to obtain features of said training tissue sample images, performing wavelet decomposition on the training tissue sample information using a one-dimensional Haar wavelet transform to obtain signal features of the training tissue sample image; 4. A method for constructing a biological tissue identification model according to claim 3.

[0081] (Appendix 5) inputting the supporting tissue sample information into the tissue identification model in which the initial parameters have been set, and fine-tuning the parameters to obtain an optimized tissue identification model; decomposing the supporting tissue sample information to obtain features of the supporting tissue sample image; establishing a mapping relationship between features of the supporting tissue sample image and the supporting sample tissue type; optimizing a tissue identification model using a plurality of the mapping relationships to obtain optimized parameters of the tissue identification model; determining an optimized tissue discrimination model based on the optimization parameters; 5. A method for constructing a biological tissue identification model according to claim 4.

[0082] (Appendix 6) After determining the optimized tissue discrimination model based on the optimization parameters, obtaining an image of the biological tissue to be identified using a miniature UHF ultrasound probe; inputting an image of the tissue to be identified into the optimized tissue identification model to obtain a tissue type. 6. A method for constructing a biological tissue identification model according to claim 5.

[0083] (Appendix 7) a collection module used to obtain training tissue sample information including a training tissue sample image and a training sample tissue type, and supporting tissue sample information; a decomposition module used for decomposing the training tissue sample information to obtain features of the training tissue sample image; a construction module used for constructing a mapping relationship between the features of the training tissue sample images and the training sample tissue types; a training module adapted to train a tissue identification model using a plurality of the mapping relationships to obtain initial parameters of the tissue identification model; a fine-tuning module for inputting the supporting tissue sample information into the tissue identification model in which the initial parameters have been set, to fine-tune the parameters to obtain an optimized tissue identification model. A biological tissue identification model construction device comprising:

[0084] (Appendix 8) the training sample tissue types include bone, muscle, fat, and blood vessels; 8. A biological tissue identification model construction device according to claim 7.

[0085] (Appendix 9) A method for constructing a tissue identification model comprising: a processor; and a memory for storing executable instructions in a computer, the executable instructions being executed by the processor to perform the method for constructing a tissue identification model according to any one of claims 1 to 6. 1. An electronic device comprising:

[0086] (Appendix 10) One or more programs are stored, and when a processor executes the one or more programs, the method for constructing a biological tissue identification model according to any one of claims 1 to 6 is performed; A computer-readable storage medium comprising:

Claims

1. obtaining training tissue sample information including a training tissue sample image and a training sample tissue type, and supporting tissue sample information; decomposing the training tissue sample information to obtain features of the training tissue sample image; building a mapping relationship between features of the training tissue sample images and the training sample tissue types; training a tissue identification model using a plurality of the mapping relationships to obtain initial parameters of the tissue identification model; inputting the supporting tissue sample information into the tissue identification model in which the initial parameters have been set, and fine-tuning the parameters to obtain an optimized tissue identification model; A method for constructing a biological tissue identification model.

2. the training sample tissue types include bone, muscle, fat, and blood vessels; The method for constructing a biological tissue identification model according to claim 1 .

3. The supporting biological tissue sample information includes a supporting biological tissue sample image and a supporting sample tissue type, and the supporting sample tissue type includes a nerve, a nucleus pulposus, an annulus fibrosus, and a spinal cord. The method for constructing a biological tissue identification model according to claim 2 .

4. said step of decomposing said training tissue sample information to obtain features of said training tissue sample images, performing wavelet decomposition on the training tissue sample information using a one-dimensional Haar wavelet transform to obtain signal features of the training tissue sample image; The method for constructing a biological tissue identification model according to claim 3 .

5. inputting the supporting tissue sample information into the tissue identification model in which the initial parameters have been set, and fine-tuning the parameters to obtain an optimized tissue identification model; decomposing the supporting tissue sample information to obtain features of the supporting tissue sample image; establishing a mapping relationship between features of the supporting biological tissue sample image and the supporting sample tissue type; optimizing a tissue identification model using a plurality of the mapping relationships to obtain optimized parameters of the tissue identification model; determining an optimized tissue discrimination model based on the optimization parameters; 5. The method for constructing a biological tissue identification model according to claim 4.

6. After determining the optimized tissue discrimination model based on the optimization parameters, obtaining an image of the biological tissue to be identified using a miniature UHF ultrasound probe; inputting an image of the tissue to be identified into the optimized tissue identification model to obtain a tissue type. The method for constructing a biological tissue identification model according to claim 5 .

7. an acquisition module used to obtain training tissue sample information including a training tissue sample image and a training sample tissue type, and supporting tissue sample information; a decomposition module used to decompose the training tissue sample information to obtain features of the training tissue sample image; a construction module used to construct a mapping relationship between the features of the training tissue sample images and the training sample tissue types; a training module adapted to train a tissue identification model using a plurality of the mapping relationships to obtain initial parameters of the tissue identification model; a fine-tuning module for inputting the supporting tissue sample information into the tissue identification model in which the initial parameters have been set, to fine-tune the parameters to obtain an optimized tissue identification model. A biological tissue identification model construction device comprising:

8. the training sample tissue types include bone, muscle, fat, and blood vessels; 8. The apparatus for constructing a biological tissue identification model according to claim 7.

9. A method for constructing a tissue identification model according to any one of claims 1 to 6, comprising: a processor; and a memory for storing executable instructions in a computer, the executable instructions being executed by the processor to perform the method for constructing a tissue identification model according to any one of claims 1 to 6.

1. An electronic device comprising:

10. One or more programs are stored, and the method for constructing a biological tissue identification model according to any one of claims 1 to 6 is performed when a processor executes the one or more programs. A computer-readable storage medium comprising:

Images