A spinal vertebral body recognition method, device and electronic equipment

Abstract

The application provides a spine vertebra recognition method, device and electronic equipment, the method comprising: acquiring a coronal picture and a sagittal picture of the same object; extracting a coronal initial feature map from the coronal picture; extracting a sagittal initial feature map from the sagittal picture; performing feature fusion processing on the sagittal initial feature map based on the coronal initial feature map to obtain a sagittal target feature map corresponding to the sagittal picture; inputting the sagittal target feature map into a first detection module to recognize the spine vertebra in the sagittal picture. Through the spine vertebra recognition method, device and electronic equipment provided in the embodiment of the application, the relationship between the spine vertebra is constrained by using the double-view features of the same object, which will not cause serious errors such as missing detection of the vertebra. Compared with the recognition of the spine vertebra in the traditional sagittal picture, the method avoids the blurred influence caused by the shielding of the patient's arm on the spine vertebra, and the detection effect is better.

Description

Technical Field

[0001] This invention relates to the field of medical imaging technology, and more specifically, to a method, device, electronic device, and computer-readable storage medium for identifying spinal vertebrae. Background Technology

[0002] Currently, AI (Artificial Intelligence) algorithms are commonly used to identify vertebrae in sagittal X-ray images. However, because the human arm can obstruct the view of the vertebrae, some areas of the cervical and thoracic vertebrae become blurred, making vertebral body detection difficult. This ultimately leads to serious errors in the detection results, such as over-detection and under-detection of vertebrae, resulting in poor detection effectiveness. Summary of the Invention

[0003] To address the existing technical problems, embodiments of the present invention provide a method, device, electronic device, and computer-readable storage medium for identifying spinal vertebrae.

[0004] In a first aspect, embodiments of the present invention provide a method for identifying spinal vertebrae, comprising: acquiring coronal and sagittal images of the same object; extracting an initial coronal feature map from the coronal image; extracting an initial sagittal feature map from the sagittal image; performing feature fusion processing on the initial sagittal feature map based on the initial coronal feature map to obtain a sagittal target feature map corresponding to the sagittal image; and inputting the sagittal target feature map into a first detection module to identify spinal vertebrae in the sagittal image.

[0005] Optionally, the initial sagittal feature map is fused based on the initial coronal feature map to obtain the target sagittal feature map corresponding to the sagittal image. This includes: having the initial sagittal feature map learn features from the initial coronal feature map to obtain a fused feature map corresponding to the sagittal image; and superimposing the fused feature map corresponding to the sagittal image with the initial sagittal feature map to obtain the target feature map of the sagittal image.

[0006] Optionally, having the sagittal initial feature map learn features from the coronal initial feature map includes: using the encoder of the converter to have the sagittal initial feature map learn features from the coronal initial feature map.

[0007] Optionally, superimposing the fused feature map corresponding to the sagittal image with the initial sagittal feature map includes: using a residual network structure to superimpose the fused feature map corresponding to the sagittal image with the initial sagittal feature map.

[0008] Optionally, an initial coronal feature map is extracted from the coronal image; an initial sagittal feature map is extracted from the sagittal image, including: extracting the initial coronal feature map from the coronal image using the same feature extractor, and extracting the initial sagittal feature map from the sagittal image.

[0009] Optionally, after extracting the initial coronal feature map from the coronal image and the initial sagittal feature map from the sagittal image, the method further includes: superimposing the fused feature map corresponding to the coronal image with the initial coronal feature map to obtain the coronal target feature map corresponding to the coronal image; the fused feature map corresponding to the coronal image represents the feature map obtained by having the initial coronal feature map learn features from the initial sagittal feature map; and inputting the coronal target feature map into the second detection module to identify the vertebral bodies of the spine in the coronal image.

[0010] Optionally, after inputting the coronal target feature map into the second detection module to identify the vertebral bodies in the coronal image, and inputting the sagittal target feature map into the first detection module to identify the vertebral bodies in the sagittal image, the method further includes: naming the identified vertebral bodies in the coronal image sequentially from top to bottom in the vertical direction to obtain the naming results of the vertebral bodies in the coronal image; and, in conjunction with the naming results of the vertebral bodies in the coronal image, naming the identified vertebral bodies in the sagittal image.

[0011] Secondly, embodiments of the present invention also provide a spinal vertebral body recognition device, comprising: an acquisition module, a feature extraction module, a feature processing module, and a recognition module; the acquisition module is used to acquire coronal and sagittal images of the same object; the feature extraction module is used to extract an initial coronal feature map from the coronal image; and extract an initial sagittal feature map from the sagittal image; the feature processing module is used to perform feature fusion processing on the initial sagittal feature map based on the initial coronal feature map to obtain a sagittal target feature map corresponding to the sagittal image; the recognition module is used to input the sagittal target feature map into a first detection module to recognize the spinal vertebrae in the sagittal image.

[0012] Thirdly, embodiments of the present invention provide an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor. The transceiver, the memory, and the processor are connected via the bus. When the computer program is executed by the processor, it implements the steps in the spinal vertebral body recognition method described above.

[0013] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps in the spinal vertebral body identification method described in any of the above claims.

[0014] The spinal vertebral body recognition method, device, electronic device, and computer-readable storage medium provided in this invention fuse initial coronal feature maps extracted from coronal images with initial sagittal feature maps extracted from sagittal images to obtain the required sagittal target feature maps for detection, ultimately improving the accuracy of the detection results for 24 sagittal vertebrae. This method utilizes coronal images of the same object (e.g., a patient) to assist in recognizing the sagittal images of that object, i.e., using the dual-view features of the same object to constrain the relationships between spinal vertebrae, avoiding serious errors such as multiple or missed vertebrae detections. Compared to traditional spinal vertebral body recognition from sagittal images, this method avoids the blurring effect caused by the patient's arm obscuring the spinal vertebrae, resulting in better detection performance. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the background art, the accompanying drawings used in the embodiments of the present invention or the background art will be described below.

[0016] Figure 1 A flowchart of a spinal vertebral body identification method provided by an embodiment of the present invention is shown;

[0017] Figure 2 This diagram illustrates a network structure for recognizing sagittal images in the spinal vertebral body recognition method provided in an embodiment of the present invention.

[0018] Figure 3 This invention provides a flowchart illustrating the process of performing feature fusion processing on the initial sagittal feature map based on the initial coronal feature map to obtain the sagittal target feature map corresponding to the sagittal image in the spinal vertebral body recognition method.

[0019] Figure 4 This diagram illustrates the process of obtaining a fused feature map in the spinal vertebral body recognition method provided in this embodiment of the invention.

[0020] Figure 5 This diagram illustrates a network structure for recognizing both sagittal and coronal images in the spinal vertebral body recognition method provided in this embodiment of the invention.

[0021] Figure 6 A schematic diagram of the structure of a spinal vertebral body identification device provided in an embodiment of the present invention is shown;

[0022] Figure 7A schematic diagram of the structure of an electronic device for performing a spinal vertebral body recognition method is shown in an embodiment of the present invention. Detailed Implementation

[0023] The embodiments of the present invention will now be described with reference to the accompanying drawings.

[0024] Figure 1 A flowchart of a spinal vertebral body identification method provided by an embodiment of the present invention is shown. Figure 1 As shown, the method includes the following steps 101-104.

[0025] Step 101: Obtain coronal and sagittal images of the same object.

[0026] In cases where it is necessary to identify individual vertebrae of the spine in a sagittal image of an object, such as identifying individual vertebrae in a patient's sagittal X-ray, in addition to obtaining the sagittal image of the object (e.g., the same patient), a coronal image of the object is also required. It should be noted that this embodiment of the invention can identify 7 cervical vertebrae, 12 thoracic vertebrae, and 5 lumbar vertebrae in a sagittal image; all 24 vertebrae can be displayed as quadrilaterals in the sagittal image.

[0027] Step 102: Extract the initial coronal feature map from the coronal image; extract the initial sagittal feature map from the sagittal image.

[0028] In this invention, a feature extractor can be used to extract features from the acquired coronal images to obtain an initial coronal feature map; correspondingly, a feature extractor can also be used to extract features from the acquired sagittal images to obtain an initial sagittal feature map. The feature extractors used for feature extraction from the coronal and sagittal images can be two or more different feature extractors. For example, the feature extractor can be any common network, such as ResNet in CNN (Convolutional Neural Network) or the Swin transformer in transformer; optionally, step 102 can also use the same feature extractor f. θ (like Figure 2 As shown, initial coronal feature maps are extracted from coronal images, and initial sagittal feature maps are extracted from sagittal images to ensure consistency and accuracy of feature extraction and reduce extraction errors between the initial coronal and sagittal feature maps.

[0029] Step 103: Perform feature fusion processing on the initial sagittal feature map based on the initial coronal feature map to obtain the sagittal target feature map corresponding to the sagittal image.

[0030] In this embodiment of the invention, the initial sagittal feature map can be fused with the initial coronal feature map obtained in step 102. This feature fusion process is a fusion process of the feature maps, which allows the information parameters carried by the initial coronal feature map to be fused with the information parameters carried by the initial sagittal feature map. The initial sagittal feature map is further refined and adjusted using the initial coronal feature map, and finally a feature map corresponding to the sagittal image with richer details and easier recognition is obtained. The feature map corresponding to the sagittal image obtained after feature fusion is used as the target feature map of the sagittal image.

[0031] Step 104: Input the sagittal target feature map into the first detection module to identify the spinal vertebrae in the sagittal image.

[0032] After obtaining the sagittal target feature map, this feature map is input into the first detection module used to identify the vertebral bodies of the spine in the sagittal image, such as a sagittal detection module based on a transformer detector and primarily using a target detection network (e.g., Figure 2 As shown in the figure); through the identification of the first detection module, the detection results of 7 cervical vertebrae, 12 thoracic vertebrae and 5 lumbar vertebrae in the sagittal image can be obtained.

[0033] The spinal vertebral body recognition method provided in this invention performs feature fusion processing based on initial coronal feature maps extracted from coronal images and initial sagittal feature maps extracted from sagittal images. This process obtains the target sagittal feature maps required for detection, ultimately improving the accuracy of the detection results for the 24 vertebral bodies in the sagittal plane. This method utilizes coronal images of the same object (e.g., a patient) to assist in recognizing the sagittal images of that object. Specifically, it uses the dual-view features of the same object to constrain the relationships between spinal vertebral bodies, avoiding serious errors such as multiple or missed vertebral body detections. Compared to traditional spinal vertebral body recognition from sagittal images, this method avoids the blurring effect caused by the patient's arm obscuring the spinal vertebral bodies, resulting in better detection performance.

[0034] Optionally, see Figure 3 As shown, step 103 above, "based on the initial coronal feature map, the initial sagittal feature map is fused to obtain the sagittal target feature map corresponding to the sagittal image", may include the following steps 1031-1032.

[0035] Step 1031: Let the initial feature map of the sagittal plane learn features from the initial feature map of the coronal plane to obtain the fused feature map corresponding to the sagittal image.

[0036] like Figure 4 As shown, given the initial coronal and sagittal feature maps, the sagittal initial feature map can learn features from the coronal initial feature map. This learning process can be understood as: the sagittal initial feature map absorbs and interacts with features from the coronal initial feature map; in other words, this learning process involves observing and learning which information parameters the coronal initial feature map (e.g., an initial feature map from another perspective) contains while the sagittal initial feature map retains its own information parameters. The information parameters carried by the coronal initial feature map are then interacted with those carried by the sagittal initial feature map to obtain a fused feature map corresponding to the sagittal image that has learned the information parameters contained in the coronal initial feature map. Optionally, step 1031 may include: using the encoder of a transformer to teach features from the coronal initial feature map using the sagittal initial feature map. It should be noted that... Figure 4 The reshape shown indicates a function call that can reshape the number of rows, columns, or dimensions of a matrix.

[0037] Step 1032: Overlay the fused feature map corresponding to the sagittal image with the initial sagittal feature map to obtain the target feature map of the sagittal image.

[0038] like Figure 2 As shown, after obtaining the fused feature map corresponding to the sagittal image, the fused feature map corresponding to the sagittal image is superimposed with the initial sagittal feature map extracted from the sagittal image. This superposition process is understood to be a feature interaction and fusion process, aimed at further integrating the more comprehensive fused feature map with the initial sagittal feature map, thereby refining the initial sagittal feature map. Specifically, it utilizes the information parameters carried by the initial coronal feature map to assist in optimizing the initial sagittal feature map, obtaining the target sagittal feature map required for detection. Optionally, superimposing the fused feature map corresponding to the sagittal image with the initial sagittal feature map may include: using a residual network structure to superimpose the fused feature map corresponding to the sagittal image with the initial sagittal feature map; that is, embodiments of the present invention can use a residual network structure to superimpose the fused feature map corresponding to the sagittal image with the initial sagittal feature map, such as... Figure 2 As shown, "+" indicates superposition. This invention, by drawing on the processing techniques of residual network structures, solves the problems of excessively deep neural networks being difficult to train, as well as the problems of gradient explosion or vanishing gradients.

[0039] In the process of recognizing vertebrae in sagittal images, this invention innovatively adopts a method of feature extraction and feature fusion for both views, which can obtain a sagittal target feature map that is more suitable for the first detection module. This method is more accurate in recognizing vertebrae in sagittal images.

[0040] Optionally, after step 102 above, "extracting the initial coronal feature map from the coronal image; extracting the initial sagittal feature map from the sagittal image", the method may further include the following steps A1-A2.

[0041] Step A1: Overlay the fused feature map corresponding to the coronal image with the initial coronal feature map to obtain the coronal target feature map corresponding to the coronal image; the fused feature map corresponding to the coronal image represents the feature map obtained by having the initial coronal feature map learn features from the initial sagittal feature map.

[0042] This invention embodiment can also employ a similar method to obtaining the sagittal target feature map to generate the coronal target feature map. Specifically, based on the method described in step 1031 above, the initial coronal feature map learns features from the initial sagittal feature map to generate a fused feature map corresponding to the coronal image. The generated fused feature map corresponding to the coronal image is then superimposed on the initial coronal feature map to obtain a feature map corresponding to the coronal image that is richer in detail and easier to identify. This feature map is then used as the coronal target feature map. The superposition method can be found in step 1032 above and will not be repeated here.

[0043] Step A2: Input the coronal target feature map into the second detection module to identify the spinal vertebrae in the coronal image.

[0044] Having obtained the coronal target feature map, embodiments of the present invention can further input the coronal target feature map into a corresponding second detection module for identification, such as... Figure 5 The coronal detection module based on a transformer detector and primarily using a target detection network, as shown, obtains the detection results of vertebral bodies in coronal images; in other words, this embodiment of the invention can identify vertebral bodies in both sagittal and coronal images. It should be noted that the second detection module for identifying vertebral bodies in coronal images and the first detection module for identifying vertebral bodies in sagittal images can use the same transformer detector (e.g., ...). Figure 5 (as shown), or the first detection module and the second detection module can be different transformer detectors, and the embodiments of the present invention do not limit this.

[0045] Optionally, after step A2 "inputting the coronal target feature map into the second detection module to identify the vertebral bodies in the coronal image" and step 104 "inputting the sagittal target feature map into the first detection module to identify the vertebral bodies in the sagittal image", the method may further include the following steps B1-B2.

[0046] Step B1: Name the vertebral bodies of the spine in the coronal image from top to bottom in the vertical direction to obtain the naming results of the vertebral bodies of the spine in the coronal image.

[0047] In this embodiment of the invention, a semantic segmentation model (such as Casecade-RCNN) can be used. First, the coronal images of the identified vertebrae of the spine are named. For example, according to the order of the vertebrae of the spine in the coronal image from top to bottom, the first 7 vertebrae (cervical vertebrae) are named C1-C7, the 8th to 19th 12 vertebrae (thoracic vertebrae) are named T1-T12, and the 20th to 24th 5 vertebrae (lumbar vertebrae) are named L1-L5, thus obtaining the naming result of the vertebrae of the spine in the coronal image.

[0048] Step B2: Based on the naming results of the vertebral bodies in the coronal images, compare and name the vertebral bodies already identified in the sagittal images.

[0049] Since the coronal and sagittal images are of the same subject (e.g., the same patient), the naming results of the vertebral bodies in the coronal and sagittal images correspond one-to-one. Therefore, the naming results of the vertebral bodies in the coronal images can be compared and matched with the individual vertebral bodies in the sagittal images to aid in the naming of the vertebral bodies in the sagittal images. For example, the first vertebral body at the top of the spine in the sagittal image should be consistent with the vertebral body named C1 in the corresponding coronal image; that is, the first vertebral body at the top of the spine in the sagittal image should also be named C1.

[0050] The spinal vertebral body recognition method provided by the embodiments of the present invention has been described in detail above. This method can also be implemented by a corresponding device. The spinal vertebral body recognition device provided by the embodiments of the present invention will be described in detail below.

[0051] Figure 6 A schematic diagram of the structure of a spinal vertebral body recognition device provided in an embodiment of the present invention is shown. Figure 6 As shown, the spinal vertebral body recognition device includes: an acquisition module 11, a feature extraction module 12, a feature processing module 13, and a recognition module 14.

[0052] The acquisition module 11 is used to acquire coronal and sagittal images of the same object.

[0053] The feature extraction module 12 is used to extract the initial coronal feature map from the coronal image and the initial sagittal feature map from the sagittal image.

[0054] The feature processing module 13 is used to perform feature fusion processing on the initial sagittal feature map based on the initial coronal feature map to obtain the sagittal target feature map corresponding to the sagittal image.

[0055] The recognition module 14 is used to input the sagittal target feature map into the first detection module to identify the spinal vertebrae in the sagittal image.

[0056] Optionally, the feature processing module 13 includes: a first fusion unit and a first overlay unit.

[0057] The first fusion unit is used to enable the initial sagittal feature map to learn features from the initial coronal feature map to obtain the fused feature map corresponding to the sagittal image.

[0058] The first overlay unit is used to overlay the fused feature map corresponding to the sagittal image with the initial sagittal feature map to obtain the target feature map of the sagittal image.

[0059] Optionally, the first fusion unit includes: using an encoder of a converter to cause the sagittal initial feature map to learn features from the coronal initial feature map.

[0060] Optionally, the first overlay unit includes: using a residual network structure to overlay the fused feature map corresponding to the sagittal image with the initial sagittal feature map.

[0061] Optionally, the feature extraction module 12 includes: extracting the initial coronal feature map from the coronal image using the same feature extractor, and extracting the initial sagittal feature map from the sagittal image.

[0062] Optionally, the device further includes a second overlay unit and a detection unit.

[0063] The second overlay unit is used to overlay the fused feature map corresponding to the coronal image with the initial coronal feature map to obtain the coronal target feature map corresponding to the coronal image; the fused feature map corresponding to the coronal image represents the feature map obtained by having the initial coronal feature map learn features from the initial sagittal feature map.

[0064] The detection unit is used to input the coronal target feature map into the second detection module to identify the spinal vertebrae in the coronal image.

[0065] Optionally, the device further includes: a first naming module and a second naming module.

[0066] The first naming module is used to name the spinal vertebrae in the coronal image in a vertical direction from top to bottom, so as to obtain the naming results of the spinal vertebrae in the coronal image.

[0067] The second naming module is used to combine the naming results of the vertebral bodies in the coronal images and name the vertebral bodies that have been identified in the sagittal images.

[0068] The device provided in this invention performs feature fusion processing based on initial coronal feature maps extracted from coronal images and initial sagittal feature maps extracted from sagittal images. This allows for the acquisition of the required sagittal target feature maps for detection, ultimately improving the accuracy of the detection results for the 24 vertebrae in the sagittal plane. This device utilizes coronal images of the same object (e.g., a patient) to assist in identifying the sagittal images of that object. Specifically, it uses dual-view features of the same object to constrain the relationships between vertebrae, avoiding serious errors such as multiple or missed vertebrae detections. Compared to traditional vertebral body identification in sagittal images, it avoids the blurring effect caused by the patient's arm obscuring the vertebrae, resulting in better detection performance.

[0069] In addition, embodiments of the present invention also provide an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor. The transceiver, the memory, and the processor are connected via the bus. When the computer program is executed by the processor, it implements the various processes of the above-described spinal vertebral body recognition method embodiments and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0070] For details, see Figure 7 As shown, this embodiment of the invention also provides an electronic device, which includes a bus 1110, a processor 1120, a transceiver 1130, a bus interface 1140, a memory 1150, and a user interface 1160.

[0071] In this embodiment of the invention, the electronic device further includes a computer program stored in a memory 1150 and executable on a processor 1120, wherein the computer program, when executed by the processor 1120, implements the various processes of the above-described spinal vertebral body recognition method embodiment.

[0072] Transceiver 1130 is used to receive and send data under the control of processor 1120.

[0073] In this embodiment of the invention, a bus architecture (represented by bus 1110) is used. Bus 1110 may include any number of interconnected buses and bridges. Bus 1110 connects various circuits, including one or more processors represented by processor 1120 and memory represented by memory 1150.

[0074] Bus 1110 represents one or more of several types of bus architectures, including memory buses and memory controllers, peripheral buses, Accelerated Graphics Port (AGP), processors, or local buses using any bus architecture from various bus architectures. As an example and not a limitation, such architectures include: Industry Standard Architecture (ISA) buses, Micro Channel Architecture (MCA) buses, Enhanced ISA (EISA) buses, Video Electronics Standards Association (VESA) buses, and Peripheral Component Interconnect (PCI) buses.

[0075] The processor 1120 can be an integrated circuit chip with signal processing capabilities. In implementation, the steps of the above method embodiments can be completed by integrated logic circuits in the processor hardware or by instructions in software form. The processors mentioned above include: general-purpose processors, central processing units (CPUs), network processors (NPs), digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), complex programmable logic devices (CPLDs), programmable logic arrays (PLAs), microcontroller units (MCUs) or other programmable logic devices, discrete gates, transistor logic devices, and discrete hardware components. They can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. For example, the processor can be a single-core processor or a multi-core processor, and the processor can be integrated on a single chip or located on multiple different chips.

[0076] Processor 1120 can be a microprocessor or any conventional processor. The method steps disclosed in the embodiments of the present invention can be directly executed by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in readable storage media known in the art, such as Random Access Memory (RAM), Flash Memory, Read-Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), registers, etc. The readable storage medium is located in the memory, and the processor reads the information in the memory and, in conjunction with its hardware, completes the steps of the above method.

[0077] Bus 1110 can also connect various other circuits, such as peripheral devices, voltage regulators, or power management circuits. Bus interface 1140 provides an interface between bus 1110 and transceiver 1130, all of which are well known in the art. Therefore, embodiments of the present invention will not be described further.

[0078] Transceiver 1130 can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. For example, transceiver 1130 receives external data from other devices, and transceiver 1130 is used to send data processed by processor 1120 to other devices. Depending on the nature of the computer system, a user interface 1160 may also be provided, such as a touchscreen, physical keyboard, monitor, mouse, speaker, microphone, trackball, joystick, or stylus.

[0079] It should be understood that, in embodiments of the present invention, memory 1150 may further include memory remotely configured relative to processor 1120, and such remotely configured memory can be connected to a server via a network. One or more portions of the aforementioned network may be an ad hoc network, intranet, extranet, virtual private network (VPN), local area network (LAN), wireless local area network (WLAN), wide area network (WAN), wireless wide area network (WWAN), metropolitan area network (MAN), Internet, public switched telephone network (PSTN), ordinary old-style telephone service (POTS), cellular telephone network, wireless network, Wi-Fi network, and combinations of two or more of the aforementioned networks. For example, cellular telephone networks and wireless networks can be Global System for Mobile Communications (GSM), Code Division Multiple Access (CDMA), WiMAX, General Packet Radio Service (GPRS), Wideband Code Division Multiple Access (WCDMA), Long Term Evolution (LTE), LTE Frequency Division Duplex (FDD), LTE Time Division Duplex (TDD), Advanced Long Term Evolution (LTE-A), Universal Mobile Telecommunications System (UMTS), Enhanced Mobile Broadband (eMBB), Massive Machine Type Communication (mMTC), Ultra Reliable Low Latency Communications (uRLLC), etc.

[0080] It should be understood that the memory 1150 in the embodiments of the present invention may be volatile memory or non-volatile memory, or may include both volatile memory and non-volatile memory. Non-volatile memory includes: read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory.

[0081] Volatile memory includes random access memory (RAM), which serves as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (SLDRAM), and direct memory bus RAM (DRRAM). The memory 1150 of the electronic device described in this embodiment includes, but is not limited to, the above-described and any other suitable types of memory.

[0082] In this embodiment of the invention, the memory 1150 stores the following elements of the operating system 1151 and the application 1152: executable modules, data structures, or subsets thereof, or extended sets thereof.

[0083] Specifically, the operating system 1151 includes various system programs, such as a framework layer, a core library layer, and a driver layer, used to implement various basic business functions and handle hardware-based tasks. The application program 1152 includes various applications, such as a media player and a browser, used to implement various application functions. Programs implementing the methods of this embodiment of the invention can be included in the application program 1152. The application program 1152 includes applets, objects, components, logic, data structures, and other computer system executable instructions that perform specific tasks or implement specific abstract data types.

[0084] In addition, this embodiment of the invention also provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements the various processes of the above-described spinal vertebral body recognition method embodiment and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0085] Computer-readable storage media include: permanent and non-permanent, removable and non-removable media, which are tangible devices capable of retaining and storing instructions for use by an instruction execution device. Computer-readable storage media include: electronic storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, and any suitable combination thereof. Computer-readable storage media include: phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), non-volatile random access memory (NVRAM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, optical disc read-only memory (CD-ROM), digital versatile optical disc (DVD) or other optical storage, magnetic tape storage, magnetic disk storage or other magnetic storage devices, memory sticks, mechanical encoding devices (e.g., punched cards or raised structures in grooves on which instructions are recorded), or any other non-transfer medium that can be used to store information accessible by a computing device. As defined in the embodiments of the present invention, computer-readable storage media do not include temporary signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses passing through fiber optic cables), or electrical signals transmitted through wires.

[0086] In the several embodiments provided in this application, it should be understood that the disclosed apparatus, electronic devices, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, or it may be an electrical, mechanical, or other form of connection.

[0087] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to solve the problems addressed by the embodiments of the present invention, depending on actual needs.

[0088] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0089] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (including: a personal computer, a server, a data center, or other network device) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media listed above that can store program code.

[0090] In the description of the embodiments of the present invention, those skilled in the art should understand that the embodiments of the present invention can be implemented as methods, apparatuses, electronic devices, and computer-readable storage media. Therefore, the embodiments of the present invention can be specifically implemented in the following forms: entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software. Furthermore, in some embodiments, the embodiments of the present invention can also be implemented as a computer program product in one or more computer-readable storage media, the computer-readable storage media containing computer program code.

[0091] The aforementioned computer-readable storage medium may be any combination of one or more computer-readable storage media. Computer-readable storage media include: electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, optical disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any combination thereof. In embodiments of the present invention, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0092] The computer program code contained in the aforementioned computer-readable storage medium may be transmitted using any suitable medium, including wireless, wire, optical fiber, radio frequency (RF), or any suitable combination thereof.

[0093] Computer program code for performing the operations of the embodiments of the present invention can be written in assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, integrated circuit configuration data, or in one or more programming languages ​​or combinations thereof. The programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The computer program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer or an external computer via any type of network, including a local area network (LAN) or a wide area network (WAN).

[0094] The embodiments of the present invention describe the provided methods, apparatus, and electronic devices through flowcharts and / or block diagrams.

[0095] It should be understood that each block of a flowchart and / or block diagram, as well as combinations of blocks in a flowchart and / or block diagram, can be implemented by computer-readable program instructions. These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine that, when executed by a computer or other programmable data processing apparatus, creates means for implementing the functions / operations specified in the blocks of the flowchart and / or block diagram.

[0096] These computer-readable program instructions may also be stored in a computer-readable storage medium that enables a computer or other programmable data processing device to function in a particular manner. In this way, the instructions stored in the computer-readable storage medium produce an instruction apparatus product that includes the functions / operations specified in the blocks of a flowchart and / or block diagram.

[0097] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus or other device to produce a computer-implemented process, such that the instructions that execute on the computer or other programmable data processing apparatus provide a process for implementing the functions / operations specified in the blocks of the flowchart and / or block diagram.

[0098] The above description is merely a specific implementation of the embodiments of the present invention, but the protection scope of the embodiments of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the embodiments of the present invention should be included within the protection scope of the embodiments of the present invention. Therefore, the protection scope of the embodiments of the present invention should be determined by the protection scope of the claims.

Claims

1. A method of identifying a spinal vertebral body, comprising: The method comprises the following steps: obtaining a coronal picture and a sagittal picture of the same object; extracting a coronal initial feature map from the coronal picture; extracting a sagittal initial feature map from the sagittal picture; learning features from the coronal initial feature map by the sagittal initial feature map to obtain a fusion feature map corresponding to the sagittal picture; superimposing the fusion feature map corresponding to the sagittal picture and the sagittal initial feature map to obtain a target feature map of the sagittal picture; inputting the target feature map of the sagittal picture into a first detection module to identify the spinal vertebrae in the sagittal picture.

2. The method of claim 1, wherein, The step of learning features from the coronal initial feature map by the sagittal initial feature map comprises the following steps: learning features from the coronal initial feature map by the sagittal initial feature map using an encoder of a converter.

3. The method of claim 1, wherein, The step of superimposing the fusion feature map corresponding to the sagittal picture and the sagittal initial feature map comprises the following steps: superimposing the fusion feature map corresponding to the sagittal picture and the sagittal initial feature map using a residual network structure.

4. The method of claim 1, wherein, The steps of extracting a coronal initial feature map from the coronal picture and extracting a sagittal initial feature map from the sagittal picture comprise the following steps: extracting the coronal initial feature map from the coronal picture and extracting the sagittal initial feature map from the sagittal picture using the same feature extractor.

5. The method of claim 1, wherein, After the steps of extracting a coronal initial feature map from the coronal picture and extracting a sagittal initial feature map from the sagittal picture, the method further comprises the following steps: superimposing a fusion feature map corresponding to the coronal picture and the coronal initial feature map to obtain a coronal target feature map corresponding to the coronal picture; the fusion feature map corresponding to the coronal picture represents a feature map obtained by learning features from the coronal initial feature map by the sagittal initial feature map; inputting the coronal target feature map into a second detection module to identify the spinal vertebrae in the coronal picture.

6. The method of claim 5, wherein, After the steps of inputting the coronal target feature map into a second detection module to identify the spinal vertebrae in the coronal picture and inputting the target feature map of the sagittal picture into a first detection module to identify the spinal vertebrae in the sagittal picture, the method further comprises the following steps: naming the identified spinal vertebrae in the coronal picture in a vertical direction from top to bottom to obtain a naming result of the spinal vertebrae in the coronal picture; naming the identified spinal vertebrae in the sagittal picture by referring to the naming result of the spinal vertebrae in the coronal picture.

7. A spinal vertebral body identification device, comprising: The method comprises the following steps: obtaining a coronal picture and a sagittal picture of the same object; extracting a coronal initial feature map from the coronal picture; extracting a sagittal initial feature map from the sagittal picture; learning features from the coronal initial feature map by the sagittal initial feature map to obtain a fusion feature map corresponding to the sagittal picture; superimposing the fusion feature map corresponding to the sagittal picture and the sagittal initial feature map to obtain a target feature map of the sagittal picture; inputting the target feature map of the sagittal picture into a first detection module to identify the spinal vertebrae in the sagittal picture. The step of learning features from the coronal initial feature map by the sagittal initial feature map comprises the following steps: learning features from the coronal initial feature map by the sagittal initial feature map using an encoder of a converter. The step of superimposing the fusion feature map corresponding to the sagittal picture and the sagittal initial feature map comprises the following steps: superimposing the fusion feature map corresponding to the sagittal picture and the sagittal initial feature map using a residual network structure. The steps of extracting a coronal initial feature map from the coronal picture and extracting a sagittal initial feature map from the sagittal picture comprise the following steps: extracting the coronal initial feature map from the coronal picture and extracting the sagittal initial feature map from the sagittal picture using the same feature extractor. After the steps of extracting a coronal initial feature map from the coronal picture and extracting a sagittal initial feature map from the sagittal picture, the method further comprises the following steps: superimposing a fusion feature map corresponding to the coronal picture and the coronal initial feature map to obtain a coronal target feature map corresponding to the coronal picture; the fusion feature map corresponding to the coronal picture represents a feature map obtained by learning features from the coronal initial feature map by the sagittal initial feature map; inputting the coronal target feature map into a second detection module to identify the spinal vertebrae in the coronal picture. After the steps of inputting the coronal target feature map into a second detection module to identify the spinal vertebrae in the coronal picture and inputting the target feature map of the sagittal picture into a first detection module to identify the spinal vertebrae in the sagittal picture, the method further comprises the following steps: naming the identified spinal vertebrae in the coronal picture in a vertical direction from top to bottom to obtain a naming result of the spinal vertebrae in the coronal picture; naming the identified spinal vertebrae in the sagittal picture by referring to the naming result of the spinal vertebrae in the coronal picture. The method comprises the following steps: obtaining a coronal picture and a sagittal picture of the same object; extracting a coronal initial feature map from the coronal picture; extracting a sagittal initial feature map from the sagittal picture; The feature processing module is configured to learn features from the coronal initial feature map by the sagittal initial feature map to obtain a fusion feature map corresponding to the sagittal picture; and superimpose the fusion feature map corresponding to the sagittal picture and the sagittal initial feature map to obtain a target feature map of the sagittal picture. The identification module is configured to input the sagittal target feature map into a first detection module to identify a spinal vertebra in the sagittal picture.

8. An electronic device comprising a bus, a transceiver, a memory, a processor, and a computer program stored on the memory and executable on the processor, the transceiver, the memory and the processor being connected by the bus, characterized in that, The computer program, when executed by the processor, implements the steps in the spinal vertebra identification method according to any one of claims 1 to 6.

9. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by the processor, implements the steps in the spinal vertebra identification method according to any one of claims 1 to 6.

Images