WO2021189914A1 - 电子设备、医学图像指标生成方法、装置及存储介质 - Google Patents

电子设备、医学图像指标生成方法、装置及存储介质 Download PDF

Info

Publication number
WO2021189914A1
WO2021189914A1 PCT/CN2020/131994 CN2020131994W WO2021189914A1 WO 2021189914 A1 WO2021189914 A1 WO 2021189914A1 CN 2020131994 W CN2020131994 W CN 2020131994W WO 2021189914 A1 WO2021189914 A1 WO 2021189914A1
Authority
WO
WIPO (PCT)
Prior art keywords
segmentation
lung field
image
index
network
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2020/131994
Other languages
English (en)
French (fr)
Inventor
刘新卉
叶苓
李楠楠
黄凌云
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ping An Technology Shenzhen Co Ltd
Original Assignee
Ping An Technology Shenzhen Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Ping An Technology Shenzhen Co Ltd filed Critical Ping An Technology Shenzhen Co Ltd
Publication of WO2021189914A1 publication Critical patent/WO2021189914A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/0002Inspection of images, e.g. flaw detection
    • G06T7/0012Biomedical image inspection
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/10Segmentation; Edge detection
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/10Image acquisition modality
    • G06T2207/10116X-ray image
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/20Special algorithmic details
    • G06T2207/20084Artificial neural networks [ANN]
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/20Special algorithmic details
    • G06T2207/20212Image combination
    • G06T2207/20221Image fusion; Image merging
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/30Subject of image; Context of image processing
    • G06T2207/30004Biomedical image processing
    • G06T2207/30008Bone
    • G06T2207/30012Spine; Backbone
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/30Subject of image; Context of image processing
    • G06T2207/30004Biomedical image processing
    • G06T2207/30048Heart; Cardiac
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/30Subject of image; Context of image processing
    • G06T2207/30004Biomedical image processing
    • G06T2207/30061Lung
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/30Subject of image; Context of image processing
    • G06T2207/30004Biomedical image processing
    • G06T2207/30096Tumor; Lesion

Definitions

  • This application relates to the field of image processing technology, and in particular to an electronic device, a method and device for generating medical image indicators, and a computer-readable storage medium.
  • chest X-ray can better show the condition of the thorax, lung tissue and pulmonary blood vessels, heart, chest blood vessels, etc., and provide a more reliable basis for clinical diagnosis.
  • An electronic device provided by the present application includes: at least one processor;
  • the computer program is executed by the at least one processor, so that the at least one processor can execute the following steps:
  • an image index model including a classification network and multiple parallel sub-networks, where the multiple parallel sub-networks include a lung field segmentation sub-network, a heart segmentation sub-network, a spine segmentation sub-network, and a lung field lesion sub-network;
  • the spine segmentation index, the cardiothoracic ratio index, and the lung field lesion segmentation index are pushed to the user.
  • the present application also provides a medical image index generating device, which includes:
  • An abnormality judgment module configured to obtain a chest cavity image, and use the classification network to determine whether the chest cavity image is abnormal
  • a lung field segmentation index generation module configured to use the lung field segmentation sub-network to perform lung field segmentation on the thoracic cavity image to obtain a lung field segmentation index when the thoracic cavity image is abnormal;
  • a heart segmentation index generating module configured to use the heart segmentation sub-network to perform heart segmentation on the thoracic cavity image to obtain a heart segmentation index
  • a spine segmentation index generating module configured to perform spine segmentation on the thoracic cavity image by using the spine segmentation sub-network to obtain a spine segmentation index
  • the cardiothoracic ratio index generating module is configured to calculate the cardiothoracic ratio index of the chest cavity image according to the lung field segmentation index and the heart segmentation index;
  • a lung field lesion segmentation index generation module configured to perform lung field lesion segmentation on the thoracic cavity image by using the lung field lesion sub-network to obtain a lung field lesion segmentation indicator
  • the index push module is used to push the spine segmentation index, the cardiothoracic ratio index, and the lung field lesion segmentation index to the user.
  • the present application also provides a computer-readable storage medium, including a storage data area and a storage program area, the storage data area stores created data, and the storage program area stores a computer program; wherein the computer program is executed when the processor is executed:
  • an image index model including a classification network and multiple parallel sub-networks, where the multiple parallel sub-networks include a lung field segmentation sub-network, a heart segmentation sub-network, a spine segmentation sub-network, and a lung field lesion sub-network;
  • the spine segmentation index, the cardiothoracic ratio index, and the lung field lesion segmentation index are pushed to the user.
  • the present application also provides a method for generating medical image indicators, the method including:
  • an image index model including a classification network and multiple parallel sub-networks, where the multiple parallel sub-networks include a lung field segmentation sub-network, a heart segmentation sub-network, a spine segmentation sub-network, and a lung field lesion sub-network;
  • the spine segmentation index, the cardiothoracic ratio index, and the lung field lesion segmentation index are pushed to the user.
  • FIG. 1 is a schematic diagram of the internal structure of an electronic device provided by an embodiment of the application.
  • FIG. 2 is a schematic flowchart of a method for generating medical image indicators provided by an embodiment of the application
  • FIG. 3 is a schematic structural diagram of an image index model provided by an embodiment of the application.
  • FIG. 4 is a schematic structural diagram of a lung field lesion sub-network provided by an embodiment of the application.
  • FIG. 5 is a schematic diagram of modules of a medical image index generating device provided by an embodiment of the application.
  • the embodiment of the present application provides an electronic device.
  • the electronic device may be, for example, at least one of a server and a terminal.
  • the server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.
  • FIG. 1 it is a schematic diagram of the internal structure of an electronic device provided by an embodiment of this application.
  • the electronic device 1 may include a processor 10, a memory 11, and a bus, and may also include a computer program stored in the memory 11 and running on the processor 10, such as a medical image indicator generating program 12.
  • the memory 11 may be volatile or non-volatile.
  • the memory 11 includes at least one type of readable storage medium.
  • the readable storage medium includes flash memory, mobile hard disk, and multimedia card. , Card-type memory (for example: SD or DX memory, etc.), magnetic memory, magnetic disk, optical disk, etc.
  • the memory 11 may be an internal storage unit of the electronic device 1 in some embodiments, for example, a mobile hard disk of the electronic device 1. In other embodiments, the memory 11 may also be an external storage device of the electronic device 1, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), and a secure digital (Secure Digital) equipped on the electronic device 1. , SD) card, flash card (Flash Card), etc.
  • the memory 11 may also include both an internal storage unit of the electronic device 1 and an external storage device.
  • the memory 11 can be used not only to store application software and various data installed in the electronic device 1, such as the code of the medical image index generating program 12, etc., but also to temporarily store data that has been output or will be output.
  • the processor 10 may be composed of integrated circuits in some embodiments, for example, may be composed of a single packaged integrated circuit, or may be composed of multiple integrated circuits with the same function or different functions, including one or more Combinations of central processing unit (CPU), microprocessor, digital processing chip, graphics processor, and various control chips, etc.
  • the processor 10 is the control core (Control Unit) of the electronic device, which uses various interfaces and lines to connect the various components of the entire electronic device, and runs or executes programs or modules stored in the memory 11 (such as executing Medical image index generation program, etc.), and call data stored in the memory 11 to execute various functions of the electronic device 1 and process data.
  • Control Unit Control Unit
  • the bus may be a peripheral component interconnection standard (peripheral component interconnection standard) interconnect, PCI for short) bus or extended industry standard structure (extended industry standard architecture, EISA for short) bus, etc.
  • the bus can be divided into address bus, data bus, control bus and so on.
  • the bus is configured to implement connection and communication between the memory 11 and at least one processor 10 and the like.
  • FIG. 1 only shows an electronic device with components. Those skilled in the art can understand that the structure shown in FIG. 1 does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown in the figure. Components, or a combination of certain components, or different component arrangements.
  • the electronic device 1 may also include a power source (such as a battery) for supplying power to various components.
  • the power source may be logically connected to the at least one processor 10 through a power management device, thereby controlling power
  • the device implements functions such as charge management, discharge management, and power consumption management.
  • the power supply may also include any components such as one or more DC or AC power supplies, recharging devices, power failure detection circuits, power converters or inverters, and power status indicators.
  • the electronic device 1 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.
  • the electronic device 1 may also include a network interface.
  • the network interface may include a wired interface and/or a wireless interface (such as a Wi-Fi interface, a Bluetooth interface, etc.), which is usually used in the electronic device 1 Establish a communication connection with other electronic devices.
  • the electronic device 1 may also include a user interface.
  • the user interface may be a display (Display) and an input unit (such as a keyboard (Keyboard)).
  • the user interface may also be a standard wired interface or a wireless interface.
  • the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, etc.
  • the display can also be appropriately called a display screen or a display unit, which is used to display the information processed in the electronic device 1 and to display a visualized user interface.
  • the medical image index generating program 12 stored in the memory 11 in the electronic device 1 is a combination of multiple computer programs.
  • a method for generating a medical image index can be implemented.
  • the medical image index generating method can refer to the following description in the flowchart shown in FIG. 2.
  • FIG. 2 it is a schematic flowchart of a method for generating a medical image index according to an embodiment of this application.
  • the method can be executed by a device, and the device can be implemented by software and/or hardware.
  • the method for generating medical image indicators includes:
  • an image index model including a classification network and multiple parallel sub-networks, where the multiple parallel sub-networks include a lung field segmentation sub-network, a heart segmentation sub-network, a spine segmentation sub-network, and a lung field lesion sub-network.
  • the image index model is used to obtain medical index data in the input medical image.
  • FIG. 2 is a schematic structural diagram of an image index model provided by an embodiment of the application.
  • the image index model includes a classification network and a plurality of parallel sub-networks, wherein the plurality of parallel sub-networks include lung fields. Segmentation sub-network, heart segmentation sub-network, spine segmentation sub-network and lung field lesion sub-network.
  • the classification network is a two-classification network with image classification function, such as VGG Net, ResNet, ResNeXt, SE-Net, etc., and the classification network is used to perform processing on the chest cavity image input into the image index model. Preliminary classification. When the classification network determines that the chest image is normal, the normal result is directly output to the user. When the classification network determines that the chest image is abnormal, the chest image is input to multiple parallel sub-networks to obtain the medicine in the medical image. Indicator data.
  • image classification function such as VGG Net, ResNet, ResNeXt, SE-Net, etc.
  • the lung field segmentation sub-network is used to segment the lung field area in the thoracic cavity image to obtain a lung field segmentation index
  • the heart segmentation sub-network is used to segment the heart area in the thoracic cavity image to obtain the heart segmentation index
  • the spine segmentation sub-network is used to segment the spine area in the thoracic cavity image to obtain the spine segmentation index
  • the lung field lesion segmentation sub-network is used to segment the lesion in the lung field area in the thoracic cavity image to obtain the lung field lesion segmentation index.
  • a python sentence with a data capture function can be used to obtain the chest cavity image from a pre-built blockchain node, and the high throughput of the blockchain node for data can be used to improve the acquisition of the chest cavity image. s efficiency.
  • the classification network is a VGG network
  • the VGG network includes a plurality of cascaded convolution kernels.
  • multiple cascaded convolution kernels with a size smaller than or equal to a preset size are used to replace a traditional larger convolution kernel to perform feature extraction on the chest cavity image.
  • the preset size may be a 3X3 convolution kernel.
  • the cascaded convolution kernel may be four 3X3 convolution kernels.
  • the judging whether the thoracic cavity image is abnormal by using the classification network includes:
  • Image classification is performed on the fusion feature map by using an image classification algorithm to obtain a classification result of whether the chest cavity image is abnormal.
  • the image classification algorithm includes but is not limited to the softmax activation function.
  • the image classification algorithm can be used to calculate the target category probability of the fused feature map as the preset target category. When the target category probability is greater than the preset target probability threshold , Determining that the classification result is abnormal, and when the target category probability is less than or equal to a preset target probability threshold, determining that the classification result is normal.
  • the forward convolution kernel and the backward convolution kernel are relative terms.
  • the classification network includes 4 cascaded convolution kernels, the first convolution that performs convolution on the chest image
  • the kernel is forward convolution compared to the second convolution kernel that convolves the chest image, the third convolution kernel that convolves the chest image, and the fourth convolution kernel that convolves the chest image.
  • Convolution kernel; the second convolution kernel that convolves the standard chest image is a backward convolution kernel compared to the first convolution kernel that convolves the chest image, and so on.
  • the backward convolution convolves the result obtained in the forward direction (forward feature map) to obtain the backward feature map.
  • the backward convolution will convolve the result obtained in the forward direction (forward feature map) to obtain the backward feature map.
  • the thoracic cavity image is convolved in the first convolution kernel to obtain the first feature map
  • the embodiment of this application adds a classification network before multiple parallel sub-networks, which can realize the screening of chest images, avoid detecting normal chest images, improve the efficiency of obtaining effective medical index data, and use more in the classification network.
  • a smaller cascaded convolution kernel replaces a common larger convolution kernel for feature extraction of the chest image, which can preserve the receptive field while increasing the non-linear ability of the classification network and improve the classification accuracy of the classification network.
  • the chest image when the chest image is abnormal, the chest image is input to the lung field segmentation sub-network, and the lung field segmentation sub-network is used to perform lung field segmentation on the chest image to obtain the lung field
  • the segmentation index wherein the segmentation index of the lung field includes but is not limited to the maximum transverse diameter of the lung field.
  • lung field segmentation sub-network to perform lung field segmentation on the thoracic cavity image to obtain a lung field segmentation index includes:
  • lung field segmentation sub-network to perform lung field down-sampling on the thoracic cavity image to obtain lung field features
  • the lung field segmentation index is the maximum transverse diameter of the lung field in the chest cavity image.
  • the segmentation processing includes classifying pixels in the lung field restored image, the preset first category includes a plurality of preset categories, and the calculation of the lung field according to the preset first category probability
  • the restored image is segmented to obtain the lung field segmentation result, that is, it is determined that the preset category with the largest probability value of the first category is the segmentation result of the lung field restored image, and the maximum transverse diameter of the lung field can be determined according to the segmentation result (That is, the index of lung field segmentation).
  • the preset first category includes category D, category E, and category F
  • the probability that the target pixel in the lung field restored image is of category D in the preset first category is 20%
  • the target pixel in the lung field restored image is
  • the default probability of category E in the first category is 70%
  • the probability that the target pixel in the lung field restoration image is of category F in the preset first category is 40%
  • the target pixel in the lung field restoration image is determined to be category E
  • the lung field segmentation index is obtained.
  • the lung field feature is up-sampled by a first threshold multiple to obtain an intermediate feature map, and then the intermediate feature map is up-sampled by a second threshold multiple to obtain a lung field restoration image, avoiding directly uploading the lung field feature
  • the up-sampling factor is too large, which leads to the loss of image features in the lung field restoration image, which improves the integrity of the feature information in the lung field restoration image.
  • the performing heart segmentation on the thoracic cavity image by using the heart segmentation sub-network includes:
  • the heart segmentation result is the largest left transverse diameter of the heart and the largest right transverse diameter of the heart in the thoracic cavity image.
  • the spine segmentation index includes the scoliosis angle.
  • said using the spine segmentation sub-network to perform spine segmentation on the thoracic cavity image to obtain a spine segmentation index includes: using the spine segmentation sub-network to perform spinal convolution processing on the thoracic cavity image to obtain a convolutional image Perform pooling processing on the convolution image to obtain a pooled image; perform full connection processing on the pooled image to obtain a full connection feature; use an image segmentation algorithm to segment the full connection feature to obtain a spine segmentation index .
  • the embodiment of the present application uses an image segmentation algorithm to segment the fully connected features to obtain the upper and lower vertebrae of scoliosis, and generate two straight lines along the vertebral surfaces of the upper and lower vertebrae to determine The intersection angle of the two straight lines is the index of spine segmentation (ie, the scoliosis angle).
  • the upper and lower vertebrae are the vertebral bodies with the largest inclination to the concave side of the scoliosis in the scoliosis.
  • the calculation of the cardio-thoracic ratio index of the chest cavity image according to the lung field segmentation index and the heart segmentation index includes:
  • F is the cardiothoracic ratio index
  • CD is the largest left transverse diameter of the heart in the heart segmentation index
  • EF is the largest right transverse diameter of the heart in the heart segmentation index
  • AB is the largest lung field in the lung field segmentation index Traverse.
  • the lung field lesion sub-network includes a shared backbone network and multiple parallel independent sub-networks.
  • the lung field lesion segmentation index includes a label of the lung field lesion.
  • Fig. 3 is a schematic structural diagram of a lung field lesion sub-network provided by an embodiment of the application.
  • the lung field lesion sub-network includes a shared backbone network and multiple parallel independent sub-networks.
  • the segmentation of lung field lesions on the thoracic cavity image by using the lung field lesion sub-network to obtain a lung field lesion segmentation index includes:
  • the preset lesion category label includes a plurality of labels, such as a first preset label, a second preset label, and a third preset label.
  • the lung field lesion segmentation index is the predicted image label corresponding to the characteristic image of the lung field lesion.
  • the multiple parallel independent sub-networks respectively use the activation function to calculate the probability value that the label of the lesion feature map is the preset label.
  • the activation function includes but is not limited to a sigmoid activation function.
  • the parallel independent sub-network 1 uses the activation function to calculate the probability that the label of the lesion feature map is the first preset label is X
  • the parallel independent sub-network 2 uses the activation function to calculate the label of the lesion feature map to be the second preset label
  • the probability value of is Y
  • the parallel independent sub-network 3 uses the activation function to calculate the probability value that the label of the lesion feature map is the third preset label is Z.
  • the predicted image label corresponding to the lesion feature map generated according to the probability value is a probability label indicating that the lesion feature map is changed to a target lesion. For example, when the probability value is "X”, the predicted image label "The probability that the lesion feature map is the target lesion is X" is generated for the lesion feature map.
  • pushing the spine segmentation index, the cardiothoracic ratio index, and the lung field lesion segmentation index to the user includes:
  • the push queue task can be uploaded by the user, and the push queue task includes a preset push sequence.
  • multiple indicator data will be obtained.
  • a large amount of data can be cut and pushed in batches by setting a push queue, avoiding the occupation and waste of computing resources due to data congestion, and improving the The efficiency of the push processing of multiple indicator data.
  • By processing multiple pieces of indicator data that need to be pushed in batches it is ensured that the previous batch of indicator data is pushed and then the next batch of indicator data is processed.
  • the electronic device of the embodiment of the present application adds a classification network in front of multiple parallel sub-networks to determine whether the chest image is abnormal, which can realize the screening of the chest image, avoid the detection of normal chest images, and improve the acquisition of effective medical indicators.
  • Data efficiency segment the thoracic cavity image through the lung field segmentation sub-network, the heart segmentation sub-network and the spine segmentation sub-network respectively to obtain the lung field segmentation index, the heart segmentation index and the spine segmentation index, and pass the lung field segmentation index and the heart segmentation index
  • the index is calculated to obtain the cardiothoracic ratio index, the lung field lesion sub-network is used to segment the thoracic cavity image, and the lung field lesion segmentation index is obtained.
  • the chest image is processed separately through multiple parallel sub-networks, and the chest image is obtained.
  • the medical image index generation method proposed in this application can obtain medical index data in medical images, and can be used for diagnosis and analysis of lung diseases.
  • FIG. 5 it is a schematic diagram of the modules of the medical image index generating device of the present application.
  • the medical image index generating device 100 described in this application can be installed in an electronic device.
  • the medical image index generation device may include a model acquisition module 101, an abnormality judgment module 102, a lung field segmentation index generation module 103, a heart segmentation index generation module 104, a spine segmentation index generation module 105, and a cardiothoracic ratio index generation module.
  • Module 106, lung field lesion segmentation index generating module 107, and index pushing module 108 can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can complete fixed functions, and are stored in the memory of the electronic device.
  • each module/unit is as follows:
  • the model acquisition module 101 is used to acquire an image index model including a classification network and multiple parallel sub-networks, wherein the multiple parallel sub-networks include lung field segmentation sub-network, heart segmentation sub-network, spine segmentation sub-network and Lung field lesion sub-network;
  • the abnormality judgment module 102 is configured to obtain a chest cavity image, and use the classification network to determine whether the chest cavity image is abnormal;
  • the lung field segmentation index generating module 103 is configured to perform lung field segmentation on the thoracic cavity image by using the lung field segmentation sub-network when the thoracic cavity image is abnormal to obtain a lung field segmentation index;
  • the heart segmentation index generating module 104 is configured to perform heart segmentation on the thoracic cavity image by using the heart segmentation sub-network to obtain a heart segmentation index;
  • the spine segmentation index generating module 105 is configured to perform spine segmentation on the thoracic cavity image by using the spine segmentation sub-network to obtain a spine segmentation index;
  • the cardio-thoracic ratio index generating module 106 is configured to calculate the cardio-thoracic ratio index of the chest cavity image according to the lung field segmentation index and the heart segmentation index;
  • the lung field lesion segmentation index generating module 107 is configured to perform lung field lesion segmentation on the thoracic cavity image by using the lung field lesion sub-network to obtain a lung field lesion segmentation index;
  • the index pushing module 108 is configured to push the spine segmentation index, the cardiothoracic ratio index, and the lung field lesion segmentation index to the user.
  • the modules in the medical image indicator generating device 100 provided by the embodiments of the present application can use the same technical means as the above-mentioned electronic device embodiments corresponding to FIGS. 1 to 4 and produce the same technical effects.
  • the medical image index generation device of the embodiment of the application judges whether the chest cavity image is abnormal by adding a classification network before multiple parallel sub-networks, which can realize the screening of the chest cavity image and avoid the detection of the normal chest cavity image. Improve the efficiency of obtaining effective medical index data; segment the thoracic cavity image through the lung field segmentation sub-network, the heart segmentation sub-network and the spine segmentation sub-network to obtain the lung field segmentation index, the heart segmentation index and the spine segmentation index, and pass The lung field segmentation index and the heart segmentation index are calculated to obtain the cardiothoracic ratio index.
  • the lung field lesion sub-network is used to segment the thoracic cavity image to obtain the lung field lesion segmentation index, which realizes the processing of the chest image separately through multiple parallel sub-networks , Get various medical index data in the thoracic cavity image. Therefore, the device for generating medical image indicators proposed in this application can acquire medical indicator data in medical images.
  • the module/unit may be stored in a computer-readable storage medium, which may be volatile Yes, it can also be non-volatile.
  • the computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory) .
  • the computer usable storage medium may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required by at least one function, etc.; the storage data area may store a block chain node Use the created data, etc.
  • modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the objectives of the solutions of the embodiments.
  • the functional modules in the various embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit.
  • the above-mentioned integrated unit may be implemented in the form of hardware, or may be implemented in the form of hardware plus software functional modules.
  • the blockchain referred to in this application is a new application mode of computer technology such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm.
  • Blockchain essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information for verification. The validity of the information (anti-counterfeiting) and the generation of the next block.
  • the blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Quality & Reliability (AREA)
  • Measuring And Recording Apparatus For Diagnosis (AREA)
  • Apparatus For Radiation Diagnosis (AREA)

Abstract

一种医学图像指标生成方法、电子设备、装置和介质,涉及图像处理技术。所述方法包括:利用分类网络判断胸腔图像是否异常(S2);当胸腔图像异常时,利用肺野分割子网络获得胸腔图像的肺野分割指标(S3);利用心脏分割子网络获得胸腔图像的心脏分割指标(S4);利用脊柱分割子网络获得胸腔图像的脊柱分割指标(S5);根据肺野分割指标、心脏分割指标计算心胸比指标(S6);利用肺野病灶子网络获得胸腔图像的肺野病灶分割指标(S7);将脊柱分割指标、心胸比指标及肺野病灶分割指标推送给用户(S8)。所述医学图像指标生成设备、装置、方法及其计算机可读存储介质可应用于医学图像分析,可以获取医学图像中的医学指标数据。

Description

电子设备、医学图像指标生成方法、装置及存储介质
本申请要求于2020年10月20日提交中国专利局、申请号为CN202011127501.8、名称为“电子设备、医学图像指标生成方法、装置及存储介质”的中国专利申请的优先权,以及要求于2020年11月24日提交中国专利局、申请号为CN202011327450.3、名称为“电子设备、医学图像指标生成方法、装置及存储介质”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及图像处理技术领域,尤其涉及一种电子设备、医学图像指标生成方法、装置及计算机可读存储介质。
背景技术
X光胸片作为一种经济简便的常规检查,可以较好的显示出胸廓、肺组织与肺血管、心脏、胸部血管等的情况,为临床诊断提供较为可靠的依据。
技术问题
发明人意识到,目前依托于X光胸片的辅助诊断技术和系统大都着眼于对医学图像进行分类,医生仅可以看到医学图像的分类结果,却无法获取分析得到具体的医学指标数据,例如,当胸腔图像中的心胸比在预设比值范围之外时,医生仅可以看到胸腔图像的分类结果为异常病变,但却无法获取心胸比的具体数值,使得医生无法对医学图像作出准确的判断。
技术解决方案
本申请提供的一种电子设备,所述电子设备包括:至少一个处理器;以及,与所述至少一个处理器通信连接的存储器;其中,所述存储器存储有可被所述至少一个处理器执行的计算机程序,所述计算机程序被所述至少一个处理器执行,以使所述至少一个处理器能够执行如下步骤:
获取包含分类网络和多个并行子网络的图像指标模型,其中,所述多个并行子网络包括肺野分割子网络、心脏分割子网络、脊柱分割子网络和肺野病灶子网络;
获取胸腔图像,利用所述分类网络判断所述胸腔图像是否异常;
当所述胸腔图像异常时,利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标;
利用所述心脏分割子网络对所述胸腔图像进行心脏分割,得到心脏分割指标;
利用所述脊柱分割子网络对所述胸腔图像进行脊柱分割,得到脊柱分割指标;
根据所述肺野分割指标和所述心脏分割指标计算所述胸腔图像的心胸比指标;
利用所述肺野病灶子网络对所述胸腔图像进行肺野病灶分割,得到肺野病灶分割指标;
将所述脊柱分割指标、心胸比指标及所述肺野病灶分割指标推送给用户。
本申请还提供一种医学图像指标生成装置,所述装置包括:
模型获取模块,用于获取包含分类网络和多个并行子网络的图像指标模型,其中,所述多个并行子网络包括肺野分割子网络、心脏分割子网络、脊柱分割子网络和肺野病灶子网络;
异常判断模块,用于获取胸腔图像,利用所述分类网络判断所述胸腔图像是否异常;
肺野分割指标生成模块,用于当所述胸腔图像异常时,利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标;
心脏分割指标生成模块,用于利用所述心脏分割子网络对所述胸腔图像进行心脏分割,得到心脏分割指标;
脊柱分割指标生成模块,用于利用所述脊柱分割子网络对所述胸腔图像进行脊柱分割,得到脊柱分割指标;
心胸比指标生成模块,用于根据所述肺野分割指标和所述心脏分割指标计算所述胸腔图像的心胸比指标;
肺野病灶分割指标生成模块,用于利用所述肺野病灶子网络对所述胸腔图像进行肺野病灶分割,得到肺野病灶分割指标;
指标推送模块,用于将所述脊柱分割指标、心胸比指标及所述肺野病灶分割指标推送给用户。
本申请还提供一种计算机可读存储介质,包括存储数据区和存储程序区,存储数据区存储创建的数据,存储程序区存储有计算机程序;其中,所述计算机程序被处理器执行时实现:
获取包含分类网络和多个并行子网络的图像指标模型,其中,所述多个并行子网络包括肺野分割子网络、心脏分割子网络、脊柱分割子网络和肺野病灶子网络;
获取胸腔图像,利用所述分类网络判断所述胸腔图像是否异常;
当所述胸腔图像异常时,利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标;
利用所述心脏分割子网络对所述胸腔图像进行心脏分割,得到心脏分割指标;
利用所述脊柱分割子网络对所述胸腔图像进行脊柱分割,得到脊柱分割指标;
根据所述肺野分割指标和所述心脏分割指标计算所述胸腔图像的心胸比指标;
利用所述肺野病灶子网络对所述胸腔图像进行肺野病灶分割,得到肺野病灶分割指标;
将所述脊柱分割指标、心胸比指标及所述肺野病灶分割指标推送给用户。医学图像指标生成
本申请还提供一种医学图像指标生成方法,所述方法包括:
获取包含分类网络和多个并行子网络的图像指标模型,其中,所述多个并行子网络包括肺野分割子网络、心脏分割子网络、脊柱分割子网络和肺野病灶子网络;
获取胸腔图像,利用所述分类网络判断所述胸腔图像是否异常;
当所述胸腔图像异常时,利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标;
利用所述心脏分割子网络对所述胸腔图像进行心脏分割,得到心脏分割指标;
利用所述脊柱分割子网络对所述胸腔图像进行脊柱分割,得到脊柱分割指标;
根据所述肺野分割指标和所述心脏分割指标计算所述胸腔图像的心胸比指标;
利用所述肺野病灶子网络对所述胸腔图像进行肺野病灶分割,得到肺野病灶分割指标;
将所述脊柱分割指标、心胸比指标及所述肺野病灶分割指标推送给用户。
附图说明
图1为本申请一实施例提供的电子设备的内部结构示意图;
图2为本申请一实施例提供的医学图像指标的生成方法的流程示意图;
图3为本申请一实施例提供的图像指标模型的结构示意图;
图4为本申请一实施例提供的肺野病灶子网络的结构示意图;
图5为本申请一实施例提供的医学图像指标生成装置的模块示意图;
如下具体实施方式将结合上述附图进一步说明本申请。
本发明的实施方式
应当理解,此处所描述的具体实施例仅仅用以解释本申请,并不用于限定本申请。
本申请实施例提供一种电子设备。所述电子设备可以是,例如服务端、终端等的至少一种。所述服务端包括但不限于:单台服务器、服务器集群、云端服务器或云端服务器集群等。
如图1所示,为本申请一实施例提供的电子设备的内部结构示意图。
所述电子设备1可以包括处理器10、存储器11和总线,还可以包括存储在所述存储器11中并可在所述处理器10上运行的计算机程序,如医学图像指标生成程序12。
其中,所述存储器11可以是易失性的,也可以是非易失性的,所述存储器11至少包括一种类型的可读存储介质,所述可读存储介质包括闪存、移动硬盘、多媒体卡、卡型存储器(例如:SD或DX存储器等)、磁性存储器、磁盘、光盘等。所述存储器11在一些实施例中可以是电子设备1的内部存储单元,例如该电子设备1的移动硬盘。所述存储器11在另一些实施例中也可以是电子设备1的外部存储设备,例如电子设备1上配备的插接式移动硬盘、智能存储卡(Smart Media Card, SMC)、安全数字(Secure Digital, SD)卡、闪存卡(Flash Card)等。进一步地,所述存储器11还可以既包括电子设备1的内部存储单元也包括外部存储设备。所述存储器11不仅可以用于存储安装于电子设备1的应用软件及各类数据,例如医学图像指标生成程序12的代码等,还可以用于暂时地存储已经输出或者将要输出的数据。
所述处理器10在一些实施例中可以由集成电路组成,例如可以由单个封装的集成电路所组成,也可以是由多个相同功能或不同功能封装的集成电路所组成,包括一个或者多个中央处理器(Central Processing unit,CPU)、微处理器、数字处理芯片、图形处理器及各种控制芯片的组合等。所述处理器10是所述电子设备的控制核心(Control Unit),利用各种接口和线路连接整个电子设备的各个部件,通过运行或执行存储在所述存储器11内的程序或者模块(例如执行医学图像指标生成程序等),以及调用存储在所述存储器11内的数据,以执行电子设备1的各种功能和处理数据。
所述总线可以是外设部件互连标准(peripheral component interconnect,简称PCI)总线或扩展工业标准结构(extended industry standard architecture,简称EISA)总线等。该总线可以分为地址总线、数据总线、控制总线等。所述总线被设置为实现所述存储器11以及至少一个处理器10等之间的连接通信。
图1仅示出了具有部件的电子设备,本领域技术人员可以理解的是,图1示出的结构并不构成对所述电子设备1的限定,可以包括比图示更少或者更多的部件,或者组合某些部件,或者不同的部件布置。
例如,尽管未示出,所述电子设备1还可以包括给各个部件供电的电源(比如电池),优选地,电源可以通过电源管理装置与所述至少一个处理器10逻辑相连,从而通过电源管理装置实现充电管理、放电管理、以及功耗管理等功能。电源还可以包括一个或一个以上的直流或交流电源、再充电装置、电源故障检测电路、电源转换器或者逆变器、电源状态指示器等任意组件。所述电子设备1还可以包括多种传感器、蓝牙模块、Wi-Fi模块等,在此不再赘述。
进一步地,所述电子设备1还可以包括网络接口,可选地,所述网络接口可以包括有线接口和/或无线接口(如WI-FI接口、蓝牙接口等),通常用于在该电子设备1与其他电子设备之间建立通信连接。
可选地,该电子设备1还可以包括用户接口,用户接口可以是显示器(Display)、输入单元(比如键盘(Keyboard)),可选地,用户接口还可以是标准的有线接口、无线接口。可选地,在一些实施例中,显示器可以是LED显示器、液晶显示器、触控式液晶显示器以及OLED(Organic Light-Emitting Diode,有机发光二极管)触摸器等。其中,显示器也可以适当的称为显示屏或显示单元,用于显示在电子设备1中处理的信息以及用于显示可视化的用户界面。
应该了解,所述实施例仅为说明之用,在专利申请范围上并不受此结构的限制。
所述电子设备1中的所述存储器11存储的医学图像指标生成程序12是多个计算机程序的组合,在所述处理器10中运行时,可以实现一种医学图像指标生成方法。详细地,所述医学图像指标生成方法可参照下述关于图2所示的流程图中的描述。
参照图2所示,为本申请一实施例提供的医学图像指标生成方法的流程示意图。该方法可以由一个装置执行,该装置可以由软件和/或硬件实现。
在本实施例中,医学图像指标生成方法包括:
S1、获取包含分类网络和多个并行子网络的图像指标模型,其中,所述多个并行子网络包括肺野分割子网络、心脏分割子网络、脊柱分割子网络和肺野病灶子网络。
本申请实施例中,所述图像指标模型用于获取输入的医学图像中的医学指标数据。
参图2所示,图2为本申请一实施例提供的图像指标模型的结构示意图,所述图像指标模型包括分类网络和多个并行子网络,其中,所述多个并行子网络包括肺野分割子网络、心脏分割子网络、脊柱分割子网络和肺野病灶子网络。
本申请实施例中,所述分类网络为具有图像分类功能的二分类网络,例如VGG Net、ResNet、ResNeXt、SE-Net等,所述分类网络用于对输入至图像指标模型中的胸腔图像进行初步分类,当所述分类网络判断胸腔图像正常时,则直接将正常结果输出给用户,当所述分类网络判断胸腔图像异常时,将胸腔图像输入至多个并行子网络以获取医学图像中的医学指标数据。
具体地,所述肺野分割子网络用于分割出胸腔图像中的肺野区域,得到肺野分割指标;所述心脏分割子网络用于分割出胸腔图像中的心脏区域,得到心脏分割指标;脊柱分割子网络用于分割出胸腔图像中的脊柱区域,得到脊柱分割指标;肺野病灶分割子网络用于分割出胸腔图像中肺野区域中的病灶,得到肺野病灶分割指标。
S2、获取胸腔图像,利用所述分类网络判断所述胸腔图像是否异常。
本申请实施例中,可利用具有数据抓取功能的python语句从预先构建的区块链节点中获取所述胸腔图像,利用区块链节点对于数据的高吞吐性,可提高获取所述胸腔图像的效率。
本申请一较佳实施例中,所述分类网络为VGG网络,所述VGG网络中包含多个级联的卷积核。
具体的,VGG网络中利用多个级联的尺寸小于等于预设尺寸的卷积核替代传统的一个较大卷积核对胸腔图像进行特征提取。
所述预设尺寸可以为3X3的卷积核。所述级联的卷积核可以为4个3X3的卷积核。
详细地,所述利用所述分类网络判断所述胸腔图像是否异常,包括:
利用所述分类网络中前向卷积核对所述胸腔图像进行卷积,得到前向特征图;
利用所述分类网络中后向卷积核对所述中间卷积图像进行卷积,得到后向特征图;
对获得的前向特征图和后向特征图进行特征融合,得到融合特征图;
利用图像分类算法对所述融合特征图进行图像分类,得到所述胸腔图像是否为异常的分类结果。
详细地,所述图像分类算法包括但不限于softmax激活函数,利用图像分类算法可计算得到融合特征图为预设的目标类别的目标类别概率,当所述目标类别概率大于预设目标概率阈值时,确定所述分类结果为异常,当所述目标类别概率小于或等于预设目标概率阈值时,确定所述分类结果为正常。
具体地,所述前向卷积核与所述后向卷积核是相对而言的,例如,分类网络包括4个级联的卷积核,第一个对胸腔图像进行卷积的卷积核相对于第二个对胸腔图像进行卷积的卷积核、第三个对胸腔图像进行卷积的卷积核和第四个对胸腔图像进行卷积的卷积核来说是前向卷积核;第二个对标准胸腔图像进行卷积的卷积核相对于第一个对胸腔图像进行卷积的卷积核来说是后向卷积核,依此类推。
当前向卷积核为初始卷积核时,后向卷积对前向得到的结果(前向特征图)进行卷积,得到后向特征图。
当前向卷积核不为初始卷积通道时,后向卷积对前向得到的结果(前向特征图)进行卷积,得到后向特征图。
具体地,例如,在第一卷积核中对所述胸腔图像进行卷积,得到第一特征图;
在第二卷积核中对所述第一特征图进行卷积,得到第二特征图;
在第三卷积核中对所述第二特征图进行卷积,得到第三特征图;
在第四卷积核中对所述第三特征图进行卷积,得到第四特征图;
对所述第一特征图、所述第二特征图、所述第三特征图和所述第四特征图进行特征融合得到融合特征图,并利用图像分类算法对所述融合特征图进行图像分类,得到所述分类结果。
本申请实施例在多个并行子网络前增加分类网络,可实现对胸腔图像的筛选,避免对正常的胸腔图像进行检测,提高了获取有效的医学指标数据的效率,且在分类网络中利用多个较小的级联卷积核替换普通的一个较大卷积核对胸腔图像进行特征提取,可保留感受野的同时增加分类网络的非线性能力,提高分类网络的分类精确度。
S3、当所述胸腔图像异常时,利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标。
本申请实施例中,当所述胸腔图像异常时,将所述胸腔图像输入至所述肺野分割子网络,利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标,其中,所述肺野分割指标包括但不限于肺野最大横径。
详细地,所述利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标,包括:
利用所述肺野分割子网络对所述胸腔图像进行肺野下采样,得到肺野特征;
将所述肺野特征进行第一阈值倍数的上采样,得到中间特征图;
将所述中间特征图进行第二阈值倍数的上采样,得到肺野复原图像;
利用第一激活函数计算所述肺野复原图像中各像素点属于预设第一类别的第一类别概率;
根据所述第一类别概率对所述肺野复原图像进行分割处理,得到肺野分割指标。
详细地,所述肺野分割指标为胸腔图像中的肺野最大横径。
具体地,所述分割处理包括对肺野复原图像中像素点进行分类,所述预设第一类别中包括多个预设类别,所述根据所述预设第一类别概率对所述肺野复原图像进行分割处理,得到肺野分割结果,即确定所述第一类别概率值最大的预设类别为肺野复原图像的分割结果,根据所述分割结果即可确定所述肺野最大横径(即肺野分割指标)。
例如,预设第一类别中包括类别D、类别E与类别F,肺野复原图像中目标像素点为预设第一类别中类别D的概率为20%,肺野复原图像中目标像素点为预设第一类别中类别E的概率为70%,肺野复原图像中目标像素点为预设第一类别中类别F的概率为40%,则确定肺野复原图像中目标像素点为类别E,当所述肺野复原图像中所有像素点均完成所述分割操作,得到肺野分割指标。
本申请实施例中将肺野特征进行第一阈值倍数的上采样,得到中间特征图,再将中间特征图进行第二阈值倍数的上采样,得到肺野复原图像,避免直接将肺野特征上采样至肺野复原图像时上采样倍数过大导致肺野复原图像中图像特征的丢失,提高了肺野复原图像中特征信息的完整性。
S4、利用所述心脏分割子网络对所述胸腔图像进行心脏分割,得到心脏分割指标。
本申请实施例中,所述利用所述心脏分割子网络对所述胸腔图像进行心脏分割,包括:
利用所述心脏分割子网络对所述胸腔图像进行心脏下采样,得到心脏特征;
将所述心脏特征进行第一阈值倍数的上采样,得到中间特征图;
将所述中间特征图进行第二阈值倍数的上采样,得到心脏复原图像;
利用第二激活函数计算所述心脏复原图像中各像素点属于预设第二类别的第二类别概率;
根据所述第二类别概率对所述心脏复原图像进行分割处理,得到心脏分割结果。
详细地,所述心脏分割结果为胸腔图像中心脏最大左横径和心脏最大右横径。
S5、利用所述脊柱分割子网络对所述胸腔图像进行脊柱分割,得到脊柱分割指标。
本申请实施例中,所述脊柱分割指标包括脊柱侧弯角度。
详细地,所述利用所述脊柱分割子网络对所述胸腔图像进行脊柱分割,得到脊柱分割指标,包括:利用所述脊柱分割子网络对所述胸腔图进行脊柱卷积处理,得到卷积图像;对所述卷积图像进行池化处理,得到池化图像;对所述池化图像进行全连接处理,得到全连接特征;利用图像分割算法对所述全连接特征进行分割,得到脊柱分割指标。
具体地,本申请实施例利用图像分割算法对所述全连接特征进行分割,得到脊柱侧弯的上端椎和下端椎,沿着所述上端椎和下端椎的椎面生成两条直线,从而确定两条直线的交角为脊柱分割指标(即脊柱侧弯角度)。
详细地,上端椎与下端椎是脊柱侧弯中,向脊柱侧弯凹侧倾斜度最大的椎体。
S6、根据所述肺野分割指标和心脏分割指标计算所述胸腔图像的心胸比指标。
本申请实施例中,所述根据所述肺野分割指标、心脏分割指标计算所述胸腔图像的心胸比指标,包括:
利用如下指标算法根据所述肺野分割指标和心脏分割指标计算所述胸腔图像的心胸比指标:
F=(CD+EF)/AB
其中,F为所述心胸比指标,CD为所述心脏分割指标中心脏最大左横径,EF为所述心脏分割指标中心脏最大右横径,AB为所述肺野分割指标中肺野最大横径。
S7、利用所述肺野病灶子网络对所述胸腔图像进行肺野病灶分割,得到肺野病灶分割指标。
本申请实施例中,所述肺野病灶子网络包括一个共享主干网络与多个并行独立子网络。所述肺野病灶分割指标包括肺野病灶的标签。
参图3所示,图3为本申请一实施例提供的肺野病灶子网络的结构示意图,所述肺野病灶子网络包括一个共享主干网络与多个并行独立子网络。
详细地,所述利用所述肺野病灶子网络对所述胸腔图像进行肺野病灶分割,得到肺野病灶分割指标,包括:
将所述胸腔图像输入至所述肺野病灶子网络进行肺野病灶特征提取,得到肺野病灶特征图像;
利用标签激活函数计算所述肺野病灶特征图像的标签为预设病灶类别标签的概率值;
根据所述概率值生成所述肺野病灶特征图像对应的肺野病灶分割指标。
优选地,所述预设病灶类别标签包括多个标签,例如第一预设标签、第二预设标签和第三预设标签等。
较佳地,所述肺野病灶分割指标即为所述肺野病灶特征图像对应的预测图像标签。
详细地,所述多个并行独立子网络分别利用激活函数计算得到病灶特征图的标签为预设标签的概率值。
所述激活函数包括但不限于sigmoid激活函数。例如,并行独立子网络1利用激活函数计算得到病灶特征图的标签为第一预设标签的概率值为X,并行独立子网络2利用激活函数计算得到病灶特征图的标签为第二预设标签的概率值为Y,并行独立子网络3利用激活函数计算得到病灶特征图的标签为第三预设标签的概率值为Z。
本申请实施例中,根据所述概率值生成的所述病灶特征图对应的预测图像标签是表示改病灶特征图为目标病灶的概率标签。例如,当概率值为“X”时,对所述病灶特征图生成预测图像标签“病灶特征图为目标病灶的概率为X”。
S8、将所述脊柱分割指标、心胸比指标及所述肺野病灶分割指标推送给用户。
本申请实施例中,所述将所述脊柱分割指标、心胸比指标及所述肺野病灶分割指标推送给用户,包括:
获取推送队列任务;
根据所述推送队列任务确定推送顺序;
根据所述推送顺序向用户推送所述脊柱分割指标、心胸比指标及所述肺野病灶分割指标。
详细地,所述推送队列任务可由用户上传,所述推送队列任务中包括预先设定的推送顺序。
实际应用中,会得到多个指标数据,在推送指标数据时,通过设置推送队列能够将大量的数据进行切割并分批进行推送,避免因为数据拥塞而导致计算资源的占用与浪费,提高了对多个指标数据进行推送处理的效率。通过分批处理多份需要推送的指标数据,从而确保前一批指标数据推送结束再继续处理后一批指标数据。
本申请实施例的电子设备通过在多个并行子网络前增加分类网络对胸腔图像是否异常进行判断,可实现对胸腔图像的筛选,避免对正常的胸腔图像进行检测,提高了获取有效的医学指标数据的效率;通过肺野分割子网络、心脏分割子网络和脊柱分割子网络分别对胸腔图像进行分割,得到肺野分割指标、心脏分割指标和脊柱分割指标,并通过肺野分割指标和心脏分割指标计算得到心胸比指标,利用肺野病灶子网络对胸腔图像进行肺野病灶分割,得到肺野病灶分割指标,实现了通过多个并行子网络分别对胸腔图像进行处理,得出胸腔图像中的各项医学指标数据。因此本申请提出的医学图像指标生成方法可以获取获取医学图像中的医学指标数据,可用于肺部疾病的诊断分析。
如图5所示,是本申请医学图像指标生成装置的模块示意图。
本申请所述医学图像指标生成装置100可以安装于电子设备中。根据实现的功能,所述医学图像指标生成装置可以包括模型获取模块101、异常判断模块102、肺野分割指标生成模块103、心脏分割指标生成模块104、脊柱分割指标生成模块105、心胸比指标生成模块106、肺野病灶分割指标生成模块107和指标推送模块108。本申请所述模块也可以称之为单元,是指一种能够被电子设备处理器所执行,并且能够完成固定功能的一系列计算机程序段,其存储在电子设备的存储器中。
在本实施例中,关于各模块/单元的功能如下:
所述模型获取模块101,用于获取包含分类网络和多个并行子网络的图像指标模型,其中,所述多个并行子网络包括肺野分割子网络、心脏分割子网络、脊柱分割子网络和肺野病灶子网络;
所述异常判断模块102,用于获取胸腔图像,利用所述分类网络判断所述胸腔图像是否异常;
所述肺野分割指标生成模块103,用于当所述胸腔图像异常时,利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标;
所述心脏分割指标生成模块104,用于利用所述心脏分割子网络对所述胸腔图像进行心脏分割,得到心脏分割指标;
所述脊柱分割指标生成模块105,用于利用所述脊柱分割子网络对所述胸腔图像进行脊柱分割,得到脊柱分割指标;
所述心胸比指标生成模块106,用于根据所述肺野分割指标和所述心脏分割指标计算所述胸腔图像的心胸比指标;
所述肺野病灶分割指标生成模块107,用于利用所述肺野病灶子网络对所述胸腔图像进行肺野病灶分割,得到肺野病灶分割指标;
所述指标推送模块108,用于将所述脊柱分割指标、心胸比指标及所述肺野病灶分割指标推送给用户。
本申请实施例所提供的医学图像指标生成装置100中的模块能够在使用时利用上述的图1至图4对应的电子设备的实施例一样的技术手段,并产生一样的技术效果。
医学图像指标生成本申请实施例的医学图像指标生成装置通过在多个并行子网络前增加分类网络对胸腔图像是否异常进行判断,可实现对胸腔图像的筛选,避免对正常的胸腔图像进行检测,提高了获取有效的医学指标数据的效率;通过肺野分割子网络、心脏分割子网络和脊柱分割子网络分别对胸腔图像进行分割,得到肺野分割指标、心脏分割指标和脊柱分割指标,并通过肺野分割指标和心脏分割指标计算得到心胸比指标,利用肺野病灶子网络对胸腔图像进行肺野病灶分割,得到肺野病灶分割指标,实现了通过多个并行子网络分别对胸腔图像进行处理,得出胸腔图像中的各项医学指标数据。因此本申请提出的用于医学图像指标生成的装置,可以获取获取医学图像中的医学指标数据。
进一步地,所述模块/单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时,可以存储在一个计算机可读取存储介质中,所述计算机可读存储介质可以是易失性的,也可以是非易失性的。所述计算机可读介质可以包括:能够携带所述计算机程序代码的任何实体或装置、记录介质、U盘、移动硬盘、磁碟、光盘、计算机存储器、只读存储器(ROM,Read-Only Memory)。
进一步地,所述计算机可用存储介质可主要包括存储程序区和存储数据区,其中,存储程序区可存储操作系统、至少一个功能所需的应用程序等;存储数据区可存储根据区块链节点的使用所创建的数据等。
在本申请所提供的几个实施例中,应该理解到,所揭露的设备,装置和方法,可以通过其它的方式实现。例如,以上所描述的装置实施例仅仅是示意性的,例如,所述模块的划分,仅仅为一种逻辑功能划分,实际实现时可以有另外的划分方式。
所述作为分离部件说明的模块可以是或者也可以不是物理上分开的,作为模块显示的部件可以是或者也可以不是物理单元,即可以位于一个地方,或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。
另外,在本申请各个实施例中的各功能模块可以集成在一个处理单元中,也可以是各个单元单独物理存在,也可以两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现,也可以采用硬件加软件功能模块的形式实现。
对于本领域技术人员而言,显然本申请不限于上述示范性实施例的细节,而且在不背离本申请的精神或基本特征的情况下,能够以其他的具体形式实现本申请。
因此,无论从哪一点来看,均应将实施例看作是示范性的,而且是非限制性的,本申请的范围由所附权利要求而不是上述说明限定,因此旨在将落在权利要求的等同要件的含义和范围内的所有变化涵括在本申请内。不应将权利要求中的任何附关联图表记视为限制所涉及的权利要求。
本申请所指区块链是分布式数据存储、点对点传输、共识机制、加密算法等计算机技术的新型应用模式。区块链(Blockchain),本质上是一个去中心化的数据库,是一串使用密码学方法相关联产生的数据块,每一个数据块中包含了一批次网络交易的信息,用于验证其信息的有效性(防伪)和生成下一个区块。区块链可以包括区块链底层平台、平台产品服务层以及应用服务层等。
此外,显然“包括”一词不排除其他单元或步骤,单数不排除复数。系统权利要求中陈述的多个单元或装置也可以由一个单元或装置通过软件或者硬件来实现。第二等词语用来表示名称,而并不表示任何特定的顺序。
最后应说明的是,以上实施例仅用以说明本申请的技术方案而非限制,尽管参照较佳实施例对本申请进行了详细说明,本领域的普通技术人员应当理解,可以对本申请的技术方案进行修改或等同替换,而不脱离本申请技术方案的精神和范围。

Claims (20)

  1. 一种电子设备,其中,所述电子设备包括:
    至少一个处理器;以及
    与所述至少一个处理器通信连接的存储器;其中,所述存储器存储有可被所述至少一个处理器执行的计算机程序,所述计算机程序被所述至少一个处理器执行,以使所述至少一个处理器执行如下步骤:
    获取包含分类网络和多个并行子网络的图像指标模型,其中,所述多个并行子网络包括肺野分割子网络、心脏分割子网络、脊柱分割子网络和肺野病灶子网络;
    获取胸腔图像,利用所述分类网络判断所述胸腔图像是否异常;
    当所述胸腔图像异常时,利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标;
    利用所述心脏分割子网络对所述胸腔图像进行心脏分割,得到心脏分割指标;
    利用所述脊柱分割子网络对所述胸腔图像进行脊柱分割,得到脊柱分割指标;
    根据所述肺野分割指标和所述心脏分割指标计算所述胸腔图像的心胸比指标;
    利用所述肺野病灶子网络对所述胸腔图像进行肺野病灶分割,得到肺野病灶分割指标;
    将所述脊柱分割指标、心胸比指标及所述肺野病灶分割指标推送给用户。
  2. 如权利要求1所述的电子设备,其中,所述利用所述分类网络判断所述胸腔图像是否异常,包括:
    利用所述分类网络中前向卷积核对所述胸腔图像进行卷积,得到前向特征图;
    利用所述分类网络中后向卷积核对所述中间卷积图像进行卷积,得到后向特征图;
    对获得的前向特征图和后向特征图进行特征融合,得到融合特征图;
    利用图像分类算法对所述融合特征图进行图像分类,得到所述胸腔图像是否为异常的分类结果。
  3. 如权利要求1所述的电子设备,其中,所述利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标,包括:
    利用所述肺野分割子网络对所述胸腔图像进行肺野下采样,得到肺野特征;
    将所述肺野特征进行第一阈值倍数的上采样,得到中间特征图;
    将所述中间特征图进行第二阈值倍数的上采样,得到肺野复原图像;
    利用第一激活函数计算所述肺野复原图像中各像素点属于预设第一类别的第一类别概率;
    根据所述第一类别概率对所述肺野复原图像进行分割处理,得到肺野分割指标。
  4. 如权利要求1所述的电子设备,其中,所述利用所述脊柱分割子网络对所述胸腔图像进行脊柱分割,得到脊柱分割指标,包括:
    利用所述脊柱分割子网络对所述胸腔图进行脊柱卷积处理,得到卷积图像;
    对所述卷积图像进行池化处理,得到池化图像;
    对所述池化图像进行全连接处理,得到全连接特征;
    利用图像分割算法对所述全连接特征进行分割,得到脊柱分割指标。
  5. 如权利要求1至4中任一项所述的电子设备,其中,所述利用所述肺野病灶子网络对所述胸腔图像进行肺野病灶分割,得到肺野病灶分割指标,包括:
    将所述胸腔图像输入至所述肺野病灶子网络进行肺野病灶特征提取,得到肺野病灶特征图像;
    利用标签激活函数计算所述肺野病灶特征图像的标签为预设病灶类别标签的概率值;
    根据所述概率值生成所述肺野病灶特征图像对应的肺野病灶分割指标。
  6. 如权利要求1至4中任一项所述的电子设备,其中,所述根据所述肺野分割指标和心脏分割指标计算所述胸腔图像的心胸比指标,包括:
    利用如下指标算法根据所述肺野分割指标、心脏分割指标计算所述胸腔图像的心胸比指标:
    F=(CD+EF)/AB
    其中,F为所述心胸比指标,CD为所述心脏分割指标中心脏最大左横径,EF为所述心脏分割指标中心脏最大右横径,AB为所述肺野分割指标中肺野最大横径。
  7. 如权利要求1所述的电子设备,其中,所述分类网络为VGG网络,所述VGG网络中包含多个级联的卷积核。
  8. 一种医学图像指标生成装置,其中,所述装置包括:
    模型获取模块,用于获取包含分类网络和多个并行子网络的图像指标模型,其中,所述多个并行子网络包括肺野分割子网络、心脏分割子网络、脊柱分割子网络和肺野病灶子网络;
    异常判断模块,用于获取胸腔图像,利用所述分类网络判断所述胸腔图像是否异常;
    肺野分割指标生成模块,用于当所述胸腔图像异常时,利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标;
    心脏分割指标生成模块,用于利用所述心脏分割子网络对所述胸腔图像进行心脏分割,得到心脏分割指标;
    脊柱分割指标生成模块,用于利用所述脊柱分割子网络对所述胸腔图像进行脊柱分割,得到脊柱分割指标;
    心胸比指标生成模块,用于根据所述肺野分割指标和所述心脏分割指标计算所述胸腔图像的心胸比指标;
    肺野病灶分割指标生成模块,用于利用所述肺野病灶子网络对所述胸腔图像进行肺野病灶分割,得到肺野病灶分割指标;
    指标推送模块,用于将所述脊柱分割指标、心胸比指标及所述肺野病灶分割指标推送给用户。
  9. 一种计算机可读存储介质,包括存储数据区和存储程序区,存储数据区存储创建的数据,存储程序区存储有计算机程序;其中,所述计算机程序被处理器执行时实现如下步骤:
    获取包含分类网络和多个并行子网络的图像指标模型,其中,所述多个并行子网络包括肺野分割子网络、心脏分割子网络、脊柱分割子网络和肺野病灶子网络;
    获取胸腔图像,利用所述分类网络判断所述胸腔图像是否异常;
    当所述胸腔图像异常时,利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标;
    利用所述心脏分割子网络对所述胸腔图像进行心脏分割,得到心脏分割指标;
    利用所述脊柱分割子网络对所述胸腔图像进行脊柱分割,得到脊柱分割指标;
    根据所述肺野分割指标和所述心脏分割指标计算所述胸腔图像的心胸比指标;
    利用所述肺野病灶子网络对所述胸腔图像进行肺野病灶分割,得到肺野病灶分割指标;
    将所述脊柱分割指标、心胸比指标及所述肺野病灶分割指标推送给用户。
  10. 如权利要求9所述的计算机可读存储介质,其中,所述利用所述分类网络判断所述胸腔图像是否异常,包括:
    利用所述分类网络中前向卷积核对所述胸腔图像进行卷积,得到前向特征图;
    利用所述分类网络中后向卷积核对所述中间卷积图像进行卷积,得到后向特征图;
    对获得的前向特征图和后向特征图进行特征融合,得到融合特征图;
    利用图像分类算法对所述融合特征图进行图像分类,得到所述胸腔图像是否为异常的分类结果。
  11. 如权利要求9所述的计算机可读存储介质,其中,所述利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标,包括:
    利用所述肺野分割子网络对所述胸腔图像进行肺野下采样,得到肺野特征;
    将所述肺野特征进行第一阈值倍数的上采样,得到中间特征图;
    将所述中间特征图进行第二阈值倍数的上采样,得到肺野复原图像;
    利用第一激活函数计算所述肺野复原图像中各像素点属于预设第一类别的第一类别概率;
    根据所述第一类别概率对所述肺野复原图像进行分割处理,得到肺野分割指标。
  12. 如权利要求9所述的计算机可读存储介质,其中,所述利用所述脊柱分割子网络对所述胸腔图像进行脊柱分割,得到脊柱分割指标,包括:
    利用所述脊柱分割子网络对所述胸腔图进行脊柱卷积处理,得到卷积图像;
    对所述卷积图像进行池化处理,得到池化图像;
    对所述池化图像进行全连接处理,得到全连接特征;
    利用图像分割算法对所述全连接特征进行分割,得到脊柱分割指标。
  13. 如权利要求9至12中任一项所述的计算机可读存储介质,其中,所述利用所述肺野病灶子网络对所述胸腔图像进行肺野病灶分割,得到肺野病灶分割指标,包括:
    将所述胸腔图像输入至所述肺野病灶子网络进行肺野病灶特征提取,得到肺野病灶特征图像;
    利用标签激活函数计算所述肺野病灶特征图像的标签为预设病灶类别标签的概率值;
    根据所述概率值生成所述肺野病灶特征图像对应的肺野病灶分割指标。
  14. 一种医学图像指标生成方法,其中,所述方法包括:
    获取包含分类网络和多个并行子网络的图像指标模型,其中,所述多个并行子网络包括肺野分割子网络、心脏分割子网络、脊柱分割子网络和肺野病灶子网络;
    获取胸腔图像,利用所述分类网络判断所述胸腔图像是否异常;
    当所述胸腔图像异常时,利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标;
    利用所述心脏分割子网络对所述胸腔图像进行心脏分割,得到心脏分割指标;
    利用所述脊柱分割子网络对所述胸腔图像进行脊柱分割,得到脊柱分割指标;
    根据所述肺野分割指标和所述心脏分割指标计算所述胸腔图像的心胸比指标;
    利用所述肺野病灶子网络对所述胸腔图像进行肺野病灶分割,得到肺野病灶分割指标;
    将所述脊柱分割指标、心胸比指标及所述肺野病灶分割指标推送给用户。
  15. 如权利要求14所述的医学图像指标生成方法,其中,所述利用所述分类网络判断所述胸腔图像是否异常,包括:
    利用所述分类网络中前向卷积核对所述胸腔图像进行卷积,得到前向特征图;
    利用所述分类网络中后向卷积核对所述中间卷积图像进行卷积,得到后向特征图;
    对获得的前向特征图和后向特征图进行特征融合,得到融合特征图;
    利用图像分类算法对所述融合特征图进行图像分类,得到所述胸腔图像是否为异常的分类结果。
  16. 如权利要求14所述的医学图像指标生成方法,其中,所述利用所述肺野分割子网络对所述胸腔图像进行肺野分割,得到肺野分割指标,包括:
    利用所述肺野分割子网络对所述胸腔图像进行肺野下采样,得到肺野特征;
    将所述肺野特征进行第一阈值倍数的上采样,得到中间特征图;
    将所述中间特征图进行第二阈值倍数的上采样,得到肺野复原图像;
    利用第一激活函数计算所述肺野复原图像中各像素点属于预设第一类别的第一类别概率;
    根据所述第一类别概率对所述肺野复原图像进行分割处理,得到肺野分割指标。
  17. 如权利要求14所述的医学图像指标生成方法,其中,所述利用所述脊柱分割子网络对所述胸腔图像进行脊柱分割,得到脊柱分割指标,包括:
    利用所述脊柱分割子网络对所述胸腔图进行脊柱卷积处理,得到卷积图像;
    对所述卷积图像进行池化处理,得到池化图像;
    对所述池化图像进行全连接处理,得到全连接特征;
    利用图像分割算法对所述全连接特征进行分割,得到脊柱分割指标。
  18. 如权利要求14至17中任一项所述的医学图像指标生成方法,其中,所述利用所述肺野病灶子网络对所述胸腔图像进行肺野病灶分割,得到肺野病灶分割指标,包括:
    将所述胸腔图像输入至所述肺野病灶子网络进行肺野病灶特征提取,得到肺野病灶特征图像;
    利用标签激活函数计算所述肺野病灶特征图像的标签为预设病灶类别标签的概率值;
    根据所述概率值生成所述肺野病灶特征图像对应的肺野病灶分割指标。
  19. 如权利要求14至17中任一项所述的医学图像指标生成方法,其中,所述根据所述肺野分割指标和心脏分割指标计算所述胸腔图像的心胸比指标,包括:
    利用如下指标算法根据所述肺野分割指标、心脏分割指标计算所述胸腔图像的心胸比指标:
    F=(CD+EF)/AB
    其中,F为所述心胸比指标,CD为所述心脏分割指标中心脏最大左横径,EF为所述心脏分割指标中心脏最大右横径,AB为所述肺野分割指标中肺野最大横径。
  20. 如权利要求14所述的医学图像指标生成方法,其中,所述分类网络为VGG网络,所述VGG网络中包含多个级联的卷积核。
PCT/CN2020/131994 2020-10-20 2020-11-27 电子设备、医学图像指标生成方法、装置及存储介质 Ceased WO2021189914A1 (zh)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN202011127501.8 2020-10-20
CN202011127501 2020-10-20
CN202011327450.3A CN112308853A (zh) 2020-10-20 2020-11-24 电子设备、医学图像指标生成方法、装置及存储介质
CN202011327450.3 2020-11-24

Publications (1)

Publication Number Publication Date
WO2021189914A1 true WO2021189914A1 (zh) 2021-09-30

Family

ID=74335544

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/131994 Ceased WO2021189914A1 (zh) 2020-10-20 2020-11-27 电子设备、医学图像指标生成方法、装置及存储介质

Country Status (2)

Country Link
CN (1) CN112308853A (zh)
WO (1) WO2021189914A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115511798A (zh) * 2022-09-08 2022-12-23 中国人民解放军陆军军医大学第一附属医院 一种基于人工智能技术的肺炎分类方法及装置
CN117274185A (zh) * 2023-09-19 2023-12-22 阿里巴巴达摩院(杭州)科技有限公司 检测方法、检测模型产品、电子设备及计算机存储介质

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113555089A (zh) * 2021-07-14 2021-10-26 江苏宏创信息科技有限公司 一种应用于临床影像的人工智能医学影像质控方法
CN113850753B (zh) * 2021-08-17 2023-09-01 苏州鸿熙融合智能医疗科技有限公司 医学图像信息计算方法、装置、边缘计算设备和存储介质
CN115019049B (zh) * 2022-08-08 2022-12-16 四川大学 基于深度神经网络的骨显像骨病灶分割方法、系统及设备

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170256056A1 (en) * 2014-09-09 2017-09-07 Icometrix Nv Method and System for Analyzing Image Data
CN109859168A (zh) * 2018-12-28 2019-06-07 上海联影智能医疗科技有限公司 一种x射线胸片图像质量确定方法及装置
CN111180046A (zh) * 2018-11-13 2020-05-19 西门子医疗有限公司 确定用于处理图像的处理序列
CN111724347A (zh) * 2020-05-22 2020-09-29 广州市康软信息科技有限公司 一种基于深度学习的胸片图像异常检测方法及系统

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018205922A1 (en) * 2017-05-08 2018-11-15 Suzhou Complexis Medical Inc. Methods and systems for pulmonary function test based on diagnostic medical imaging and machine learning
CN110766682A (zh) * 2019-10-29 2020-02-07 慧影医疗科技(北京)有限公司 肺结核定位筛查装置及计算机设备
CN111325739B (zh) * 2020-02-28 2020-12-29 推想医疗科技股份有限公司 肺部病灶检测的方法及装置,和图像检测模型的训练方法
CN111428709B (zh) * 2020-03-13 2023-10-24 平安科技(深圳)有限公司 图像处理方法、装置、计算机设备和存储介质
CN111754453A (zh) * 2020-05-11 2020-10-09 佛山市第四人民医院(佛山市结核病防治所) 基于胸透图像的肺结核检测方法、系统和存储介质

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170256056A1 (en) * 2014-09-09 2017-09-07 Icometrix Nv Method and System for Analyzing Image Data
CN111180046A (zh) * 2018-11-13 2020-05-19 西门子医疗有限公司 确定用于处理图像的处理序列
CN109859168A (zh) * 2018-12-28 2019-06-07 上海联影智能医疗科技有限公司 一种x射线胸片图像质量确定方法及装置
CN111724347A (zh) * 2020-05-22 2020-09-29 广州市康软信息科技有限公司 一种基于深度学习的胸片图像异常检测方法及系统

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115511798A (zh) * 2022-09-08 2022-12-23 中国人民解放军陆军军医大学第一附属医院 一种基于人工智能技术的肺炎分类方法及装置
CN117274185A (zh) * 2023-09-19 2023-12-22 阿里巴巴达摩院(杭州)科技有限公司 检测方法、检测模型产品、电子设备及计算机存储介质
CN117274185B (zh) * 2023-09-19 2024-05-07 阿里巴巴达摩院(杭州)科技有限公司 检测方法、检测模型产品、电子设备及计算机存储介质

Also Published As

Publication number Publication date
CN112308853A (zh) 2021-02-02

Similar Documents

Publication Publication Date Title
WO2021189914A1 (zh) 电子设备、医学图像指标生成方法、装置及存储介质
CN111932482B (zh) 图像中目标物的检测方法、装置、电子设备及存储介质
CN111915609B (zh) 病灶检测分析方法、装置、电子设备及计算机存储介质
Warin et al. Maxillofacial fracture detection and classification in computed tomography images using convolutional neural network-based models
WO2022121156A1 (zh) 图像中目标物检测方法、装置、电子设备及可读存储介质
WO2022247007A1 (zh) 医学图像分级方法、装置、电子设备及可读存储介质
WO2021189855A1 (zh) 基于ct序列的图像识别方法、装置、电子设备及介质
Seo et al. A deep learning algorithm for automated measurement of vertebral body compression from X-ray images
CN111696084B (zh) 细胞图像分割方法、装置、电子设备及可读存储介质
JP6727176B2 (ja) 学習支援装置、学習支援装置の作動方法、学習支援プログラム、学習支援システム、および端末装置
Guler Ayyildiz et al. Comparison of deep learning methods for the radiographic detection of patients with different periodontitis stages
WO2021189913A1 (zh) 图像中目标物的分割方法、装置、电子设备及存储介质
CN111933274B (zh) 疾病分类诊断方法、装置、电子设备及存储介质
WO2020110774A1 (ja) 画像処理装置、画像処理方法、及びプログラム
CN110969623B (zh) 一种肺部ct多征象自动检测方法、系统、终端及存储介质
CN111862096A (zh) 图像分割方法、装置、电子设备及存储介质
WO2019146358A1 (ja) 学習システム、方法及びプログラム
CN114511569B (zh) 基于肿瘤标志物的医学图像识别方法、装置、设备及介质
Yasin et al. Automatic mandibular third molar and mandibular canal relationship determination based on deep learning models for preoperative risk reduction
CN112017763B (zh) 医疗影像数据传输方法、装置、设备及介质
CN115206512B (zh) 基于物联网的医院信息管理方法及装置
CN113284047A (zh) 基于多重特征的目标物分割方法、装置、设备及存储介质
CN115578370B (zh) 一种基于脑影像的代谢区域异常检测方法及装置
CN111462203A (zh) Dr病灶演化分析装置和方法
CN115861283A (zh) 医学影像的分析方法、装置、设备及存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20927873

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20927873

Country of ref document: EP

Kind code of ref document: A1