WO2023071114A1 - 基于人工智能的结石图像识别方法、装置及设备 - Google Patents

基于人工智能的结石图像识别方法、装置及设备 Download PDF

Info

Publication number
WO2023071114A1
WO2023071114A1 PCT/CN2022/089994 CN2022089994W WO2023071114A1 WO 2023071114 A1 WO2023071114 A1 WO 2023071114A1 CN 2022089994 W CN2022089994 W CN 2022089994W WO 2023071114 A1 WO2023071114 A1 WO 2023071114A1
Authority
WO
WIPO (PCT)
Prior art keywords
stone
image
contour
calculus
bounding box
Prior art date
Application number
PCT/CN2022/089994
Other languages
English (en)
French (fr)
Inventor
袁超
徐介夫
Original Assignee
平安科技(深圳)有限公司
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 平安科技(深圳)有限公司 filed Critical 平安科技(深圳)有限公司
Publication of WO2023071114A1 publication Critical patent/WO2023071114A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR 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; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/10Segmentation; Edge detection
    • G06T7/181Segmentation; Edge detection involving edge growing; involving edge linking
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/60Analysis of geometric attributes
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/70Determining position or orientation of objects or cameras
    • GPHYSICS
    • G16INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16HHEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00ICT specially adapted for the handling or processing of medical images
    • G16H30/20ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/10Image acquisition modality
    • G06T2207/10132Ultrasound image
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/20Special algorithmic details
    • G06T2207/20081Training; Learning

Definitions

  • the present application belongs to the field of artificial intelligence, and in particular relates to a stone image recognition method, device and equipment based on artificial intelligence.
  • B-ultrasound examination items When people undergo physical examinations in hospitals, B-ultrasound examination items are often included. For example, B-ultrasound is used to check people's stones, including kidney stones and gallstones. Through the inspection of these items, people can understand their physical condition more clearly according to the inspection results, or formulate corresponding prevention and control measures to avoid further deterioration of the disease.
  • B-ultrasound examination usually an inspector, such as a doctor, uses a B-ultrasound sensor to press on the inspection site, and then marks the combination on the displayed image, including parameters such as the maximum diameter of the mark, and then saves the marked image. Due to the three-dimensional shape of the stone, the inventors found that inspectors often need to collect images at different angles, mark the collected images multiple times and then save the required images.
  • the present application proposes an artificial intelligence-based stone image recognition method, device and equipment to solve the problems of cumbersome operation and low inspection efficiency in stone detection in the prior art.
  • the first aspect of the present application provides an artificial intelligence-based calculus image recognition method, the method comprising:
  • Collect images to be identified including stones through sensing equipment;
  • a recognition report corresponding to the stone image is generated according to the multiple stone sizes at different acquisition angles.
  • the second aspect of the present application provides an artificial intelligence-based calculus image recognition device, the artificial intelligence-based calculus image recognition device includes a memory and a processor, the memory is used to store at least one computer-readable instruction, and the processing The device is used to execute the at least one computer readable instruction to achieve the following steps:
  • Collect images to be identified including stones through sensing equipment;
  • a recognition report corresponding to the stone image is generated according to the multiple stone sizes at different acquisition angles.
  • a third aspect of the present application provides a computer-readable storage medium, the computer-readable storage medium stores at least one computer-readable instruction, and when the at least one computer-readable instruction is executed by a processor, the following steps are implemented:
  • Collect images to be identified including stones through sensing equipment;
  • a recognition report corresponding to the stone image is generated according to the multiple stone sizes at different acquisition angles.
  • a fourth aspect of the present application provides an artificial intelligence-based calculus image recognition device, the device comprising:
  • the stone image acquisition unit is used to collect images to be identified including stones through the sensing device;
  • a calculus position determining unit configured to perform target detection on the image to be recognized including a calculus, and determine the position of the calculus included in the image to be recognized;
  • a stone contour recognition unit configured to recognize the contour of the stone according to the determined location of the stone
  • a stone size identification unit configured to determine the stone size in the stone image according to the outline of the stone
  • the recognition report generation unit is configured to generate a recognition report corresponding to the stone image according to the multiple stone sizes at different acquisition angles.
  • the stone image recognition method, device and equipment based on artificial intelligence described in this application improve the efficiency of stone detection.
  • Fig. 1 is a schematic diagram of the implementation process of an artificial intelligence-based calculus image recognition method provided in the embodiment of the present application;
  • Fig. 2 is a schematic flow diagram of the implementation of a method for determining the location of calculus provided by the embodiment of the present application;
  • FIG. 3 is a schematic diagram of a bounding box identified by an image to be identified provided in an embodiment of the present application
  • Fig. 4 is a schematic diagram of an artificial intelligence-based calculus image recognition device provided in an embodiment of the present application.
  • Fig. 5 is a schematic diagram of an artificial intelligence-based calculus image recognition device provided by an embodiment of the present application.
  • the embodiment of the present application proposes an artificial intelligence-based calculus image recognition method, as shown in Figure 1, the method includes:
  • an image to be recognized including stones is collected by a sensing device.
  • the sensing device may include stone detection equipment such as B-ultrasound equipment and CT equipment.
  • stone detection equipment such as B-ultrasound equipment and CT equipment.
  • multiple stone images can be collected according to different collection angles, so that the size information of the stone can be better restored.
  • the collected stone image may not be able to reflect the length information of the stone.
  • the obtained stone image can effectively reflect the length information of the stone. for accurate stone size.
  • the inspector can touch the detection site through the B-ultrasound probe, and detect the detection site from different angles, so as to obtain more comprehensive stone size information.
  • an angle sensor may be set on the B-ultrasound probe, for example, a compass or a gyroscope may be set.
  • a compass or a gyroscope may be set.
  • the detection angle required for a stone inspection operation can be set, and after the image acquisition of an acquisition angle is completed, the angle at which the outline acquisition is not completed is determined according to the preset angle for outline acquisition.
  • the prompt information of the operation angle that needs to be contour collected can be generated, so that the staff can perform the angle collection operation according to the prompt information, and avoid collecting images of repeated angles, which will affect the accuracy of the image.
  • the screen may display an image of the operating angle, text and/or voice prompts to prompt the inspector to collect images according to the next collection angle.
  • the corresponding relationship between the collection angle and the collection location can also be set, and according to the collection location corresponding to the next collection angle, the inspector is prompted to collect according to the collection location corresponding to the next collection angle.
  • the inspector is prompted to collect according to the collection location corresponding to the next collection angle. For example, in the process of gallstone detection, when the inspector completes the detection on the front of the abdomen, he can be prompted to collect images on the side of the abdomen according to a predetermined angle of inclination.
  • the image to be identified can also be stored in a node of a block chain.
  • object detection is performed on the image to be recognized, and the position of the stone included in the image to be recognized is determined.
  • object detection is performed on the image to be recognized, that is, stones included in the image to be recognized are detected. Therefore, it is convenient to analyze the size of the stone according to the detection result.
  • the image to be recognized is divided into a first predetermined number of grids.
  • the image to be recognized including stones may be evenly divided into multiple grids according to a preset grid division manner.
  • an image to be recognized including stones may be divided into S*S grids, where S is a predetermined value.
  • the divided grid is input to the pre-trained bounding box generation network, and a second predetermined number of bounding boxes and the probability that the bounding boxes belong to stones are obtained.
  • the divided grid can be input into the pre-trained bounding box generation network, and the bounding box generation network is calculated and processed to generate a second predetermined number of bounding boxes in each grid, and each grid can be obtained.
  • the probability that the bounding box belongs to a stone can include two aspects, namely, the possibility that the bounding box contains the target and the accuracy of the bounding box.
  • the probability that the bounding box contains the object is the probability that the bounding box contains a stone.
  • the probability of the bounding box is 1, and if it does not contain stones, the probability of the bounding box is 0.
  • the accuracy of the bounding box can be expressed by the intersection ratio of the predicted box and the actual box.
  • the intersection ratio is the ratio of the intersection and union of the predicted frame and the actual frame.
  • the bounding box of the stone in the image to be recognized is determined by non-maximum suppression, and the location of the stone is determined according to the identified bounding box.
  • non-maximum value suppression algorithm to determine the location of stones can be used to solve the problem that the same stone is detected multiple times. For example, the same stone belongs to multiple grids and may be detected by multiple bounding boxes. The most accurate bounding box can be determined through the non-maximum value suppression algorithm.
  • the undeleted bounding boxes select the bounding box with the second highest accuracy, and perform calculations with other undeleted bounding boxes in turn, delete the bounding boxes whose intersection ratio is greater than a certain value, until the calculation process of all bounding boxes is completed, Thus, the redundant bounding box is removed.
  • the determined bounding box it can be used as the position of the stone.
  • the position of the stone is within the bounding box, or the center point of the stone is located at the center of the bounding box.
  • a bounding box including stones included in the image to be recognized may be obtained, and the bounding box includes stones.
  • the contour of the stone is identified according to the determined location of the stone.
  • the present application After determining the position of the calculus in the calculus image, in order to facilitate the inspector to determine the size of the calculus, the present application further determines the contour of the recognized calculus.
  • binarization may be performed on the determined image within the bounding box including the calculus first, to obtain a binarized image of the image within the bounding box.
  • the image in the bounding box is binarized, and the image in the bounding box can be processed in grayscale, combined with the preset grayscale threshold, the grayscale value of the pixel in the bounding box is compared with the grayscale threshold, Determines the value of each pixel within the bounding box to be 0 or 1.
  • the intersection of the two values can be determined as a contour point, and the contour of the calculus in the bounding box can be obtained according to the determined multiple contour points.
  • optimization processing such as filtering may also be performed on the identified contours to obtain the contours of stones with higher precision.
  • the size of the stone in the stone image is determined according to the outline of the stone.
  • the center position of the stone can be determined according to the contour.
  • a line connecting any two contour points passing through the central position can be obtained.
  • the line connecting two contour points passing through the center is the diameter.
  • the distance between the calculus and the sensing device can be determined according to the signal detected by the sensing device, combined with the size of the calculus image collected by the sensing device, the ratio of the calculus image to the actual image can be determined. According to the size of the diameter of the stone in the image of the stone, combined with the ratio of the image of the stone to the actual image, the real diameter of the stone can be calculated.
  • a plurality of diameters present in the obtained stone outline can be compared to determine the largest diameter and the smallest diameter of the stone in the stone image.
  • an identification report corresponding to the image of the stone is generated according to the multiple stone sizes at different acquisition angles.
  • the shape of the stone at the detected site may be irregular, or there may be multiple stones at the detected site, in order to obtain the size information of the stone more accurately, it is necessary to collect images of the stone at different acquisition angles.
  • the detection angle necessary for the same detection object can be set.
  • inspectors inspect the object to be inspected they can collect angles one by one according to the inspection sequence for inspection.
  • the detection angle during image collection can be detected.
  • the collection angle of the inspector when collecting the current image can be obtained through the angle sensor provided in the sensing device.
  • the angle that still needs to be acquired can be determined, and the inspection personnel can be prompted to collect images at the next acquisition angle through sound, image or text prompts.
  • the clarity of the stone image at the third angle may not meet the predetermined requirements, and the inspector can be prompted by the system Acquisition of calculi images is carried out again according to the third angle.
  • correspondences between different collection angles and parts may also be set.
  • the location corresponding to the collection angle may be prompted. For example, the left side of the waist, the front of the waist, the front left of the waist, etc.
  • the first detection point, the second detection point, and the third detection point can be prompted, so that the detection personnel can quickly locate the detection site.
  • the pattern determines the next acquisition angle, so that the next acquisition angle can more accurately collect images of overlapping stones, so as to generate a more accurate stone report.
  • the bounding box of the stone can be removed, and the outline of the stone can be directly displayed in the image, so that the inspector can view the information of the stone more clearly.
  • the contour and size of the calculus in the image can also be directly displayed.
  • the displayed dimensions may include the maximum and minimum diameters of the profile. Inspection personnel can flexibly adjust the collection angle according to the displayed size information, so as to obtain more accurate collection images.
  • the obtained multiple contours can be fused to obtain a three-dimensional image of the stone, and an electronic detection report can be generated based on the obtained three-dimensional image, which is convenient for the detected object to be more accurate. To view the detection results intuitively.
  • Fig. 4 is a schematic diagram of an artificial intelligence-based calculus image recognition device provided in an embodiment of the present application. As shown in Fig. 4, the device includes:
  • Calculus image acquisition unit 401 configured to acquire an image to be recognized including calculus through a sensing device
  • a stone position determining unit 402 configured to perform target detection on the image to be recognized including stones, and determine the position of the stone included in the image to be recognized;
  • a stone contour recognition unit 403, configured to recognize the contour of the stone according to the determined location of the stone
  • a stone size identification unit 404 configured to determine the stone size in the stone image according to the outline of the stone
  • the recognition report generating unit 405 is configured to generate a recognition report corresponding to the stone image according to the multiple stone sizes at different acquisition angles.
  • the artificial intelligence-based stone image recognition device shown in FIG. 4 corresponds to the artificial intelligence-based stone image recognition method shown in FIG. 1 .
  • Fig. 5 is a schematic diagram of an artificial intelligence-based calculus image recognition device provided by an embodiment of the present application.
  • the stone image recognition device 5 based on artificial intelligence of this embodiment includes: a processor 50, a memory 51, and a computer program 52 stored in the memory 51 and operable on the processor 50, For example, an artificial intelligence-based stone image recognition program.
  • the processor 50 executes the computer program 52, the steps in the above-mentioned embodiments of the artificial intelligence-based stone image recognition method are realized.
  • the processor 50 executes the computer program 52, the functions of the modules/units in the above-mentioned device embodiments are realized.
  • the computer program 52 can be divided into one or more modules/units, and the one or more modules/units are stored in the memory 51 and executed by the processor 50 to complete this application.
  • the one or more modules/units may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program 52 in the artificial intelligence-based stone image recognition device 5 .
  • the stone image recognition device based on artificial intelligence may include, but not limited to, a processor 50 and a memory 51 .
  • Fig. 5 is only an example of the stone image recognition device 5 based on artificial intelligence, and does not constitute a limitation to the stone image recognition device 5 based on artificial intelligence, and may include more or less components, or a combination of certain components, or different components, for example, the artificial intelligence-based calculus image recognition device may also include input and output devices, network access devices, buses, and the like.
  • the so-called processor 50 can be a central processing unit (Central Processing Unit, CPU), and can also be other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), Field-Programmable Gate Array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
  • a general-purpose processor may be a microprocessor, or the processor may be any conventional processor, or the like.
  • the memory 51 may be an internal storage unit of the artificial intelligence-based calculus image recognition device 5 , such as a hard disk or memory of the artificial intelligence-based calculus image recognition device 5 .
  • the memory 51 can also be an external storage device of the artificial intelligence-based calculus image recognition device 5, such as a plug-in hard disk equipped on the artificial intelligence-based calculus image recognition device 5, a smart memory card (Smart Media Card) , SMC), Secure Digital (Secure Digital, SD) card, Flash Card (Flash Card), etc.
  • the memory 51 can also include both the internal storage unit of the artificial intelligence-based stone image recognition device 5 and an external storage device.
  • the memory 51 is used to store the computer program and other programs and data required by the artificial intelligence-based calculus image recognition device.
  • the memory 51 can also be used to temporarily store data that has been output or will be output.
  • the disclosed apparatus/terminal device and method may be implemented in other ways.
  • the device/terminal device embodiments described above are only illustrative.
  • the division of the modules or units is only a logical function division.
  • the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be in electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.
  • each functional unit in each embodiment of the present application may be integrated into one processing unit, each unit may exist separately physically, or two or more units may be integrated into one unit.
  • the above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.
  • the integrated module/unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments in this application can also be completed by hardware related to computer program instructions.
  • the computer program can be stored in a computer-readable storage medium.
  • the computer program When executed by a processor, the steps in the above-mentioned various method embodiments can be realized.
  • the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file or some intermediate form.
  • the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a removable hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM, Read-Only Memory) , Random Access Memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal and software distribution medium, etc.
  • ROM Read-Only Memory
  • RAM Random Access Memory
  • electrical carrier signal telecommunication signal and software distribution medium, etc.
  • the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable Excluding electrical carrier signals and telecommunication signals.
  • the computer-readable storage medium may be non-volatile or volatile.
  • the computer-readable storage medium may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required by at least one function, etc.; The data created using the node, etc.
  • Blockchain essentially a decentralized database, is a series of data blocks associated with each other using cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify its Validity of information (anti-counterfeiting) and 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)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Medical Informatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Quality & Reliability (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Public Health (AREA)
  • Geometry (AREA)
  • Apparatus For Radiation Diagnosis (AREA)

Abstract

本申请属于人工智能领域,提出了一种基于人工智能的结石图像识别方法、装置及设备。该方法包括:通过传感设备采集包括结石的待识别图像;对所述包括结石的待识别图像进行目标检测,确定所述待识别图像中所包括的结石的位置;根据所确定的结石的位置,识别所述结石的轮廓;根据所述结石的轮廓,确定所述结石图像中的结石尺寸;根据不同采集角度的多个所述结石尺寸,生成所述结石图像对应的识别报告。由于该方法可以自动对图像中的结石位置、轮廓进行识别检测,并自动记录不同采集角度的结石尺寸,从而可以大大的降低检测人员的操作难度,提高结石检测的效率。本申请还涉及区块链技术,待识别图像存储于区块链节点中。

Description

基于人工智能的结石图像识别方法、装置及设备
本申请要求于2021年10月29日提交中国专利局,申请号为202111270917.X申请名称为“基于人工智能的结石图像识别方法、装置及设备”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
本申请属于人工智能领域,尤其涉及基于人工智能的结石图像识别方法、装置及设备。
背景技术
当人们在医院进行身体检查时,经常会包括B超检查项目。比如,通过B超检查人们的结石,包括如肾结石、胆结石等。通过对这些项目的检查,使得人们能够根据检查结果,更为清楚的了解自己的身体状况,或者制定相应的防治措施,避免病情进一步恶化。
在进行B超检查时,通常由检查人员,比如医生使用B超传感器在检查部位按压,然后在显示的图像上对结合进行标记,包括如标记最大直径等参数,然后保存所标记的图像。由于结石是立体形状,发明人发现检查人员往往需要在不同角度采集图像,对采集的图像多次标记后再保存所需要的图像,整个操作过程较为麻烦,使得B超检查效率较低。
发明内容
本申请提出一种基于人工智能的结石图像识别方法、装置及设备,以解决现有技术中在进行结石检测时,操作过程较为麻烦,检查效率较低的问题。
本申请的第一方面提供一种基于人工智能的结石图像识别方法,所述方法包括:
通过传感设备采集包括结石的待识别图像;
对所述待识别图像进行目标检测,确定所述待识别图像中所包括的结石的位置;
根据所确定的结石的位置,识别所述结石的轮廓;
根据所述结石的轮廓,确定所述结石图像中的结石尺寸;
根据不同采集角度的多个所述结石尺寸,生成所述结石图像对应的识别报告。
本申请的第二方面提供一种基于人工智能的结石图像识别设备,所述基于人工智能的结石图像识别设备包括存储器及处理器,所述存储器用于存储至少一个计算机可读指令,所述处理器用于执行所述至少一个计算机可读指令以实现以下步骤:
通过传感设备采集包括结石的待识别图像;
对所述待识别图像进行目标检测,确定所述待识别图像中所包括的结石的位置;
根据所确定的结石的位置,识别所述结石的轮廓;
根据所述结石的轮廓,确定所述结石图像中的结石尺寸;
根据不同采集角度的多个所述结石尺寸,生成所述结石图像对应的识别报告。
本申请的第三方面提供一种计算机可读存储介质,所述计算机可读存储介质存储有至少一个计算机可读指令,所述至少一个计算机可读指令被处理器执行时实现以下步骤:
通过传感设备采集包括结石的待识别图像;
对所述待识别图像进行目标检测,确定所述待识别图像中所包括的结石的位置;
根据所确定的结石的位置,识别所述结石的轮廓;
根据所述结石的轮廓,确定所述结石图像中的结石尺寸;
根据不同采集角度的多个所述结石尺寸,生成所述结石图像对应的识别报告。
本申请的第四方面提供一种基于人工智能的结石图像识别装置,所述装置包括:
结石图像采集单元,用于通过传感设备采集包括结石的待识别图像;
结石位置确定单元,用于对所述包括结石的待识别图像进行目标检测,确定所述待识别图像中所包括的结石的位置;
结石轮廓识别单元,用于根据所确定的结石的位置,识别所述结石的轮廓;
结石尺寸识别单元,用于根据所述结石的轮廓,确定所述结石图像中的结石尺寸;
识别报告生成单元,用于根据不同采集角度的多个所述结石尺寸,生成所述结石图像对应的识别报告。
本申请所述的基于人工智能的结石图像识别方法、装置及设备,提高结石检测的效率。
附图说明
图1是本申请实施例提供的一种基于人工智能的结石图像识别方法的实现流程示意图;
图2是本申请实施例提供的一种确定结石位置方法的实现流程示意图;
图3是本申请实施例提供的待识别图像所识别的边界框示意图;
图4是本申请实施例提供的一种基于人工智能的结石图像识别装置的示意图;
图5是本申请实施例提供的基于人工智能的结石图像识别设备的示意图。
具体实施方式
以下描述中,为了说明而不是为了限定,提出了诸如特定系统结构、技术之类的具体细节,以便透彻理解本申请实施例。然而,本领域的技术人员应当清楚,在没有这些具体细节的其它实施例中也可以实现本申请。在其它情况中,省略对众所周知的系统、装置、电路以及方法的详细说明,以免不必要的细节妨碍本申请的描述。
为了说明本申请所述的技术方案,下面通过具体实施例来进行说明。
目前,在检查机构,包括如医院或体检中心进行结石检测时,通常需要由检测人员根据自己的经验,将检测设备对准结石位置,通过显示设备显示当前所采集的图像。检测人员根据图像中所显示的结石的大小,估计结石的尺寸。并且,由于检测设备与结石的距离及检测角度的不同,均会影响到结石图像中所显示的结石的尺寸,不利于检测人员准确的评估结石尺寸的同时,也不利于提高结石检查操作的效率。
基于此,本申请实施例提出了一种基于人工智能的结石图像识别方法,如图1所示,该方法包括:
在S101中,通过传感设备采集包括结石的待识别图像。
本申请实施例中,所述传感设备可以包括如B超设备、CT设备等结石检测设备。通过传感设备进行结石图像采集时,可以按照不同的采集角度,采集多个结石图像,从而能够更好的还原结石的尺寸信息。比如,结石的长度方向与采集角度的方向一致时,所采集的结石图像可能无法反应结石的长度信息,通过切换采集角度,使得所获得的结石图像能够有效的反应结石的长度信息,从而得到更为准确的结石尺寸。
当所述传感设备为B超设备时,检测人员可以通过B超探头接触检测部位,通过对该检测部位进行不同角度的检测,从而得到更为全面的结石尺寸信息。
在可能的实现方式中,可在B超探头设置角度传感器,比如可以设置罗盘或陀螺仪等。当检测人员通过B超探头进行检测时,在采集到结石图像时,读取角度传感器的角度数据,记录结石图像与采集角度的对应关系,从而便于后续对结石的尺寸计算和三维重建处理。
在可能的实现方式中,可以设定一次结石检查操作所需要的检测角度,当完成一个采集角度的图像采集后,根据预先设定的需要进行轮廓采集的角度,确定未完成轮廓采集的角度。 可以根据未完成轮廓采集的角度,生成需要进行轮廓采集的操作角度的提示信息,以便工作人员可以根据该提示信息进行角度采集操作,避免采集到重复角度的图像,影响图像的准确性。比如,可以通过屏幕显示操作角度的图像、文字和/或语音提示的方式,提示检测人员按照下一个采集角度进行图像采集。
或者,也可以设定采集角度与采集部位的对应关系,根据下一个采集角度所对应的采集部位,提示检测人员按照下一个采集角度对应的采集部位进行采集。比如,在胆结石检测过程,当检测人员在腹部正面检测完成后,可以提示在腹部侧面,按照预定的倾斜角度进行图像采集。
需要强调的是,为进一步保证上述待识别图像的私密和安全性,上述待识别图像还可以存储于一区块链的节点中。
在S102中,对所述待识别图像进行目标检测,确定所述待识别图像中所包括的结石的位置。
在本申请实施例中,对所述待识别图像进行目标检测,即检测待识别图像中包括的结石。从而便于根据检测结果对结石的大小进行分析。
其中,对待识别的图像进行目标检测时,可以如图2所示,包括:
在S201中,将所述待识别图像划分为第一预定数量的栅格。
可以按照预设的网格划分方式,将包括结石的待识别图像均匀划分为多个栅格。比如,可以将包括结石的待识别图像划分为S*S个网格,其中,S为预定的数值。
在S202中,将所划分的栅格输入到预先训练完成的边界框生成网络,获取第二预定数量的边界框,以及所述边界框属于结石的概率。
在进行结石检测时,可以将所划分的栅格输入到预先训练完成的边界框生成网络,通过边界框生成网络计算处理,在每个栅格中生成第二预定数量的边界框,以及得到每个边界框所框选的对象属于结石的概率。
其中,边界框属于结石的概率,也即边界框的置信度,可以包含两个方面,分别为边界框包含目标的可能性和边界框的准确度。边界框包含目标的可能性,即为边界框包含结石的可能性大小。当边界框中包含结石时,那该边界框的可能性为1,如果不包含结石,则边界框的可能性为0。边界框的准确度可以通过预测框与实际框的交并比来表示。交并比即预测框与实际框的交集与并集的比值。
在S203中,根据所述边界框属于结石的概率,通过非极大值抑制确定所述待识别图像中的结石的边界框,根据所识别的边界框确定所述结石的位置。
采用非极大值抑制算法确定结石的位置,可以用于解决同一个结石被多次检测的问题。比如,同一个结石属于多个栅格中,可能被多个边界框检测,通过非极大值抑制算法可以确定最准确的边界框。
采用非极大值抑制算法进行边界框选择时,可以先根据边界框的准确度,对边界框进行排序,选择出准确度最大的第一边界框,然后计算该第一边界框与其它边界框的交并比,如果计算到第一边界框与其它某个边界框,比如第二边界框的交并比大于一定值,则将该第二边界框剔除,重复上述过程,直到处理完成所有的边界框与第一边界框的计算检测。在未删除的边界框中,选择准确度第二大的边界框,依次与其它未删除的边界框进行计算,删除交并比大于一定值的边界框,直到完成所有的边界框的计算处理,从而将多余的边界框删除。根据所确定的边界框,即可作为结石的位置。比如,结石的位置处于边界框内,或者结石的中心点位于边界框的中心位置。如图3所示,通过非极大值抑制处理后,可以得到包括结石的待识别图像中所包括的结石的边界框,在所述边界框中包括结石。
在S103中,根据所确定的结石的位置,识别所述结石的轮廓。
在确定了结石图像中的结石的位置后,为了便于检测人员确定结石的大小,本申请进一步确定所识别的结石的轮廓。
在识别结石的轮廓时,可以先对所确定的包括结石的边界框内的图像进行二值化处理,得到边界框内图像的二值化图像。
对边界框内的图像进行二值化处理,可以将边界框内的图像进行灰度处理,结合预先设定的灰度阈值,将边界框内的像素的灰度值与灰度阈值进行比较,确定边界框内的每个像素的值为0或1。
在对边界框内图像的像素进行二值化处理后,可以确定二值相交处为轮廓点,根据所确定的多个轮廓点,即可得到边界框内的结石的轮廓。
当然,在可能的实现方式中,还可以对所识别的轮廓进行过滤等优化处理,得到精度更高的结石的轮廓。
在S104中,根据所述结石的轮廓,确定所述结石图像中的结石尺寸。
在得到边界框内的结石的轮廓后,可以根据轮廓确定结石的中心位置。根据所确定的结石的中心位置,即可得到通过所述中心位置的任意两个轮廓点的连线。为了表述方便,我们称通过中心位置的两个轮廓点的连线为直径。可以根据传感设备所探测的信号,确定结石与传感设备之间的距离,结合传感设备所采集的结石图像的大小,即可确定结石图像与实际图像的比例大小。根据结石图像中的结石的直径的大小,结合结石图像与实际图像的比例,即可计算得到结石的真实直径。
由于结石的形状通常为不规则形状,因此,通过结石的中心点的直径中,会存在长度不同的多个直径。可以将所得到的结石轮廓中存在的多个直径进行比较,确定结石图像中的结石的最大直径和最小直径。
在S105中,根据不同采集角度的多个所述结石尺寸,生成所述结石图像对应的识别报告。
由于被检测部位的结石的形状可能不规则,或者被检测部位可能存在多个结石,因此,为了更为准确的得到结石的尺寸信息,需要在不同采集角度进行结石图像的采集。
在一般情况下,可以设定同一检测对象所必需的检测角度。当检测人员对检测对象进行检测时,可以按照检测顺序逐个采集角度进行检测。为了减少检测人员失误的几率,简化检测过程,在检测人员进行图像采集时,可以检测图像采集时的检测角度。比如,可以通过传感设备中设置的角度传感器,获取检测人员在采集当前图像时的采集角度。当已使用的采集角度与必须的采集角度进行比较,即可确定还需要进行采集的角度,可以通过声音、图像或文字提示的方式,提示检测人员进行下一个采集角度的图像采集。
比如,检测人员根据第一角度、第二角度和第三角度进行结石图像的采集后,由于操作失误,可能第三角度的结石图像的清晰度未能达到预定的要求,可以由系统提示检测人员重新按照第三角度进行结石图像的采集。
在可能的实现方式中,还可以设定不同采集角度与部位的对应关系。当需要提示检测人员进行所需要的采集角度进行检测时,可以提示采集角度所对应的部位。比如腰部左方、腰部前方、腰部左前方等。或者还可以根据预设定的检测部位的标识,通过提示第一检测点、第二检测点、第三检测点的方式,以使得检测人员能够快速的定位需要检测的部位。
在可能的实现方式中,还可以根据检测到的结石的距离的大小,确定是否出现重叠的结石。比如,可以在检测同一个边界框中的结石与传感设备之间的距离的差值大于预定的距离阈值时,则认为该边界框中的结石为存在遮挡或重叠的结石,可以根据重叠的样式,确定下一个采集角度,使得下一个采集角度能够更为准确的对重叠的结石进行图像的采集,以便生成更为准确的结石报告。
比如,当检测到边界框中存在重叠的结石,且在当前的图像中,被遮挡的结石的相邻位置是否存在其它结石,根据不存在其它结石的方位,确定下一个采集角度,从而能够对被遮挡的结石进行图像采集时,得到更为准确的反应结石尺寸的图像。
在本申请实施例中,通过结石的轮廓确定结石的尺寸后,可以将结石的边界框去除,直 接在图像中显示结石的轮廓,从而便于检测人员更为清晰的查看结石信息。在可能的实现方式中,还可以直接显示图像中的结石的轮廓和尺寸。所显示的尺寸可以包括轮廓的最大直径和最小直径。检测人员可以根据显示的尺寸信息,灵活的对采集角度进行调整,从而得到更为准确的采集图像。
在可能的实现方式中,通过不同角度采集得到结石的轮廓后,可以将所得到的多个轮廓进行融合,得到结石的三维图像,根据所得到的三维图像生成电子检测报告,便于被检测对象更为直观的查看检测结果。
应理解,上述实施例中各步骤的序号的大小并不意味着执行顺序的先后,各过程的执行顺序应以其功能和内在逻辑确定,而不应对本申请实施例的实施过程构成任何限定。
图4为本申请实施例提供的一种基于人工智能的结石图像识别装置的示意图,如图4所示,该装置包括:
结石图像采集单元401,用于通过传感设备采集包括结石的待识别图像;
结石位置确定单元402,用于对所述包括结石的待识别图像进行目标检测,确定所述待识别图像中所包括的结石的位置;
结石轮廓识别单元403,用于根据所确定的结石的位置,识别所述结石的轮廓;
结石尺寸识别单元404,用于根据所述结石的轮廓,确定所述结石图像中的结石尺寸;
识别报告生成单元405,用于根据不同采集角度的多个所述结石尺寸,生成所述结石图像对应的识别报告。
图4所示的基于人工智能的结石图像识别装置,与图1所示的基于人工智能的结石图像识别方法对应。
图5是本申请一实施例提供的基于人工智能的结石图像识别设备的示意图。如图5所示,该实施例的基于人工智能的结石图像识别设备5包括:处理器50、存储器51以及存储在所述存储器51中并可在所述处理器50上运行的计算机程序52,例如基于人工智能的结石图像识别程序。所述处理器50执行所述计算机程序52时实现上述各个基于人工智能的结石图像识别方法实施例中的步骤。或者,所述处理器50执行所述计算机程序52时实现上述各装置实施例中各模块/单元的功能。
示例性的,所述计算机程序52可以被分割成一个或多个模块/单元,所述一个或者多个模块/单元被存储在所述存储器51中,并由所述处理器50执行,以完成本申请。所述一个或多个模块/单元可以是能够完成特定功能的一系列计算机程序指令段,该指令段用于描述所述计算机程序52在所述基于人工智能的结石图像识别设备5中的执行过程。
所述基于人工智能的结石图像识别设备可包括,但不仅限于,处理器50、存储器51。本领域技术人员可以理解,图5仅仅是基于人工智能的结石图像识别设备5的示例,并不构成对基于人工智能的结石图像识别设备5的限定,可以包括比图示更多或更少的部件,或者组合某些部件,或者不同的部件,例如所述基于人工智能的结石图像识别设备还可以包括输入输出设备、网络接入设备、总线等。
所称处理器50可以是中央处理单元(Central Processing Unit,CPU),还可以是其他通用处理器、数字信号处理器(Digital Signal Processor,DSP)、专用集成电路(Application Specific Integrated Circuit,ASIC)、现场可编程门阵列(Field-Programmable Gate Array,FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
所述存储器51可以是所述基于人工智能的结石图像识别设备5的内部存储单元,例如基于人工智能的结石图像识别设备5的硬盘或内存。所述存储器51也可以是所述基于人工智能的结石图像识别设备5的外部存储设备,例如所述基于人工智能的结石图像识别设备5上配备的插接式硬盘,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51还可以既包括所述基于人工智能的结石 图像识别设备5的内部存储单元也包括外部存储设备。所述存储器51用于存储所述计算机程序以及所述基于人工智能的结石图像识别设备所需的其他程序和数据。所述存储器51还可以用于暂时地存储已经输出或者将要输出的数据。
所属领域的技术人员可以清楚地了解到,为了描述的方便和简洁,仅以上述各功能单元、模块的划分进行举例说明,实际应用中,可以根据需要而将上述功能分配由不同的功能单元、模块完成,即将所述装置的内部结构划分成不同的功能单元或模块,以完成以上描述的全部或者部分功能。实施例中的各功能单元、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,也可以采用软件功能单元的形式实现。另外,各功能单元、模块的具体名称也只是为了便于相互区分,并不用于限制本申请的保护范围。上述系统中单元、模块的具体工作过程,可以参考前述方法实施例中的对应过程,在此不再赘述。
在上述实施例中,对各个实施例的描述都各有侧重,某个实施例中没有详述或记载的部分,可以参见其它实施例的相关描述。
本领域普通技术人员可以意识到,结合本文中所公开的实施例描述的各示例的单元及算法步骤,能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行,取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能,但是这种实现不应认为超出本申请的范围。
在本申请所提供的实施例中,应该理解到,所揭露的装置/终端设备和方法,可以通过其它的方式实现。例如,以上所描述的装置/终端设备实施例仅仅是示意性的,例如,所述模块或单元的划分,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,例如多个单元或组件可以结合或者可以集成到另一个系统,或一些特征可以忽略,或不执行。另一点,所显示或讨论的相互之间的耦合或直接耦合或通讯连接可以是通过一些接口,装置或单元的间接耦合或通讯连接,可以是电性,机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的,作为单元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单元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单元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,也可以采用软件功能单元的形式实现。
所述集成的模块/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。基于这样的理解,本申请实现上述实施例方法中的全部或部分流程,也可以通过计算机程序指令相关的硬件来完成,所述的计算机程序可存储于一计算机可读存储介质中,该计算机程序在被处理器执行时,可实现上述各个方法实施例的步骤。其中,所述计算机程序包括计算机程序代码,所述计算机程序代码可以为源代码形式、对象代码形式、可执行文件或某些中间形式等。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)、随机存取存储器(RAM,Random Access Memory)、电载波信号、电信信号以及软件分发介质等。需要说明的是,所述计算机可读介质包含的内容可以根据司法管辖区内立法和专利实践的要求进行适当的增减,例如在某些司法管辖区,根据立法和专利实践,计算机可读介质不包括是电载波信号和电信信号。
进一步地,所述计算机可读存储介质可以是非易失性,也可以是易失性。
进一步地,所述计算机可读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。
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。
以上所述实施例仅用以说明本申请的技术方案,而非对其限制;尽管参照前述实施例对本申请进行了详细的说明,本领域的普通技术人员应当理解:其依然可以对前述各实施例所记载的技术方案进行修改,或者对其中部分技术特征进行等同替换;而这些修改或者替换,并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围,均应包含在本申请的保护范围之内。

Claims (20)

  1. 一种基于人工智能的结石图像识别方法,其中,所述方法包括:
    通过传感设备采集包括结石的待识别图像;
    对所述待识别图像进行目标检测,确定所述待识别图像中所包括的结石的位置;
    根据所确定的结石的位置,识别所述结石的轮廓;
    根据所述结石的轮廓,确定所述结石图像中的结石尺寸;
    根据不同采集角度的多个所述结石尺寸,生成所述结石图像对应的识别报告。
  2. 根据权利要求1所述的基于人工智能的结石图像识别方法,其中,所述对所述待识别图像进行目标检测,确定所述待识别图像中所包括的结石的位置,包括:
    将所述待识别图像划分为第一预定数量的栅格;
    将所划分的栅格输入到预先训练完成的边界框生成网络,获取第二预定数量的边界框,以及所述边界框属于结石的概率;
    根据所述边界框属于结石的概率,通过非极大值抑制确定所述待识别图像中的结石的边界框,根据所识别的边界框确定所述结石的位置。
  3. 根据权利要求2所述的基于人工智能的结石图像识别方法,其中,所述根据所确定的结石的位置,识别所述结石的轮廓,包括:
    对所确定的结石的边界框内的图像进行二值化处理,得到边界框内图像的二值化图像;
    根据所述二值化图像,寻找边界框中的轮廓点,根据所述轮廓点绘制生成结石轮廓。
  4. 根据权利要求2所述的基于人工智能的结石图像识别方法,其中,在所述根据所述结石的轮廓,确定所述结石图像中的结石尺寸之后,所述方法还包括:
    去除所确定的结石的边界框,显示所述结石图像在当前的采集角度下的结石的轮廓,以及所述轮廓的尺寸。
  5. 根据权利要求1所述的基于人工智能的结石图像识别方法,其中,所述根据所述结石的轮廓,确定所述结石图像中的结石尺寸,包括:
    根据所述结石的轮廓,确定所述结石的中心位置;
    根据所述结石的中心位置确定在当前采集角度下的结石图像的多个直径,确定所述结石的最大直径和最小直径。
  6. 根据权利要求1所述的基于人工智能的结石图像识别方法,其中,所述方法还包括:
    获取传感设备在不同的采集角度所获取的结石的轮廓;
    根据不同采集角度所获取的结石的轮廓,生成所述结石的三维图像。
  7. 根据权利要求6所述的基于人工智能的结石图像识别方法,其中,所述获取传感设备在不同的采集角度所获取的结石的轮廓,包括:
    当传感设备采集到结石的第一轮廓时,通过角度传感器获取采集所述第一轮廓的第一角度;
    根据预设的需要进行轮廓采集的角度,确定当前未进行轮廓采集的角度;
    根据预先设定的角度与采集部位的对应关系,生成未进行轮廓采集的部位的提示信息,在检测到处于需要进行轮廓采集的部位时获取结石的轮廓。
  8. 一种基于人工智能的结石图像识别设备,其中,所述基于人工智能的结石图像识别设备包括存储器及处理器,所述存储器用于存储至少一个计算机可读指令,所述处理器用于执行所述至少一个计算机可读指令以实现以下步骤:
    通过传感设备采集包括结石的待识别图像;
    对所述待识别图像进行目标检测,确定所述待识别图像中所包括的结石的位置;
    根据所确定的结石的位置,识别所述结石的轮廓;
    根据所述结石的轮廓,确定所述结石图像中的结石尺寸;
    根据不同采集角度的多个所述结石尺寸,生成所述结石图像对应的识别报告。
  9. 根据权利要求8所述的基于人工智能的结石图像识别设备,其中,所述处理器执行所述至少一个计算机可读指令以实现所述对所述待识别图像进行目标检测,确定所述待识别图像中所包括的结石的位置时,具体包括:
    将所述待识别图像划分为第一预定数量的栅格;
    将所划分的栅格输入到预先训练完成的边界框生成网络,获取第二预定数量的边界框,以及所述边界框属于结石的概率;
    根据所述边界框属于结石的概率,通过非极大值抑制确定所述待识别图像中的结石的边界框,根据所识别的边界框确定所述结石的位置。
  10. 根据权利要求9所述的基于人工智能的结石图像识别设备,其中,所述处理器执行所述至少一个计算机可读指令以实现所述根据所确定的结石的位置,识别所述结石的轮廓时,具体包括:
    对所确定的结石的边界框内的图像进行二值化处理,得到边界框内图像的二值化图像;
    根据所述二值化图像,寻找边界框中的轮廓点,根据所述轮廓点绘制生成结石轮廓。
  11. 根据权利要求9所述的基于人工智能的结石图像识别设备,其中,在所述根据所述结石的轮廓,确定所述结石图像中的结石尺寸之后,所述处理器执行所述至少一个计算机可读指令还用以实现以下步骤:
    去除所确定的结石的边界框,显示所述结石图像在当前的采集角度下的结石的轮廓,以及所述轮廓的尺寸。
  12. 根据权利要求8所述的基于人工智能的结石图像识别设备,其中,所述处理器执行所述至少一个计算机可读指令以实现所述根据所述结石的轮廓,确定所述结石图像中的结石尺寸时,具体包括:
    根据所述结石的轮廓,确定所述结石的中心位置;
    根据所述结石的中心位置确定在当前采集角度下的结石图像的多个直径,确定所述结石的最大直径和最小直径。
  13. 根据权利要求8所述的基于人工智能的结石图像识别设备,其中,所述处理器执行所述至少一个计算机可读指令还用以实现以下步骤:
    获取传感设备在不同的采集角度所获取的结石的轮廓,包括:当传感设备采集到结石的第一轮廓时,通过角度传感器获取采集所述第一轮廓的第一角度;根据预设的需要进行轮廓采集的角度,确定当前未进行轮廓采集的角度;根据预先设定的角度与采集部位的对应关系,生成未进行轮廓采集的部位的提示信息,在检测到处于需要进行轮廓采集的部位时获取结石的轮廓;
    根据不同采集角度所获取的结石的轮廓,生成所述结石的三维图像。
  14. 一种计算机可读存储介质,其中,所述计算机可读存储介质存储有至少一个计算机可读指令,所述至少一个计算机可读指令被处理器执行时实现以下步骤:
    通过传感设备采集包括结石的待识别图像;
    对所述待识别图像进行目标检测,确定所述待识别图像中所包括的结石的位置;
    根据所确定的结石的位置,识别所述结石的轮廓;
    根据所述结石的轮廓,确定所述结石图像中的结石尺寸;
    根据不同采集角度的多个所述结石尺寸,生成所述结石图像对应的识别报告。
  15. 根据权利要求14所述的存储介质,其中,所述至少一个计算机可读指令被所述处理器执行以实现所述对所述待识别图像进行目标检测,确定所述待识别图像中所包括的结石的位置时,具体包括:
    将所述待识别图像划分为第一预定数量的栅格;
    将所划分的栅格输入到预先训练完成的边界框生成网络,获取第二预定数量的边界框,以及所述边界框属于结石的概率;
    根据所述边界框属于结石的概率,通过非极大值抑制确定所述待识别图像中的结石的边界框,根据所识别的边界框确定所述结石的位置。
  16. 根据权利要求15所述的存储介质,其中,所述至少一个计算机可读指令被所述处理器执行以实现所述根据所确定的结石的位置,识别所述结石的轮廓时,具体包括:
    对所确定的结石的边界框内的图像进行二值化处理,得到边界框内图像的二值化图像;
    根据所述二值化图像,寻找边界框中的轮廓点,根据所述轮廓点绘制生成结石轮廓。
  17. 根据权利要求15所述的存储介质,其中,在所述根据所述结石的轮廓,确定所述结石图像中的结石尺寸之后,所述至少一个计算机可读指令被处理器执行时还用以实现以下步骤:
    去除所确定的结石的边界框,显示所述结石图像在当前的采集角度下的结石的轮廓,以及所述轮廓的尺寸。
  18. 根据权利要求14所述的存储介质,其中,所述至少一个计算机可读指令被所述处理器执行以实现所述根据所述结石的轮廓,确定所述结石图像中的结石尺寸时,具体包括:
    根据所述结石的轮廓,确定所述结石的中心位置;
    根据所述结石的中心位置确定在当前采集角度下的结石图像的多个直径,确定所述结石的最大直径和最小直径。
  19. 根据权利要求14所述的存储介质,其中,所述至少一个计算机可读指令被处理器执行时还用以实现以下步骤:
    获取传感设备在不同的采集角度所获取的结石的轮廓,包括:当传感设备采集到结石的第一轮廓时,通过角度传感器获取采集所述第一轮廓的第一角度;根据预设的需要进行轮廓采集的角度,确定当前未进行轮廓采集的角度;根据预先设定的角度与采集部位的对应关系,生成未进行轮廓采集的部位的提示信息,在检测到处于需要进行轮廓采集的部位时获取结石的轮廓;
    根据不同采集角度所获取的结石的轮廓,生成所述结石的三维图像。
  20. 一种基于人工智能的结石图像识别装置,其中,所述装置包括:
    结石图像采集单元,用于通过传感设备采集包括结石的待识别图像;
    结石位置确定单元,用于对所述包括结石的待识别图像进行目标检测,确定所述待识别图像中所包括的结石的位置;
    结石轮廓识别单元,用于根据所确定的结石的位置,识别所述结石的轮廓;
    结石尺寸识别单元,用于根据所述结石的轮廓,确定所述结石图像中的结石尺寸;
    识别报告生成单元,用于根据不同采集角度的多个所述结石尺寸,生成所述结石图像对应的识别报告。
PCT/CN2022/089994 2021-10-29 2022-04-28 基于人工智能的结石图像识别方法、装置及设备 WO2023071114A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111270917.XA CN113781477A (zh) 2021-10-29 2021-10-29 基于人工智能的结石图像识别方法、装置及设备
CN202111270917.X 2021-10-29

Publications (1)

Publication Number Publication Date
WO2023071114A1 true WO2023071114A1 (zh) 2023-05-04

Family

ID=78873533

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/089994 WO2023071114A1 (zh) 2021-10-29 2022-04-28 基于人工智能的结石图像识别方法、装置及设备

Country Status (2)

Country Link
CN (1) CN113781477A (zh)
WO (1) WO2023071114A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113781477A (zh) * 2021-10-29 2021-12-10 平安科技(深圳)有限公司 基于人工智能的结石图像识别方法、装置及设备
CN113962992A (zh) * 2021-12-21 2022-01-21 青岛大学附属医院 基于深度学习的泌尿结石平扫ct图像识别系统及训练方法
CN117788474B (zh) * 2024-02-27 2024-05-03 陕西省人民医院(陕西省临床医学研究院) 一种基于机器视觉的泌尿结石图像识别方法

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100092085A1 (en) * 2008-10-13 2010-04-15 Xerox Corporation Content-based image harmonization
CN111380785A (zh) * 2020-03-30 2020-07-07 中南大学 岩石颗粒的二维几何特征参数获取系统及方法
CN113034426A (zh) * 2019-12-25 2021-06-25 飞依诺科技(苏州)有限公司 超声图像病灶描述方法、装置、计算机设备和存储介质
CN113112443A (zh) * 2019-12-25 2021-07-13 飞依诺科技(苏州)有限公司 超声图像病灶的分割方法、装置和计算机设备
CN113158869A (zh) * 2021-04-15 2021-07-23 深圳市优必选科技股份有限公司 图像识别方法、装置、终端设备及计算机可读存储介质
CN113781477A (zh) * 2021-10-29 2021-12-10 平安科技(深圳)有限公司 基于人工智能的结石图像识别方法、装置及设备

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113344911B (zh) * 2021-07-06 2023-05-12 北京大都正隆医疗科技有限公司 一种结石尺寸的测量方法以及测量装置

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100092085A1 (en) * 2008-10-13 2010-04-15 Xerox Corporation Content-based image harmonization
CN113034426A (zh) * 2019-12-25 2021-06-25 飞依诺科技(苏州)有限公司 超声图像病灶描述方法、装置、计算机设备和存储介质
CN113112443A (zh) * 2019-12-25 2021-07-13 飞依诺科技(苏州)有限公司 超声图像病灶的分割方法、装置和计算机设备
CN111380785A (zh) * 2020-03-30 2020-07-07 中南大学 岩石颗粒的二维几何特征参数获取系统及方法
CN113158869A (zh) * 2021-04-15 2021-07-23 深圳市优必选科技股份有限公司 图像识别方法、装置、终端设备及计算机可读存储介质
CN113781477A (zh) * 2021-10-29 2021-12-10 平安科技(深圳)有限公司 基于人工智能的结石图像识别方法、装置及设备

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
ZHAO, LIANHENG ET AL.: "A digitalized 2D particle database for statistical shape analysis and discrete modeling of rock aggregate", CONSTRUCTION AND BUILDING MATERIALS, vol. 247, 3 March 2020 (2020-03-03), pages 2 - 6, XP086142771, ISSN: 0950-0618, DOI: 10.1016/j.conbuildmat.2019.117906 *

Also Published As

Publication number Publication date
CN113781477A (zh) 2021-12-10

Similar Documents

Publication Publication Date Title
WO2023071114A1 (zh) 基于人工智能的结石图像识别方法、装置及设备
US11494902B2 (en) Systems and methods for automatic detection and quantification of pathology using dynamic feature classification
CN108133476B (zh) 一种肺结节自动检测方法及系统
CN107480677B (zh) 一种识别三维ct图像中感兴趣区域的方法及装置
JP2020507836A (ja) 重複撮像を予測した手術アイテムの追跡
CN111862044B (zh) 超声图像处理方法、装置、计算机设备和存储介质
Yang et al. A two-stage convolutional neural network for pulmonary embolism detection from CTPA images
CN111476774B (zh) 基于新型冠状病毒肺炎ct检测的智能征象识别装置
CN111696083B (zh) 一种图像处理方法、装置、电子设备及存储介质
US11756199B2 (en) Image analysis in pathology
US20190138840A1 (en) Automatic ruler detection
CN103886602A (zh) 一种基于纹理的射线图像缺陷检测方法
CN104272094A (zh) 缺陷判定装置、放射线摄像系统及缺陷判定方法
CN111372042B (zh) 故障检测方法、装置、计算机设备和存储介质
CN115115841A (zh) 一种阴影斑点图像处理分析方法及系统
KR20180045473A (ko) 이미지 분석을 이용한 흑색종 검사 시스템, 방법 및 컴퓨터 프로그램
CN110211200B (zh) 一种基于神经网络技术的牙弓线生成方法及其系统
CN108399354A (zh) 计算机视觉识别肿瘤的方法和装置
CN107767961B (zh) 面向接骨板类型判别的骨骼模板设计方法
US11728056B2 (en) Method for detecting thicknesses of coating layers of nuclear fuel particles
JP2012143387A (ja) 骨粗鬆症診断支援装置及び骨粗鬆症診断支援プログラム
CN109685798A (zh) 一种确定有效医学图像的方法及装置
WO2019058963A1 (ja) 医用画像処理装置及び医用画像処理方法及びそれに用いる処理プログラム
CN109166108B (zh) 一种ct影像肺部异常组织的自动识别方法
CN114119588A (zh) 一种训练眼底黄斑病变区域检测模型的方法、装置及系统

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: 22885028

Country of ref document: EP

Kind code of ref document: A1