WO2022048540A1 - 胃肠标记物自动识别方法及识别方法 - Google Patents

胃肠标记物自动识别方法及识别方法 Download PDF

Info

Publication number
WO2022048540A1
WO2022048540A1 PCT/CN2021/115724 CN2021115724W WO2022048540A1 WO 2022048540 A1 WO2022048540 A1 WO 2022048540A1 CN 2021115724 W CN2021115724 W CN 2021115724W WO 2022048540 A1 WO2022048540 A1 WO 2022048540A1
Authority
WO
WIPO (PCT)
Prior art keywords
gastrointestinal
marker
suspected
area
array
Prior art date
Application number
PCT/CN2021/115724
Other languages
English (en)
French (fr)
Other versions
WO2022048540A9 (zh
Inventor
侯晓华
谢小平
金玉
王廷旗
高飞
段晓东
Original Assignee
安翰科技(武汉)股份有限公司
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 安翰科技(武汉)股份有限公司 filed Critical 安翰科技(武汉)股份有限公司
Priority to EP21863604.1A priority Critical patent/EP4209992A4/en
Priority to US18/024,215 priority patent/US20230306589A1/en
Publication of WO2022048540A1 publication Critical patent/WO2022048540A1/zh
Publication of WO2022048540A9 publication Critical patent/WO2022048540A9/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
    • A61B6/12—Arrangements for detecting or locating foreign bodies
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
    • A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
    • A61B6/5211—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data
    • A61B6/5217—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data extracting a diagnostic or physiological parameter from medical diagnostic data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/136—Segmentation; Edge detection involving thresholding
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B90/00—Instruments, implements or accessories specially adapted for surgery or diagnosis and not covered by any of the groups A61B1/00 - A61B50/00, e.g. for luxation treatment or for protecting wound edges
    • A61B90/39—Markers, e.g. radio-opaque or breast lesions markers
    • A61B2090/3904—Markers, e.g. radio-opaque or breast lesions markers specially adapted for marking specified tissue
    • A61B2090/3912—Body cavities
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B90/00—Instruments, implements or accessories specially adapted for surgery or diagnosis and not covered by any of the groups A61B1/00 - A61B50/00, e.g. for luxation treatment or for protecting wound edges
    • A61B90/39—Markers, e.g. radio-opaque or breast lesions markers
    • A61B2090/3966—Radiopaque markers visible in an X-ray image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10116—X-ray image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30092—Stomach; Gastric
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30204—Marker

Definitions

  • the invention relates to the technical field of medical devices, in particular to an automatic identification method and an identification system for a gastrointestinal marker that can automatically detect the position of a gastrointestinal marker in an image.
  • Gastrointestinal marker capsule is a capsule containing 20 to 24 slices of X-ray opaque gastrointestinal markers, the capsule shell can dissolve naturally in the gastrointestinal tract, and the gastrointestinal markers are biocompatible materials that can be used for detection Gastrointestinal transit time is an important diagnostic tool for judging whether there is constipation or other gastrointestinal diseases.
  • the X-ray film taken depends on the doctor to determine the position and type of the gastrointestinal markers, which greatly increases the workload of the doctor.
  • the purpose of the present invention is to provide an automatic identification method and identification system for gastrointestinal markers, which can automatically detect the position of the gastrointestinal markers in an image.
  • a method for automatic identification of gastrointestinal markers comprising the steps of:
  • Determining the suspected gastrointestinal marker area in the image includes: segmenting the image using the maximum stable extreme value area method to determine the suspected gastrointestinal marker area in the image; calculating the minimum rectangular frame of the suspected gastrointestinal marker area in the image to form the first an array;
  • Removing overlapping suspected gastrointestinal marker areas including: S3.1 creating an empty second array; S3.2 putting the rectangle with the largest area into the second array, and removing the rectangle with the largest area from the first array; S3.3 traverse the first array, calculate the intersection and union ratio of the selected rectangular box and the rectangular box with the largest area, if the intersection ratio is greater than a certain threshold T1, delete the selected rectangular box from the first array Repeat steps S3.2 and S3.3 to the first array, until the first array becomes an empty array, and the second array finally formed is the array formed after removing the overlapping suspected gastrointestinal marker area;
  • it also includes improving the recognizability of the gastrointestinal markers in the image by enhancing the image contrast.
  • enhancing the image contrast includes the following steps:
  • the maximum value of the regional grayscale is Max_R
  • the minimum value of the regional grayscale is Min_R
  • the area of the suspected gastrointestinal marker is the grayscale value greater than the grayscale threshold T and the grayscale maximum value and the grayscale minimum value.
  • the value of the difference is less than the grayscale change threshold T_change; the value range of the grayscale threshold T is: 150 ⁇ T ⁇ 200, and the value range of the grayscale change threshold T_change is: 10 ⁇ T_change ⁇ 20.
  • L the length of the longest side of the suspected gastrointestinal marker area
  • the value range of L1 pixel is: 30 ⁇ L1 ⁇ 40
  • the value range of L2 pixel is: 20 ⁇ L2 ⁇ 30
  • the value range of L3 pixel is: 10 ⁇ L3 ⁇ 20
  • M represents the error range coefficient of L
  • the value range is: 1.0 ⁇ M ⁇ 1.2
  • the suspected gastrointestinal marker region is not a gastrointestinal marker.
  • the shape of the gastrointestinal marker or the type of the gastrointestinal marker is determined according to the length of the longest side of the smallest rectangular frame of the suspected gastrointestinal marker area;
  • the suspected gastrointestinal marker area is a combination of multiple gastrointestinal markers
  • the suspected gastrointestinal marker area is a three-compartment gastrointestinal marker
  • the suspected gastrointestinal marker area is an "O" ring gastrointestinal marker
  • An automatic identification system for gastrointestinal markers comprising:
  • the suspected gastrointestinal marker area identification module is used to determine the suspected gastrointestinal marker area in the image; the suspected gastrointestinal marker area identification module includes: a suspected gastrointestinal marker area determination module, using the maximum stable extreme value area method Segmenting the image, and determining the suspected gastrointestinal marker area in the image; the suspected gastrointestinal marker area marking module is used to calculate the smallest rectangular frame of the suspected gastrointestinal marker area, and form a first array;
  • the de-overlapping module is used to remove the overlapping suspected gastrointestinal marker regions;
  • the de-overlapping module includes: a rectangular frame area acquisition module, used to obtain the area of the rectangular frame in the first array; an overlapping rectangular frame analysis and removal module, with To perform the following steps: S3.1 creates an empty second array; S3.2 puts the rectangle with the largest area in the first array into the second array, and removes the rectangle with the largest area from the first array; S3.3 Traverse the first array, calculate the intersection of the selected rectangular box and the rectangular box with the largest area and the intersection ratio of the union, if the intersection ratio is greater than a certain threshold T1, delete the selected rectangular box from the first array; Repeat steps S3.2 and S3.3 for an array until the first array becomes an empty array, and the second array finally formed is an array formed after removing the overlapping suspected gastrointestinal marker regions;
  • the gastrointestinal marker determination module is used to determine whether the suspected gastrointestinal marker area belongs to the gastrointestinal marker.
  • the automatic identification system for gastrointestinal markers further includes an image processing module for improving the recognizability of the gastrointestinal markers in the image, and the image processing module includes:
  • the gray value calculation module is used to calculate the gray value range in the image, and obtain the gray minimum value g min and the gray maximum value g max ;
  • the image contrast enhancement module is used to stretch the gray value of the image to the interval of [0,255].
  • the suspected gastrointestinal marker area determination module is used to obtain the maximum value Max_R of the regional gray level and the minimum value Min_R of the regional gray level of the image area R, and it is considered to be a suspected gastrointestinal marker area if the following formula is satisfied: Min_R> T, and Max_R-Min_R ⁇ T_change; the value range of the grayscale threshold T is: 150 ⁇ T ⁇ 200, and the value range of the grayscale change threshold T_change is: 10 ⁇ T_change ⁇ 20.
  • the gastrointestinal marker determination module includes:
  • a gastrointestinal marker determination module which determines whether the suspected gastrointestinal marker area belongs to a gastrointestinal marker according to the length of the longest side; the gastrointestinal marker determination module further includes a gastrointestinal marker type judgment module, The length of the longest side determines which type of gastrointestinal marker the suspected gastrointestinal marker region belongs to.
  • the gastrointestinal marker judgment module judging which type of gastrointestinal marker the suspected gastrointestinal marker region belongs to according to the length of the longest side includes: defining L as the length of the longest side of the suspected gastrointestinal marker region; Length, L1 pixel value range is: 30 ⁇ L1 ⁇ 40, L2 pixel value range is: 20 ⁇ L2 ⁇ 30, L3 pixel value range is: 10 ⁇ L3 ⁇ 20, M represents the error range coefficient of L, take The value range is: 1.0 ⁇ M ⁇ 1.2; if L>L1*2 or L ⁇ L3*(M-1), the suspected gastrointestinal marker area is not a gastrointestinal marker.
  • the gastrointestinal marker judging module determines which type of gastrointestinal marker the suspected gastrointestinal marker region belongs to according to the length of the longest side, further comprising: if L1*M ⁇ L ⁇ L1*2, the suspected gastric The intestinal marker area is a combination of multiple gastrointestinal markers; (0.5*L1+0.5*L2) ⁇ L ⁇ L1*M, the suspected gastrointestinal marker area is a three-compartment gastrointestinal marker; (0.5*L2+ 0.5*L3) ⁇ L ⁇ L2*M, the suspected gastrointestinal marker area is an "O" ring type gastrointestinal marker; L3*(M-1) ⁇ L ⁇ L3*M, the suspected gastrointestinal marker area is a circle Dot-type gastrointestinal markers.
  • An electronic device includes a memory and a processor, wherein the memory stores a computer program that can be executed on the processor, and when the processor executes the program, the steps in the above image recognition method are implemented.
  • the beneficial effects of the present invention are: the automatic identification method of gastrointestinal markers of the present invention can automatically detect the position of the gastrointestinal markers in the image by processing and analyzing the image, and judge the gastrointestinal motility, The workload of doctors is greatly reduced.
  • FIG. 1 is a flow chart of a method for automatic identification of gastrointestinal markers according to a preferred embodiment of the present invention
  • Fig. 2 is the flow chart of the automatic identification method of gastrointestinal markers according to another specific embodiment of the present invention.
  • Figure 3 is a schematic diagram of the result after the original picture is processed in steps S1 and S2;
  • FIG. 4 is a schematic diagram of the result after the processing of step S3 is performed on the basis of FIG. 3 .
  • the automatic identification method of the gastrointestinal marker of the present invention is to identify the position of the gastrointestinal marker in the image.
  • the gastrointestinal marker can be X-ray contrast agents such as barium sulfate, bismuth salt or tungsten body in the prior art, and the shape and structure of the gastrointestinal marker is not limited; it can also be a newly developed gastrointestinal marker.
  • the automatic identification method of gastrointestinal markers mainly includes the following steps: S1 to improve the recognizability of gastrointestinal markers in the image, so as to be able to quickly determine the area of suspected gastrointestinal markers; S2 Determine the suspected gastrointestinal marker area in the image; S3 remove the overlapping suspected gastrointestinal marker area; S4 determine whether the suspected gastrointestinal marker area belongs to the gastrointestinal marker.
  • the method can automatically detect the position of the gastrointestinal marker in the image, judge the gastrointestinal motility, and greatly reduce the workload of the doctor.
  • Step S1 specifically improves the recognizability of the gastrointestinal marker in the image by enhancing the image contrast; that is, the image contrast enhancement is performed on the acquired original image to improve the recognizability of the gastrointestinal marker in the image.
  • step S1 includes the following steps: S1.1 calculates the gray value range in the image, and obtains the gray value minimum value g min and the gray value maximum value g max ; S1.2 stretches the gray value of the image to [0,255 ] to enhance the recognition of gastrointestinal markers in the image.
  • step S1 may be omitted.
  • Step S2 uses the Maximum Stable Extremal Regions (Maximally Stable Extremal Regions, MSER) method to segment the image, determines the suspected gastrointestinal marker region in the image, and calculates the minimum rectangular frame of the suspected gastrointestinal marker region.
  • MSER Maximum Stable Extremal Regions
  • the gastrointestinal marker area is the area with little change in gray value and above the background.
  • the maximum value of the regional gray level is Max_R
  • the minimum value of the regional gray level is Min_R
  • the area of the suspected gastrointestinal marker is set so that the gray value is greater than the gray threshold T, and the maximum gray value and the minimum gray value are The difference is less than the grayscale change threshold T_change area.
  • step S2 also includes calculating the smallest rectangular frame of the suspected gastrointestinal marker region in the image, and the result is to form a first array (rectangular frame array rectangles[]).
  • the value range of the grayscale threshold T is: 150 ⁇ T ⁇ 200, and the value range of the grayscale change threshold T_change is: 10 ⁇ T_change ⁇ 20.
  • non-maximum suppression (Non-Maximum Suppression, NMS) is used to remove the overlapping suspected gastrointestinal marker regions.
  • NMS Non-Maximum Suppression
  • the present invention adopts non-maximum value suppression to suppress those redundant regions, and the suppression process is an iterative-traversal-elimination process.
  • step S3 includes the following steps: S3.1 creates an empty second array (rectangles_keep[]); S3.2 sorts the rectangles in the first array according to the area from large to small, this sorting process is not a necessary step , can also be omitted; record the rectangle with the largest area (rectangle_max) as the first rectangle and put it into the second array, correspondingly, remove the rectangle with the largest area from the first array; S3.3 traverse the first array, calculate The intersection of the selected rectangle (rectangle_select) and the rectangle with the largest area and the intersection over union (IOU) of the union, if the intersection over union (IOU) is greater than a certain threshold T1, the selected rectangle will be selected from the first delete from an array.
  • the threshold T1 can be set according to actual requirements.
  • step S3.3 sequentially calculates the IOUs of the remaining rectangles in the first array and the first rectangles placed in the second array, and deletes the corresponding rectangles with IOU>T1, thereby forming a new first array.
  • step S3.2 marks the rectangle with the largest area in the new first array as the second rectangle and puts it into the second array, correspondingly deleting the second rectangle in the new first array; at this time, The second array contains two rectangles: the first rectangle and the second rectangle.
  • step S3.3 sequentially calculates the IOUs of the remaining rectangles in the new first array and the second rectangles placed in the second array, and deletes the corresponding rectangles with IOU>T1, thereby forming a new first array.
  • the rectangular boxes in the first array are continuously reduced until it becomes an empty array; while the rectangular boxes in the second array are continuously accumulated, and the finally formed second array is to remove the overlapping suspected gastrointestinal The array formed after the marker area.
  • FIG. 3 and FIG. 4 there are three suspected gastrointestinal marker regions, which are marked as A, B and C, respectively.
  • the regions where B and C are located in Fig. 3 both have two region boxes representing suspected gastrointestinal marker regions, representing overlapping suspected gastrointestinal marker regions; then after step S3, in Fig. 4
  • the overlapping suspected gastrointestinal marker regions in the regions of B and C are removed, that is, there is only one area box representing the suspected gastrointestinal marker region, which represents the suspected gastrointestinal marker region from which the overlapping region has been removed; it is convenient for accurate To identify whether the suspected gastrointestinal marker area belongs to the gastrointestinal marker and its kind.
  • the boxes in Figures 3 and 4 represent the actual suspected gastrointestinal marker area, and the rectangular box described above is the smallest area that can cover the actual suspected gastrointestinal marker area during the above-mentioned identification method. , not shown in the figure.
  • Step S4 determines whether the suspected gastrointestinal marker region belongs to the gastrointestinal marker. Specifically, whether the suspected gastrointestinal marker region belongs to the gastrointestinal marker is determined according to the length of the longest side of each rectangular frame in the second array.
  • the present invention defines L as the length of the longest side of the suspected gastrointestinal marker area, the value range of L1 pixel is: 30 ⁇ L1 ⁇ 40, the value range of L2 pixel is: 20 ⁇ L2 ⁇ 30, and the value range of L3 pixel is : 10 ⁇ L3 ⁇ 20, M represents the error range coefficient of L, the value range is: 1.0 ⁇ M ⁇ 1.2; If L>L1*2 or L ⁇ L3*(M-1), the suspected gastrointestinal marker area Not a gastrointestinal marker.
  • the shape of the gastrointestinal marker or the type of the gastrointestinal marker can also be determined according to the length of the longest side of the smallest rectangular frame of the suspected gastrointestinal marker region.
  • the suspected gastrointestinal marker area is the overlap of multiple gastrointestinal markers; (0.5*L1+0.5*L2) ⁇ L ⁇ L1*M, the suspected gastrointestinal marker The marker area is a three-compartment gastrointestinal marker; (0.5*L2+0.5*L3) ⁇ L ⁇ L2*M, the suspected gastrointestinal marker area is an "O" ring gastrointestinal marker; L3*(M-1) ⁇ L ⁇ L3*M, the suspected gastrointestinal marker area is a dot-type gastrointestinal marker.
  • the present invention also provides an automatic identification system for gastrointestinal markers, comprising:
  • Suspected gastrointestinal marker area identification module used to determine the suspected gastrointestinal marker area in the image
  • the gastrointestinal marker determination module is used to determine whether the suspected gastrointestinal marker area belongs to the gastrointestinal marker.
  • the automatic identification system for gastrointestinal markers further includes an image processing module for improving the recognizability of the gastrointestinal markers in the image;
  • the image processing module includes: a gray value calculation module for calculating the image In the gray value range, the gray minimum value g min and the gray maximum value g max are obtained;
  • the image contrast enhancement module is used to stretch the gray value of the image to the interval of [0,255].
  • the automatic identification system for the gastrointestinal marker may not include the image processing module.
  • the suspected gastrointestinal marker region identification module includes:
  • the suspected gastrointestinal marker region determination module uses the maximum stable extreme value region method to segment the image, and determines the suspected gastrointestinal marker region in the image. Specifically, the suspected gastrointestinal marker region determination module is used to obtain the maximum value of regional gray level Max_R and the minimum value of regional gray level Min_R of the image region R, and it is considered to be a suspected region of gastrointestinal markers if the following formula is satisfied: Min_R> T, and Max_R-Min_R ⁇ T_change; the value range of the grayscale threshold T is: 150 ⁇ T ⁇ 200, and the value range of the grayscale change threshold T_change is: 10 ⁇ T_change ⁇ 20.
  • the suspected gastrointestinal marker area marking module is used to calculate the smallest rectangular frame of the suspected gastrointestinal marker area, and form a first array.
  • step S2 For the method for the suspected gastrointestinal marker region identification module to determine the suspected gastrointestinal marker region in the image, refer to the above step S2.
  • the de-overlapping module includes: a rectangular frame area acquisition module, used to obtain the area of the rectangular frame in the first array; an overlapping rectangular frame analysis and removal module, used to create an empty second array; Put the selected rectangle into the second array, correspondingly, remove the rectangle with the largest area from the first array; traverse the first array, calculate the intersection and union ratio of the selected rectangle and the rectangle with the largest area, If the intersection ratio is greater than a certain threshold T 1 , the selected rectangle is deleted from the first array.
  • the method for removing overlapping suspected gastrointestinal marker regions by the de-overlapping module refers to the above step S3.
  • the gastrointestinal marker determination module includes: an acquisition of the longest side module for acquiring the longest side of each rectangular frame in the second array; a gastrointestinal marker determination module for judging the suspected stomach according to the length of the longest side Whether the intestinal marker region is a gastrointestinal marker.
  • the gastrointestinal marker determination module further includes a gastrointestinal marker type judgment module, which determines which type of gastrointestinal marker the gastrointestinal marker region belongs to according to the length of the longest side.
  • the gastrointestinal marker judgment module judging which type of gastrointestinal marker the suspected gastrointestinal marker region belongs to according to the length of the longest side includes: defining L as the length of the longest side of the suspected gastrointestinal marker region; Length, L1 pixel value range is: 30 ⁇ L1 ⁇ 40, L2 pixel value range is: 20 ⁇ L2 ⁇ 30, L3 pixel value range is: 10 ⁇ L3 ⁇ 20, M represents the error range coefficient of L, take The value range is: 1.0 ⁇ M ⁇ 1.2; if L>L1*2 or L ⁇ L3*(M-1), the suspected gastrointestinal marker area is not a gastrointestinal marker.
  • the gastrointestinal marker judging module determines which type of gastrointestinal marker the suspected gastrointestinal marker region belongs to according to the length of the longest side, further comprising: if L1*M ⁇ L ⁇ L1*2, the suspected gastric The intestinal marker area is a combination of multiple gastrointestinal markers; (0.5*L1+0.5*L2) ⁇ L ⁇ L1*M, the suspected gastrointestinal marker area is a three-compartment gastrointestinal marker; (0.5*L2+ 0.5*L3) ⁇ L ⁇ L2*M, the suspected gastrointestinal marker area is an "O" ring type gastrointestinal marker; L3*(M-1) ⁇ L ⁇ L3*M, the suspected gastrointestinal marker area is a circle Dot gastrointestinal marker
  • step S4 for the method for determining the gastrointestinal marker by the gastrointestinal marker determination module, refer to step S4 of the above-mentioned automatic identification method for gastrointestinal markers, which will not be repeated here.
  • the present invention also provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and the processor implements the steps in the image recognition method when the processor executes the program.
  • the present invention also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, implements the steps in the above-mentioned image recognition method.
  • the automatic identification method of gastrointestinal markers of the present invention can automatically detect the position and type of gastrointestinal markers in the image, so as to identify the positions of different types of gastrointestinal markers in the digestive tract, and judge the subject. gastrointestinal motility.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Medical Informatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Quality & Reliability (AREA)
  • Geometry (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Surgery (AREA)
  • Optics & Photonics (AREA)
  • Pathology (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • High Energy & Nuclear Physics (AREA)
  • Animal Behavior & Ethology (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Physiology (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

一种胃肠标记物自动识别方法及识别系统，包括如下步骤：在图像中确定疑似胃肠标记物区域；去除重叠的疑似胃肠标记物区域；判断疑似胃肠标记物区域是否属于胃肠标记物。该方法通过对图像的处理和分析，能够自动检测胃肠标记物在图像中的位置。

Description

胃肠标记物自动识别方法及识别方法
本申请要求了申请日为2020年09月01日，申请号为202010902829.6，发明名称为“胃肠标记物自动识别方法及识别系统”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
本发明涉及医疗器械技术领域，尤其涉及一种可以自动检测胃肠标记物在图像中的位置的胃肠标记物自动识别方法及识别系统。
背景技术
通过X射线胃肠标记物检查肠胃动力，是检查肠胃病的重要诊断手段之一。胃肠标记物胶囊是一种包含20～24个薄片不透X射线胃肠标记物的胶囊，胶囊壳可在胃肠道内自然溶解，而胃肠标记物为生物相容性材料，可用于检测胃肠道通过时间，是判断是否有便秘或其他胃肠道疾病的重要诊断手段。
目前，胃肠动力检查中，拍摄的X光片需要依靠医生判断胃肠标记物的位置和类别，大大增加了医生的工作量。
鉴于此，有必要提出一种胃肠标记物自动识别方法及识别系统，以解决上述问题。
发明内容
本发明的目的在于提供一种胃肠标记物自动识别方法及识别系统，可以自动检测胃肠标记物在图像中的位置。
为了实现上述发明目的，本发明采用如下技术方案：
一种胃肠标记物自动识别方法，其特征在于，包括如下步骤：
在图像中确定疑似胃肠标记物区域，包括：使用最大稳定极值区域方法分割图像，在图像中确定疑似胃肠标记物区域；计算图像中疑似胃肠标记物区域的最小矩形框，形成第一数组；
去除重叠的疑似胃肠标记物区域，包括：S3.1创建一个空的第二数组；S3.2将面积最大的矩形框放入第二数组中，第一数组中去掉面积最大的矩形框；S3.3遍历第一数组，计算选中的矩形框和面积最大的矩形框的交集和并集的交并比值，如果交并比值大于一定阈值T1，则将选中的矩形框从第一数组中删除；对第一数组重复步骤S3.2和S3.3，直至第一数组成为空数组，最终形成的第二数组为去除重叠的疑似胃肠标记物区域后构成的数组；
判断疑似胃肠标记物区域是否属于胃肠标记物。
进一步地，还包括通过增强图像对比度提高胃肠标记物在图像中的可辨识度。
进一步地，增强图像对比度包括如下步骤：
计算图像中灰度值范围，求出灰度最小值g min和灰度最大值g max；
将图像的灰度值拉伸到[0,255]的区间。
进一步地，在图像区域R中，区域灰度最大值是Max_R，区域灰度最小值是Min_R，疑似胃肠标记物区域是灰度值大于灰度阈值T且灰度最大值与灰度最小值的差值小于灰度变化阈值T_change的区域；灰度阈值T的取值范围为：150≤T≤200，灰度变化阈值T_change的取值范围为：10≤T_change≤20。
进一步地，根据第二数组中每一个矩形框的最长边的长度判断疑似胃肠标记物区域是否属于胃肠标记物。
进一步地，定义L为疑似胃肠标记物区域的最长边的长度，L1像素取值范围为：30＜L1≤40，L2像素取值范围为：20＜L2≤30，L3像素 取值范围为：10＜L3≤20，M表示L的误差范围系数，取值范围为：1.0≤M≤1.2；
若L>L1*2或L<L3*(M-1)，疑似胃肠标记物区域不是胃肠标记物。
进一步地，根据疑似胃肠标记物区域的最小矩形框的最长边的长度判断胃肠标记物的形状或胃肠标记物的种类；
若L1*M<L<L1*2,疑似胃肠标记物区域是多个胃肠标记物重合在一起；
(0.5*L1+0.5*L2)<L<L1*M，疑似胃肠标记物区域是三室型胃肠标记物；
(0.5*L2+0.5*L3)<L<L2*M,疑似胃肠标记物区域是“O”环型胃肠标记物；
L3*(M-1)<L<L3*M,疑似胃肠标记物区域是圆点型胃肠标记物。
一种胃肠标记物自动识别系统，包括：
疑似胃肠标记物区域识别模块，用以在图像中确定疑似胃肠标记物区域；所述疑似胃肠标记物区域识别模块包括：疑似胃肠标记物区域确定模块，使用最大稳定极值区域方法分割图像，在图像中确定疑似胃肠标记物区域；疑似胃肠标记物区域标记模块，用以计算疑似胃肠标记物区域的最小矩形框，并形成第一数组；
去重叠模块，用以去除重叠的疑似胃肠标记物区域；所述去重叠模块包括：矩形框面积获取模块，用以获取第一数组中的矩形框的面积；重叠矩形框分析去除模块，用以执行如下步骤：S3.1创建一个空的第二数组；S3.2将第一数组中面积最大的矩形框放入第二数组中，第一数组中去掉面积最大的矩形框；S3.3遍历第一数组，计算选中的矩形框和面积最大的矩形框的交集和并集的交并比值，如果交并比值大于一定阈值T1，则将选中的矩形框从第一数组中删除；对第一数组重复步骤S3.2和S3.3，直至第一数组成为空数组，最终形成的第二数组为去除重叠的疑 似胃肠标记物区域后构成的数组；
胃肠标记物确定模块，用以判断疑似胃肠标记物区域是否属于胃肠标记物。
进一步地，所述胃肠标记物自动识别系统还包括用以提高胃肠标记物在图像中的可辨识度的图像处理模块，所述图像处理模块包括：
灰度值计算模块，用以计算图像中灰度值范围，求出灰度最小值g min和灰度最大值g max；
图像对比度增强模块，用以将图像的灰度值拉伸到[0,255]的区间。
进一步地，所述疑似胃肠标记物区域确定模块，用以获取图像区域R的区域灰度最大值Max_R和区域灰度最小值Min_R，满足下面公式则认为是疑似胃肠标记物区域：Min_R>T，且Max_R-Min_R<T_change；灰度阈值T的取值范围为：150≤T≤200，灰度变化阈值T_change的取值范围为：10≤T_change≤20。
进一步地，所述胃肠标记物确定模块包括：
获取最长边模块，用以获取第二数组中每一个矩形框的最长边；
胃肠标记物判断模块，根据所述最长边的长度判断疑似胃肠标记物区域是否属于胃肠标记物；所述胃肠标记物确定模块还包括胃肠标记物种类判断模块，根据所述最长边的长度判断疑似胃肠标记物区域属于哪一类胃肠标记物。
进一步地，所述胃肠标记物判断模块根据所述最长边的长度判断疑似胃肠标记物区域属于哪一类胃肠标记物包括：定义L为疑似胃肠标记物区域的最长边的长度，L1像素取值范围为：30＜L1≤40，L2像素取值范围为：20＜L2≤30，L3像素取值范围为：10＜L3≤20，M表示L的误差范围系数，取值范围为：1.0≤M≤1.2；若L>L1*2或L<L3*(M-1)，疑似胃肠标记物区域不是胃肠标记物。
进一步地，所述胃肠标记物判断模块根据所述最长边的长度判断疑 似胃肠标记物区域属于哪一类胃肠标记物还包括：若L1*M<L<L1*2,疑似胃肠标记物区域是多个胃肠标记物重合在一起；(0.5*L1+0.5*L2)<L<L1*M，疑似胃肠标记物区域是三室型胃肠标记物；(0.5*L2+0.5*L3)<L<L2*M,疑似胃肠标记物区域是“O”环型胃肠标记物；L3*(M-1)<L<L3*M,疑似胃肠标记物区域是圆点型胃肠标记物。一种电子设备，包括存储器和处理器，所述存储器存储有可在所述处理器上运行的计算机程序，所述处理器执行所述程序时实现上述图像识别方法中的步骤。
一种计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时实现上述图像识别方法中的步骤。
与现有技术相比，本发明的有益效果在于：本发明的胃肠标记物自动识别方法，通过对图像的处理和分析，能够自动检测胃肠标记物在图像中的位置，判断肠胃动力，大大减少了医生的工作量。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据提供的附图获得其他的附图。
图1是本发明一较佳实施例的胃肠标记物自动识别方法的流程图；
图2是本发明另一具体实施例的胃肠标记物自动识别方法的流程图；
图3对原始图片进行步骤S1、S2处理后的结果示意图；
图4是在图3基础上进行步骤S3处理后的结果示意图。
具体实施方式
以下将结合附图所示的具体实施方式对本申请进行详细描述。但这些实施方式并不限制本申请，本领域的普通技术人员根据这些实施方式所做出的结构、方法、或功能上的变换均包含在本申请的保护范围内。
本发明的胃肠标记物自动识别方法，是为了识别图像中胃肠标记物的位置。胃肠标记物可以采用现有技术中的硫酸钡、铋盐或钨体等X线造影剂，且胃肠标记物的形状结构不限；也可以为新研发的胃肠标记物。
请参考图1～图4所示，所述胃肠标记物自动识别方法主要包括如下步骤：S1提高胃肠标记物在图像中的可辨识度，以能够迅速判断疑似胃肠标记物区域；S2在图像中确定疑似胃肠标记物区域；S3去除重叠的疑似胃肠标记物区域；S4判断疑似胃肠标记物区域是否属于胃肠标记物。该方法通过对图像的处理，能够自动检测胃肠标记物在图像中的位置，判断肠胃动力，大大减少了医生的工作量。
步骤S1具体通过增强图像对比度提高胃肠标记物在图像中的可辨识度；即对获取的原始图像做图像对比度增强，提高胃肠标记物在图像中的可辨识度。
具体地，步骤S1包括如下步骤：S1.1计算图像中灰度值范围，求出灰度最小值g min和灰度最大值g max；S1.2将图像的灰度值拉伸到[0,255]的区间，增强图像中胃肠标记物辨识度。
当然，当胃肠标记物在图像中的可辨识度较高，满足后续步骤的需求时，可以省略步骤S1。
步骤S2使用最大稳定极值区域(Maximally Stable Extremal Regions，MSER)方法分割图像，在图像中确定疑似胃肠标记物区域，并计算疑似胃肠标记物区域的最小矩形框。在X光图像中，胃肠标记物区域是灰度值变化不大而且高于背景的区域。在图像区域R中，区域灰度最大值 是Max_R，区域灰度最小值是Min_R，设定疑似胃肠标记物区域是灰度值大于灰度阈值T，且灰度最大值与灰度最小值的差值小于灰度变化阈值T_change的区域。即满足下面公式可以认为是疑似胃肠标记物区域：Min_R>T，且Max_R-Min_R<T_change。如此，使用寻找图像中最大稳定极值区域的方法可以找到疑似胃肠标记物的多个区域。进一步地，步骤S2还包括计算图像中疑似胃肠标记物区域的最小矩形框，结果为形成第一数组(矩形框数组rectangles[])。
一具体实施例中，灰度阈值T的取值范围为：150≤T≤200，灰度变化阈值T_change的取值范围为：10≤T_change≤20。
步骤S2获取的疑似胃肠标记物区域有很多区域可能存在重叠的现象，步骤S3使用非极大值抑制(Non-Maximum Suppression，NMS)去除重叠的疑似胃肠标记物区域。本发明采用非极大值抑制是为了抑制那些冗余的区域，抑制的过程是一个迭代-遍历-消除的过程。
具体地，步骤S3包括如下步骤：S3.1创建一个空的第二数组(rectangles_keep[])；S3.2将第一数组中的矩形框按照面积从大到小排序，该排序过程并非必需步骤，也可以省略；将面积最大的矩形框(rectangle_max)记作第一矩形框放入第二数组中，对应的，第一数组中去掉面积最大的矩形框；S3.3遍历第一数组，计算选中的矩形框(rectangle_select)和面积最大的矩形框的交集和并集的交并比值(intersection over union，IOU)，如果交并比值(IOU)大于一定阈值T1，就将选中的矩形框从第一数组中删除。其中，阈值T1可以根据实际需求进行设定。
即，步骤S3.3依次计算第一数组中剩余矩形框与放入第二数组中的第一矩形框的IOU，并删除IOU>T1的对应矩形框，从而形成新的第一数组。
然后，对新的第一数组重复步骤S3.2和S3.3。具体地，步骤S3.2 将新的第一数组中面积最大的矩形框记作第二矩形框放入第二数组中，对应在新的第一数组中删除该第二矩形框；此时，第二数组中包含两个矩形框：第一矩形框和第二矩形框。步骤S3.3依次计算新的第一数组中剩余矩形框与放入第二数组中的第二矩形框的IOU，并删除IOU>T1的对应矩形框，从而形成又一新的第一数组。
以此类推，在遍历过程中，第一数组中的矩形框不断减少，直至成为空数组；而第二数组中的矩形框则不断累加，最终形成的第二数组即为去除重叠的疑似胃肠标记物区域后构成的数组。
由图3和图4比对可知，经过上述步骤，B和C处重合的疑似胃肠标记物区域被过滤掉了。
如图3和图4所示，一具体实施例中，有三处疑似胃肠标记物区域，分别标记为A、B和C。经过步骤S1和S2后，图3中B、C所在区域均具有两个代表疑似胃肠标记物区域的区域框，代表有重叠的疑似胃肠标记物区域；然后经过步骤S3后，图4中去除了B、C所在区域的重叠的疑似胃肠标记物区域，即仅剩下了一个代表疑似胃肠标记物区域的区域框，代表已经去除了重叠区的疑似胃肠标记物区域；便于精确地识别疑似胃肠标记物区域是否属于胃肠标记物及其种类。需要说明的是：图3和图4中的框代表实际的疑似胃肠标记物区域，而上文描述的矩形框为在上述识别方法过程中，能够覆盖实际的疑似胃肠标记物区域的最小的矩形框，未在图中显示。
步骤S4判断疑似胃肠标记物区域是否属于胃肠标记物。具体地，根据第二数组中每一个矩形框的最长边的长度判断疑似胃肠标记物区域是否属于胃肠标记物。
本发明定义L为疑似胃肠标记物区域的最长边的长度，L1像素取值范围为：30＜L1≤40，L2像素取值范围为：20＜L2≤30，L3像素取值范围为：10＜L3≤20，M表示L的误差范围系数，取值范围为：1.0≤M≤1.2； 若L>L1*2或L<L3*(M-1)，则疑似胃肠标记物区域不是胃肠标记物。
进一步地，根据疑似胃肠标记物区域的最小矩形框的最长边的长度还能够判断胃肠标记物的形状或胃肠标记物的种类。具体地，L1*M<L<L1*2，疑似胃肠标记物区域是多个胃肠标记物重合在一起；(0.5*L1+0.5*L2)<L<L1*M，疑似胃肠标记物区域是三室型胃肠标记物；(0.5*L2+0.5*L3)<L<L2*M，疑似胃肠标记物区域是“O”环型胃肠标记物；L3*(M-1)<L<L3*M，疑似胃肠标记物区域是圆点型胃肠标记物。
本发明还提供一种胃肠标记物自动识别系统，包括：
疑似胃肠标记物区域识别模块，用以在图像中确定疑似胃肠标记物区域；
去重叠模块，用以去除重叠的疑似胃肠标记物区域；
胃肠标记物确定模块，用以判断疑似胃肠标记物区域是否属于胃肠标记物。
进一步地，所述胃肠标记物自动识别系统还包括用以提高胃肠标记物在图像中的可辨识度的图像处理模块；所述图像处理模块包括：灰度值计算模块，用以计算图像中灰度值范围，求出灰度最小值g min和灰度最大值g max；图像对比度增强模块，用以将图像的灰度值拉伸到[0,255]的区间。所述图像处理模块提高胃肠标记物在图像中的可辨识度的方法参考上述步骤S1。
当然，当胃肠标记物在图像中的可辨识度较高，满足后续步骤的需求时，所述胃肠标记物自动识别系统可以不包括所述图像处理模块。
所述疑似胃肠标记物区域识别模块包括：
疑似胃肠标记物区域确定模块，使用最大稳定极值区域方法分割图像，在图像中确定疑似胃肠标记物区域。具体地，所述疑似胃肠标记物区域确定模块，用以获取图像区域R的区域灰度最大值Max_R和区域灰度最小值Min_R，满足下面公式则认为是疑似胃肠标记物区域： Min_R>T，且Max_R-Min_R<T_change；灰度阈值T的取值范围为：150≤T≤200，灰度变化阈值T_change的取值范围为：10≤T_change≤20。
疑似胃肠标记物区域标记模块，用以计算疑似胃肠标记物区域的最小矩形框，并形成第一数组。
所述疑似胃肠标记物区域识别模块在图像中确定疑似胃肠标记物区域的方法参考上述步骤S2。
所述去重叠模块包括：矩形框面积获取模块，用以获取第一数组中的矩形框的面积；重叠矩形框分析去除模块，用以创建一个空的第二数组；将第一数组中面积最大的矩形框放入第二数组中，相应的，第一数组中去掉面积最大的矩形框；遍历第一数组，计算选中的矩形框和面积最大的矩形框的交集和并集的交并比值，如果交并比值大于一定阈值T 1，则将选中的矩形框从第一数组中删除。具体地，所述去重叠模块去除重叠的疑似胃肠标记物区域的方法参考上述步骤S3。
所述胃肠标记物确定模块包括：获取最长边模块，用以获取第二数组中每一个矩形框的最长边；胃肠标记物判断模块，根据所述最长边的长度判断疑似胃肠标记物区域是否属于胃肠标记物。
进一步地，所述胃肠标记物确定模块还包括胃肠标记物种类判断模块，根据所述最长边的长度判断胃肠标记物区域属于哪一类胃肠标记物。
进一步地，所述胃肠标记物判断模块根据所述最长边的长度判断疑似胃肠标记物区域属于哪一类胃肠标记物包括：定义L为疑似胃肠标记物区域的最长边的长度，L1像素取值范围为：30＜L1≤40，L2像素取值范围为：20＜L2≤30，L3像素取值范围为：10＜L3≤20，M表示L的误差范围系数，取值范围为：1.0≤M≤1.2；若L>L1*2或L<L3*(M-1)，疑似胃肠标记物区域不是胃肠标记物。
进一步地，所述胃肠标记物判断模块根据所述最长边的长度判断疑似胃肠标记物区域属于哪一类胃肠标记物还包括：若L1*M<L<L1*2,疑 似胃肠标记物区域是多个胃肠标记物重合在一起；(0.5*L1+0.5*L2)<L<L1*M，疑似胃肠标记物区域是三室型胃肠标记物；(0.5*L2+0.5*L3)<L<L2*M,疑似胃肠标记物区域是“O”环型胃肠标记物；L3*(M-1)<L<L3*M,疑似胃肠标记物区域是圆点型胃肠标记物
具体地，所述胃肠标记物确定模块确定胃肠标记物的方法参考上述胃肠标记物自动识别方法步骤S4，于此不再赘述。
本发明还提供一种电子设备，包括存储器和处理器，所述存储器存储有可在所述处理器上运行的计算机程序，所述处理器执行所述程序时实现上述图像识别方法中的步骤。
本发明还提供一种计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时实现上述图像识别方法中的步骤。综上所述，本发明的胃肠标记物自动识别方法，可以自动检测胃肠标记物在图像中的位置和种类，从而识别不同种类的胃肠标记物在消化道中的部位，判断被检者的胃肠动力。
应当理解，虽然本说明书按照实施方式加以描述，但并非每个实施方式仅包含一个独立的技术方案，说明书的这种叙述方式仅仅是为清楚起见，本领域技术人员应当将说明书作为一个整体，各实施方式中的技术方案也可以经适当组合，形成本领域技术人员可以理解的其他实施方式。
上文所列出的一系列的详细说明仅仅是针对本发明的可行性实施方式的具体说明，并非用以限制本发明的保护范围，凡未脱离本发明技艺精神所作的等效实施方式或变更均应包含在本发明的保护范围之内。

Claims (15)

  1. 一种胃肠标记物自动识别方法，其特征在于，包括如下步骤：
    在图像中确定疑似胃肠标记物区域，包括：使用最大稳定极值区域方法分割图像，在图像中确定疑似胃肠标记物区域；计算图像中疑似胃肠标记物区域的最小矩形框，形成第一数组；
    去除重叠的疑似胃肠标记物区域，包括：S3.1创建一个空的第二数组；S3.2将面积最大的矩形框放入第二数组中，第一数组中去掉面积最大的矩形框；S3.3遍历第一数组，计算选中的矩形框和面积最大的矩形框的交集和并集的交并比值，如果交并比值大于一定阈值T1，则将选中的矩形框从第一数组中删除；对第一数组重复步骤S3.2和S3.3，直至第一数组成为空数组，最终形成的第二数组为去除重叠的疑似胃肠标记物区域后构成的数组；
    判断疑似胃肠标记物区域是否属于胃肠标记物。
  2. 根据权利要求1所述的胃肠标记物自动识别方法，其特征在于，还包括通过增强图像对比度提高胃肠标记物在图像中的可辨识度。
  3. 根据权利要求2所述的胃肠标记物自动识别方法，其特征在于，增强图像对比度包括如下步骤：
    计算图像中灰度值范围，求出灰度最小值g min和灰度最大值g max；
    将图像的灰度值拉伸到[0,255]的区间。
  4. 根据权利要求1所述的胃肠标记物自动识别方法，其特征在于，在图像区域R中，区域灰度最大值是Max_R，区域灰度最小值是Min_R，疑似胃肠标记物区域是灰度值大于灰度阈值T且灰度最大值与灰度最小值的差值小于灰度变化阈值T_change的区域；灰度阈值T的取值范围为：150≤T≤200，灰度变化阈值T_change的取值范围为：10≤T_change≤20。
  5. 根据权利要求1所述的胃肠标记物自动识别方法，其特征在于， 根据第二数组中每一个矩形框的最长边的长度判断疑似胃肠标记物区域是否属于胃肠标记物。
  6. 根据权利要求5所述的胃肠标记物自动识别方法，其特征在于，定义L为疑似胃肠标记物区域的最长边的长度，L1像素取值范围为：30＜L1≤40，L2像素取值范围为：20＜L2≤30，L3像素取值范围为：10＜L3≤20，M表示L的误差范围系数，取值范围为：1.0≤M≤1.2；
    若L>L1*2或L<L3*(M-1)，疑似胃肠标记物区域不是胃肠标记物。
  7. 根据权利要求6所述的胃肠标记物自动识别方法，其特征在于，根据疑似胃肠标记物区域的最小矩形框的最长边的长度判断胃肠标记物的形状或胃肠标记物的种类；
    若L1*M<L<L1*2,疑似胃肠标记物区域是多个胃肠标记物重合在一起；
    (0.5*L1+0.5*L2)<L<L1*M，疑似胃肠标记物区域是三室型胃肠标记物；
    (0.5*L2+0.5*L3)<L<L2*M,疑似胃肠标记物区域是“O”环型胃肠标记物；
    L3*(M-1)<L<L3*M,疑似胃肠标记物区域是圆点型胃肠标记物。
  8. 一种胃肠标记物自动识别系统，其特征在于，包括：
    疑似胃肠标记物区域识别模块，用以在图像中确定疑似胃肠标记物区域；所述疑似胃肠标记物区域识别模块包括：疑似胃肠标记物区域确定模块，使用最大稳定极值区域方法分割图像，在图像中确定疑似胃肠标记物区域；疑似胃肠标记物区域标记模块，用以计算疑似胃肠标记物区域的最小矩形框，并形成第一数组；
    去重叠模块，用以去除重叠的疑似胃肠标记物区域；所述去重叠模块包括：矩形框面积获取模块，用以获取第一数组中的矩形框的面积；重叠矩形框分析去除模块，用以执行如下步骤：S3.1创建一个空的第二 数组；S3.2将第一数组中面积最大的矩形框放入第二数组中，第一数组中去掉面积最大的矩形框；S3.3遍历第一数组，计算选中的矩形框和面积最大的矩形框的交集和并集的交并比值，如果交并比值大于一定阈值T1，则将选中的矩形框从第一数组中删除；对第一数组重复步骤S3.2和S3.3，直至第一数组成为空数组，最终形成的第二数组为去除重叠的疑似胃肠标记物区域后构成的数组；
    胃肠标记物确定模块，用以判断疑似胃肠标记物区域是否属于胃肠标记物。
  9. 根据权利要求8所述的胃肠标记物自动识别系统，其特征在于，所述胃肠标记物自动识别系统还包括用以提高胃肠标记物在图像中的可辨识度的图像处理模块，所述图像处理模块包括：
    灰度值计算模块，用以计算图像中灰度值范围，求出灰度最小值g min和灰度最大值g max；
    图像对比度增强模块，用以将图像的灰度值拉伸到[0,255]的区间。
  10. 根据权利要求8所述的胃肠标记物自动识别系统，其特征在于，所述疑似胃肠标记物区域确定模块用以获取图像区域R的区域灰度最大值Max_R和区域灰度最小值Min_R，满足下面公式则认为是疑似胃肠标记物区域：Min_R>T，且Max_R-Min_R<T_change；灰度阈值T的取值范围为：150≤T≤200，灰度变化阈值T_change的取值范围为：10≤T_change≤20。
  11. 根据权利要求8所述的胃肠标记物自动识别系统，其特征在于，所述胃肠标记物确定模块包括：
    获取最长边模块，用以获取第二数组中每一个矩形框的最长边；
    胃肠标记物判断模块，根据所述最长边的长度判断疑似胃肠标记物区域是否属于胃肠标记物；所述胃肠标记物确定模块还包括胃肠标记物种类判断模块，根据所述最长边的长度判断疑似胃肠标记物区域属于哪 一类胃肠标记物。
  12. 根据权利要求11所述的胃肠标记物自动识别系统，其特征在于，所述胃肠标记物判断模块根据所述最长边的长度判断疑似胃肠标记物区域属于哪一类胃肠标记物包括：定义L为疑似胃肠标记物区域的最长边的长度，L1像素取值范围为：30＜L1≤40，L2像素取值范围为：20＜L2≤30，L3像素取值范围为：10＜L3≤20，M表示L的误差范围系数，取值范围为：1.0≤M≤1.2；若L>L1*2或L<L3*(M-1)，疑似胃肠标记物区域不是胃肠标记物。
  13. 根据权利要求12所述的胃肠标记物自动识别系统，其特征在于，所述胃肠标记物判断模块根据所述最长边的长度判断疑似胃肠标记物区域属于哪一类胃肠标记物还包括：若L1*M<L<L1*2,疑似胃肠标记物区域是多个胃肠标记物重合在一起；(0.5*L1+0.5*L2)<L<L1*M，疑似胃肠标记物区域是三室型胃肠标记物；(0.5*L2+0.5*L3)<L<L2*M,疑似胃肠标记物区域是“O”环型胃肠标记物；L3*(M-1)<L<L3*M,疑似胃肠标记物区域是圆点型胃肠标记物。
  14. 一种电子设备，包括存储器和处理器，所述存储器存储有可在所述处理器上运行的计算机程序，其特征在于，所述处理器执行所述程序时实现权利要求1所述图像识别方法中的步骤。
  15. 一种计算机可读存储介质，其上存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现权利要求1所述图像识别方法中的步骤。
PCT/CN2021/115724 2020-09-01 2021-08-31 胃肠标记物自动识别方法及识别方法 WO2022048540A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP21863604.1A EP4209992A4 (en) 2020-09-01 2021-08-31 AUTOMATIC IDENTIFICATION METHOD AND IDENTIFICATION SYSTEM FOR GASTROINTESTINAL MARKERS
US18/024,215 US20230306589A1 (en) 2020-09-01 2021-08-31 Automatic identification method and identification system for gastrointestinal marker

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010902829.6A CN111768408B (zh) 2020-09-01 2020-09-01 胃肠标记物自动识别方法及识别系统
CN202010902829.6 2020-09-01

Publications (2)

Publication Number Publication Date
WO2022048540A1 true WO2022048540A1 (zh) 2022-03-10
WO2022048540A9 WO2022048540A9 (zh) 2022-04-28

Family

ID=72729616

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/115724 WO2022048540A1 (zh) 2020-09-01 2021-08-31 胃肠标记物自动识别方法及识别方法

Country Status (4)

Country Link
US (1) US20230306589A1 (zh)
EP (1) EP4209992A4 (zh)
CN (1) CN111768408B (zh)
WO (1) WO2022048540A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116029977A (zh) * 2022-11-08 2023-04-28 安徽萍聚德医疗科技股份有限公司 测定胃肠动力标记物在结直肠传输时间的鉴定分析方法

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111768408B (zh) * 2020-09-01 2020-11-27 安翰科技(武汉)股份有限公司 胃肠标记物自动识别方法及识别系统
CN113658694A (zh) * 2021-07-19 2021-11-16 博迈医疗科技(常州)有限公司 一种胃肠动力评估设备及系统
CN117408998B (zh) * 2023-12-13 2024-03-12 真健康(广东横琴)医疗科技有限公司 一种体表定位标记物的分割方法及设备

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110144480A1 (en) * 2009-12-10 2011-06-16 Siemens Corporation Stent marker detection using a learning based classifier in medical imaging
CN104644202A (zh) * 2013-11-20 2015-05-27 株式会社东芝 医用图像数据处理装置、医用图像数据处理方法以及医用图像数据处理程序
CN109934276A (zh) * 2019-03-05 2019-06-25 安翰科技(武汉)股份有限公司 基于迁移学习的胶囊内窥镜图像分类系统及方法
CN111768408A (zh) * 2020-09-01 2020-10-13 安翰科技(武汉)股份有限公司 胃肠标记物自动识别方法及识别系统

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9526587B2 (en) * 2008-12-31 2016-12-27 Intuitive Surgical Operations, Inc. Fiducial marker design and detection for locating surgical instrument in images
US7564999B2 (en) * 2005-07-25 2009-07-21 Carestream Health, Inc. Method for identifying markers in radiographic images
CN100530222C (zh) * 2007-10-18 2009-08-19 清华大学 图像匹配方法
US8348831B2 (en) * 2009-12-15 2013-01-08 Zhejiang University Device and method for computer simulated marking targeting biopsy
US8308632B2 (en) * 2010-06-15 2012-11-13 Siemens Aktiengesellschaft Method and apparatus for displaying information in magnetically guided capsule endoscopy
CN102799901B (zh) * 2012-07-10 2015-07-15 陈遇春 一种多角度人脸检测方法
CN104123540B (zh) * 2014-07-15 2015-09-30 北京天智航医疗科技股份有限公司 手术机器人定位标志点自动识别方法
CN108042093A (zh) * 2017-11-14 2018-05-18 重庆金山医疗器械有限公司 一种胶囊内镜的控制方法、装置及系统
CN109829501B (zh) * 2019-02-01 2021-02-19 北京市商汤科技开发有限公司 图像处理方法及装置、电子设备和存储介质
CN110522927A (zh) * 2019-09-03 2019-12-03 安徽萍聚德医疗科技股份有限公司 一种胃肠动力标记物胶囊

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110144480A1 (en) * 2009-12-10 2011-06-16 Siemens Corporation Stent marker detection using a learning based classifier in medical imaging
CN104644202A (zh) * 2013-11-20 2015-05-27 株式会社东芝 医用图像数据处理装置、医用图像数据处理方法以及医用图像数据处理程序
CN109934276A (zh) * 2019-03-05 2019-06-25 安翰科技(武汉)股份有限公司 基于迁移学习的胶囊内窥镜图像分类系统及方法
CN111768408A (zh) * 2020-09-01 2020-10-13 安翰科技(武汉)股份有限公司 胃肠标记物自动识别方法及识别系统

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
JIANG KUI, GENG XINGYUN: "AUTOMATIC QUANTITATIVE ANALYSIS OF MEDICAL ULTRASONIC IMAGE BASED ON HUMAN-COMPUTER INTERACTION ", COMPUTER APPLICATIONS AND SOFTWARE, vol. 26, no. 5, 15 May 2009 (2009-05-15), XP055906916 *
See also references of EP4209992A4 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116029977A (zh) * 2022-11-08 2023-04-28 安徽萍聚德医疗科技股份有限公司 测定胃肠动力标记物在结直肠传输时间的鉴定分析方法
CN116029977B (zh) * 2022-11-08 2024-03-15 安徽萍聚德医疗科技股份有限公司 测定胃肠动力标记物在结直肠传输时间的鉴定分析方法

Also Published As

Publication number Publication date
CN111768408B (zh) 2020-11-27
EP4209992A4 (en) 2024-02-28
CN111768408A (zh) 2020-10-13
WO2022048540A9 (zh) 2022-04-28
US20230306589A1 (en) 2023-09-28
EP4209992A1 (en) 2023-07-12

Similar Documents

Publication Publication Date Title
WO2022048540A1 (zh) 胃肠标记物自动识别方法及识别方法
JP4601134B2 (ja) 形状特徴に基づく欠陥検出の方法及び装置
Jin et al. Pulmonary nodule detection based on CT images using convolution neural network
US5668888A (en) Method and system for automatic detection of ribs and pneumothorax in digital chest radiographs
JPH09511077A (ja) 異なる２種の方式による画像整合および画像相関のための自動化方法およびシステム
CN106778668A (zh) 一种联合ransac和cnn的鲁棒的车道线检测方法
CN110866932A (zh) 多通道舌体边缘检测装置、方法及存储介质
JP2005296605A (ja) 放射線写真画像を診断関連領域と診断非関連領域とにセグメント化する方法
CN108510493A (zh) 医学图像内目标对象的边界定位方法、存储介质及终端
CN108510489B (zh) 一种基于深度学习的尘肺检测方法与系统
CN111652844B (zh) 一种基于数字图像区域增长的x射线缺陷检测方法及系统
US9672600B2 (en) Clavicle suppression in radiographic images
JP2002183708A (ja) ディジタル胸部画像のリブケイジ境界検出方法及び装置
CN111401102B (zh) 深度学习模型训练方法及装置、电子设备及存储介质
CN114463310A (zh) 一种dr图像增强方法及装置
CN111105427A (zh) 一种基于连通区域分析的肺部图像分割方法及系统
CN114022554A (zh) 一种基于yolo的按摩机器人穴位检测与定位方法
CN111223117A (zh) 舌面图像分割装置、方法及计算机存储介质
CN113012127A (zh) 基于胸部医学影像的心胸比测量方法
JPH11238129A (ja) 眼科用画像処理装置
US20190043198A1 (en) Automated segmentation of histological sections for vasculature quantification
CN110555850B (zh) 识别图像中肋骨区域的方法、装置、电子设备和存储介质
CN115908361A (zh) 一种口腔全景片的龋齿识别方法
CN113744849B (zh) 一种ct图像半自动标注方法和系统
CN115965603A (zh) 内窥镜图像的图像处理方法、装置、终端及可读存储介质

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: 21863604

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2021863604

Country of ref document: EP

Effective date: 20230403