CN114627056A - Real-time high-accuracy detection method for auricle deformity of child - Google Patents

Real-time high-accuracy detection method for auricle deformity of child Download PDF

Info

Publication number
CN114627056A
CN114627056A CN202210148662.8A CN202210148662A CN114627056A CN 114627056 A CN114627056 A CN 114627056A CN 202210148662 A CN202210148662 A CN 202210148662A CN 114627056 A CN114627056 A CN 114627056A
Authority
CN
China
Prior art keywords
image
point
calculating
value
deformity
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202210148662.8A
Other languages
Chinese (zh)
Other versions
CN114627056B (en
Inventor
彭雪
钟诚
陈晨
李媛媛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
First Affiliated Hospital of Army Medical University
Original Assignee
First Affiliated Hospital of Army Medical University
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 First Affiliated Hospital of Army Medical University filed Critical First Affiliated Hospital of Army Medical University
Priority to CN202210148662.8A priority Critical patent/CN114627056B/en
Publication of CN114627056A publication Critical patent/CN114627056A/en
Application granted granted Critical
Publication of CN114627056B publication Critical patent/CN114627056B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/725—Details of waveform analysis using specific filters therefor, e.g. Kalman or adaptive filters
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7253—Details of waveform analysis characterised by using transforms
    • A61B5/7257—Details of waveform analysis characterised by using transforms using Fourier transforms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
    • 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
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B2503/00—Evaluating a particular growth phase or type of persons or animals
    • A61B2503/06—Children, e.g. for attention deficit diagnosis
    • 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

Landscapes

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

Abstract

The invention relates to a real-time high-precision children auricle malformation detection method, which solves the technical problem of low real-time performance and low accuracy, and separates the image background by adopting multi-angle extraction to take the ears of children as the images including the target image, wherein each angle corresponds to one image, and one image is selected; defining the top left corner of the minimum circumscribed rectangle of the target image, intercepting an optimized image calculation center, calculating the center of gravity of the target image, translating the center of gravity of the target image to the center of the optimized image, and determining the center as a centroid point; the distance from the sampling boundary point to the centroid point is calculated to be the radius, the normalized radius sequence is calculated, and the absolute value of the difference between the normalized radii of the adjacent 2 sampling boundary points is calculated, if the absolute value is larger than the predefined threshold range, the technical scheme is considered to be possible for malformation, the problem is well solved, and the method can be used for detecting the ear malformation of children.

Description

Real-time high-accuracy detection method for auricle deformity of child
Technical Field
The invention relates to the field of detection of ear deformities of children, in particular to a real-time high-accuracy detection method for the ear pinna deformities of the children.
Background
The small ear deformity is congenital auricular development deformity, and the degree of the deformity can be greatly different. The most severe deformities are presented without ears, the lightest presenting a morphology approximating that of an auricle, but significantly less than normal. Both of these situations are rare. Most of the auricular malformations consist of small cartilage masses that collapse without the pinna morphology, and a more normal, but anteriorly and superiorly displaced, lobe. No external auditory canal and tympanum, auditory ossicles dysplasia, and hearing impairment.
The existing children ear deformity detection technology has the technical problems of low real-time performance and low accuracy. The invention provides a real-time high-accuracy detection method for auricle deformity of children, which aims to solve the technical problems.
Disclosure of Invention
The invention aims to solve the technical problems of low real-time performance and low accuracy in the prior art. The real-time high-accuracy detection method for the auricle deformity of the child has the characteristics of high real-time performance and high accuracy.
In order to solve the technical problems, the technical scheme is as follows:
a real-time high-precision children auricle deformity detection method comprises the following steps:
step S1, extracting images containing the ears of the children as target images from multiple angles, separating the image backgrounds, wherein each angle corresponds to one image, and optionally selecting one image to execute step S2;
step S2, defining (N, M) as the number of horizontal and vertical coordinate pixels of the minimum circumscribed rectangle of the target image;
step S3, intercepting the size of the image as (L, L) as an optimized image by using the top left corner vertex of the minimum circumscribed matrix of the target image;
Figure BDA0003509704480000021
rmaxthe maximum distance from the boundary point of the object to the center is K max (N/2, M/2);
step S4, calculating the center of gravity of the target image
Figure BDA0003509704480000022
Optimizing the center of the image (x)0，y0) According to a translation function
Figure BDA0003509704480000023
Translating the center of gravity of the target image to the center (x) of the optimized image0，y0) Then, will (x)0，y0) Determining the centroid point;
step S5, determining sampling boundary points by using equal pixel spacing as sampling step length, and calculating the kth sampling boundary point (x)k，yk) To the centroid point (x)0，y0) Defined as the radius:
Figure BDA0003509704480000024
step S6, calculating the normalized radius sequence as
Figure BDA0003509704480000025
Calculating the absolute value delta r of the difference between the normalized radii of every two adjacent 2 sampling boundary pointsn(m)＝|rn(k+1)-rn(k) Where k is a positive integer greater than 1, |, m ═ 0,1,. k-1, and Δ r is determinedn(m) size, e.g. Δ rn(m) regarding the points larger than the predefined threshold range as possible deformities, and defining the 2 adjacent sampling boundary points as possible points of deformities;
and step S7, reselecting an image, and returning to step S2 until all images are traversed, wherein the possible distortion point is defined as ear distortion, otherwise, the image is normal.
The working principle of the invention is as follows: the invention changes the gene detection and the image contrast detection in the current children ear deformity detection, converts the deformity detection into the shape mutation or deletion detection, and improves the detection efficiency and the real-time property. On the basis, the shape of the profile abrupt change is detected by adopting a mode of calculating the absolute value of the difference of normalized radii of every adjacent 2 sampling boundary points, and the shape is defined as a possible distortion point. And further, the number of possible distortion points is counted and calculated, a threshold value is preset, and the detection accuracy is improved.
In the above scheme, for optimization, further, the method for detecting ear deformity of children further includes:
and step S8, calculating the deformity possibility rate after all the images are detected, wherein the deformity possibility rate is larger than a predefined threshold value and is regarded as the ear deformity of the child.
Further, step S1 further includes performing image enhancement filtering on the image, including:
s11, use (-1)x+yMultiplying the input image F (x, y) by the obtained image F (x, y), performing central transformation to complete image preprocessing, and then calculating Fourier transformation to obtain F (u, v);
s12, multiplying F (u, v) by H (u, v) to complete filtering G (u, v) × F (u, v);
Figure BDA0003509704480000031
wherein D (u, v) is the distance from the Fourier transform origin, and σ is the Gaussian curve expansion coefficient;
s13, performing Fourier inversion;
s14, taking the real part of the Fourier inverse transformation, multiplying by (-1)x+yImage post-processing is completed to obtain an enhanced image f' (x, y).
Further, step S6 further includes:
step S61, taking any possible distortion point as a midpoint, intercepting an image with the size of N multiplied by M as a thin image, dividing the thin image into blocks with the size of w multiplied by w, and calculating the pixel mean value avg _ local and the mean square error local of each block;
Figure BDA0003509704480000041
Figure BDA0003509704480000042
blob (i, j) refers to the pixel value of the pixel point (i, j), wherein i is less than or equal to M, and j is less than or equal to N;
step S62, comparing the mean square deviation value local of each block with a predefined threshold value, defining k1 times of the maximum value of each block as the threshold value, if the mean square deviation value local is greater than the threshold value, taking the value 1, otherwise taking the value 0, and traversing to obtain a logic matrix, wherein k1 is a constant;
step S63, defining a closed operation coefficient and a corrosion operation coefficient, initially performing closed operation on the logic matrix, and then performing corrosion operation;
step S64, calculating the number of non-0 points in the dot product of the logic matrix after corrosion operation and the filtered image, adjusting the closing operation coefficient and the corrosion operation coefficient if the number of the non-0 points is more than or equal to a predefined threshold value, and returning to execute the step S63; when the number of the non-0 points is smaller than a predefined threshold value, recording the position of the current non-0 point and the gray value of the corresponding pixel point;
step S65, calculating the direction field of the thin image, and calculating the gradient of each pixel point (i, j) as
Figure BDA0003509704480000043
Calculating the local direction FX (i, j) of each block, and combining the local directions of all the blocks into a fine image direction field map;
Figure BDA0003509704480000044
Figure BDA0003509704480000051
Figure BDA0003509704480000052
step S66, finding out the area with the angle less than or equal to the angle pi/2 in the fine image direction field image, and setting the value of the area position as 1, otherwise setting the value as 0, and obtaining a corresponding logic matrix;
step S67, calculating the boundary of the logic matrix corresponding to the thin image direction field map, performing point multiplication operation on the obtained matrix and the logic matrix of the pixel position in the step S64 to obtain a new logic matrix, performing point multiplication operation on the new logic matrix and the image which is in accordance with the filtering, calculating the position and the gray value of the maximum pixel point, and defining the point as an alternative shape point;
step S68, updating the angle to be pi/2 +/-g multiplied by delta angle, repeating the steps S66-S67 to obtain a candidate shape point set, wherein delta angle is any angle number, and g is an integer;
step S69, classifying the alternative shape points according to distance, calculating the average value of the alternative shape point positions in the classification with the point number larger than the predefined threshold value, and determining the maximum alternative shape point in the class with the maximum average value as the coordinate of the shape point of the thin image
The invention has the beneficial effects that: the invention changes the gene detection and the image contrast detection in the current ear deformity detection of children, converts the deformity detection into the shape mutation or deletion detection, and improves the detection efficiency and the real-time property. On the basis, the shape of the profile abrupt change is detected by adopting a mode of calculating the absolute value of the difference of normalized radii of every adjacent 2 sampling boundary points, and the shape is defined as a possible distortion point. And further, the number of possible distortion points is counted and calculated, a threshold value is preset, and the detection accuracy is improved. And meanwhile, image enhancement filtering is carried out on the image, so that the image quality is improved, and the detection precision is increased. Finally, the high-precision children deformed ear detection is realized by accurately positioning, intercepting and detecting the possibly distorted area.
Drawings
The invention is further illustrated by the following examples in conjunction with the drawings.
Fig. 1 is a schematic view of a method for detecting ear deformity in children in example 1.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Example 1
The embodiment provides a real-time high-accuracy detection method for auricle deformity of children, which is characterized in that: the real-time high-accuracy detection method for the auricle deformity of the child comprises the following steps:
step S1, extracting images containing the ears of the children as target images from multiple angles, separating the image backgrounds, wherein each angle corresponds to one image, and optionally selecting one image to execute step S2;
step S2, defining (N, M) as the number of horizontal and vertical coordinate pixels of the minimum circumscribed rectangle of the target image;
step S3, intercepting the size of the image as (L, L) as an optimized image by using the top left corner vertex of the minimum circumscribed matrix of the target image;
Figure BDA0003509704480000061
rmaxthe maximum distance from the boundary point of the object to the center is K max (N/2, M/2);
step S4, calculating the center of gravity of the target image
Figure BDA0003509704480000071
Optimizing the center of the image (x)0，y0) According to a translation function
Figure BDA0003509704480000072
Translating the center of gravity of the target image to the center (x) of the optimized image0，y0) Then, will (x)0，y0) Determining as a centroid point;
step S5, determining sampling boundary points by using equal pixel spacing as sampling step length, and calculating the kth sampling boundary point (x)k，yk) To the centroid point (x)0，y0) Defined as the radius:
Figure BDA0003509704480000073
step S6, calculating the normalized radius sequence as
Figure BDA0003509704480000074
Calculating the absolute value delta r of the difference between the normalized radii of every two adjacent 2 sampling boundary pointsn(m)＝|rn(k+1)-rn(k) I, m is 0,1,. k-1, k is a positive integer greater than 1, and Δ r is determinedn(m) size, e.g. Δ rn(m) regarding the sample boundary points as possible distortion if the sample boundary points are larger than the predefined threshold range, and defining the 2 adjacent sample boundary points as possible distortion points;
and step S7, reselecting an image, and returning to step S2 until all images are traversed, wherein the possible distortion point is defined as ear distortion, otherwise, the image is normal.
In the embodiment, the gene detection and the image contrast detection in the existing children ear deformity detection are changed, and the deformity detection is converted into shape mutation or deletion detection, so that the detection efficiency and the real-time property are improved. On the basis, the shape of the profile abrupt change is detected by adopting a mode of calculating the absolute value of the difference of normalized radii of every adjacent 2 sampling boundary points, and the shape is defined as a possible distortion point. And further, the number of possible distortion points is counted and calculated, a threshold value is preset, and the detection accuracy is improved.
Preferably, in order to improve the detection accuracy, the method for detecting ear deformity of a child preferably further comprises:
and step S8, calculating the deformity possibility rate after all the images are detected, wherein the deformity possibility rate is larger than a predefined threshold value and is regarded as the ear deformity of the child.
In order to improve the image quality and to improve the detection accuracy, step S1 preferably further includes performing image enhancement filtering on the image, including:
s11, use (-1)x+yMultiplying the input image F (x, y) by the obtained image F (x, y), performing center transformation to complete image preprocessing, and then calculating Fourier transformation to obtain F (u, v);
s12, multiplying F (u, v) by H (u, v) to complete filtering G (u, v) × F (u, v);
Figure BDA0003509704480000081
wherein D (u, v) is the distance from the Fourier transform origin, and σ is the Gaussian curve expansion coefficient;
s13, performing Fourier inversion;
s14, taking the real part of the Fourier inverse transformation, multiplying by (-1)x+yAnd finishing image post-processing to obtain an enhanced image f' (x, y).
Preferably, in order to improve the detection accuracy, if an incoming stroke accurately detects a possible distortion point, step S6 further includes:
step S61, taking any possible distortion point as a midpoint, intercepting an image with the size of N multiplied by M to be defined as a fine image, partitioning the fine image into blocks with the size of w multiplied by w, and calculating the mean value avg _ local and the mean square error local of each block;
Figure BDA0003509704480000082
Figure BDA0003509704480000083
blob (i, j) refers to the pixel value of the pixel point (i, j), wherein i is less than or equal to M, and j is less than or equal to N;
step S62, comparing the mean square deviation value local of each block with a predefined threshold value, defining k1 times of the maximum value of each block as the threshold value, if the mean square deviation value local is greater than the threshold value, taking the value 1, otherwise taking the value 0, and traversing to obtain a logic matrix, wherein k1 is a constant;
step S63, defining a closed operation coefficient and a corrosion operation coefficient, initially performing closed operation on the logic matrix, and then performing corrosion operation;
step S64, calculating the number of non-0 points in the dot product of the logic matrix after corrosion operation and the filtered image, adjusting the closing operation coefficient and the corrosion operation coefficient if the number of the non-0 points is more than or equal to a predefined threshold value, and returning to execute the step S63; when the number of the non-0 points is smaller than a predefined threshold value, recording the position of the current non-0 point and the gray value of the corresponding pixel point;
step S65, calculating the direction field of the thin image, and calculating the gradient of each pixel point (i, j) as
Figure BDA0003509704480000091
Calculating the local direction FX (i, j) of each block, and combining the local directions of all the blocks to form a fine image direction field map;
Figure BDA0003509704480000092
Figure BDA0003509704480000093
Figure BDA0003509704480000094
step S66, finding out the area with the angle less than or equal to the angle pi/2 in the fine image direction field image, and setting the value of the area position as 1, otherwise setting the value as 0, and obtaining a corresponding logic matrix;
step S67, calculating the boundary of the logic matrix corresponding to the thin image direction field map, performing point multiplication operation on the obtained matrix and the logic matrix of the pixel position in the step S64 to obtain a new logic matrix, performing point multiplication operation on the new logic matrix and the image which is in accordance with the filtering, calculating the position and the gray value of the maximum pixel point, and defining the point as an alternative shape point;
step S68, updating the angle to be pi/2 +/-g multiplied by delta angle, repeating the steps S66-S67 to obtain a candidate shape point set, wherein delta angle is any angle number, and g is an integer;
step S69, classifying the alternative shape points according to distance, calculating the average value of the alternative shape point positions in the classification with the point number larger than the predefined threshold value, and determining the maximum alternative shape point in the class with the maximum average value as the coordinates of the thin image shape point
In the embodiment, the gene detection and the image contrast detection in the existing children ear deformity detection are changed, and the deformity detection is converted into shape mutation or deletion detection, so that the detection efficiency and the real-time property are improved. On the basis, the shape of the profile abrupt change is detected by adopting a mode of calculating the absolute value of the difference of normalized radii of every adjacent 2 sampling boundary points, and the shape is defined as a possible distortion point. And further, the number of possible distortion points is counted and calculated, a threshold value is preset, and the detection accuracy is improved. And meanwhile, image enhancement filtering is carried out on the image, so that the image quality is improved, and the detection precision is increased. Finally, the high-precision children deformed ear detection is realized by accurately positioning, intercepting and detecting the possibly distorted area.
Although the illustrative embodiments of the present invention have been described above to enable those skilled in the art to understand the present invention, the present invention is not limited to the scope of the embodiments, and it is apparent to those skilled in the art that all the inventive concepts using the present invention are protected as long as they can be changed within the spirit and scope of the present invention as defined and defined by the appended claims.

Claims (4)

1. A real-time high-accuracy detection method for auricle deformity of children is characterized by comprising the following steps: the real-time high-accuracy detection method for the auricle deformity of the child comprises the following steps:
step S1, extracting images containing the ears of the children as target images from multiple angles, separating the image backgrounds, wherein each angle corresponds to one image, and optionally selecting one image to execute step S2;
step S2, defining (N, M) as the number of horizontal and vertical coordinate pixels of the minimum circumscribed rectangle of the target image;
step S3, intercepting the size of the image as (L, L) as an optimized image by using the top left corner vertex of the minimum circumscribed matrix of the target image;
Figure FDA0003509704470000011
rmaxis the maximum distance from the boundary point to the center of the object, K ═ max(N/2，M/2)；
Step S4, calculating the center of gravity of the target image
Figure FDA0003509704470000012
Optimizing center of image (x)0，y0) According to a translation function
Figure FDA0003509704470000013
Translating the center of gravity of the target image to the center (x) of the optimized image0，y0) Then, will (x)0，y0) Determining the centroid point;
step S5, determining sampling boundary points by using equal pixel spacing as sampling step length, and calculating the kth sampling boundary point (x)k，yk) To the centroid point (x)0，y0) Defined as the radius:
Figure FDA0003509704470000014
step S6, calculating the normalized radius sequence as
Figure FDA0003509704470000015
Calculating the absolute value delta r of the difference between the normalized radiuses of every two adjacent 2 sampling boundary pointsn(m)＝|rn(k+1)-rn(k) I, m is 0,1,. k-1, k is a positive integer greater than 1, and Δ r is determinedn(m) size, e.g. Δ rn(m) regarding the sample boundary points as possible distortion if the sample boundary points are larger than the predefined threshold range, and defining the 2 adjacent sample boundary points as possible distortion points;
and step S7, reselecting an image, and returning to step S2 until all images are traversed, wherein the possible distortion point is defined as ear distortion, otherwise, the image is normal.
2. The method for detecting the auricular deformity of the child in real time with high accuracy according to claim 1, wherein: the method for detecting the ear deformity of the child further comprises the following steps:
and step S8, calculating the deformity possibility rate after all the images are detected, wherein the deformity possibility rate is larger than a predefined threshold value and is regarded as the ear deformity of the child.
3. The method for detecting the auricular deformity of the child in real time with high accuracy according to claim 1, wherein: step S1 further includes image enhancement filtering the image, including:
s11, use (-1)x+yMultiplying the input image F (x, y) by the obtained image F (x, y), performing central transformation to complete image preprocessing, and then calculating Fourier transformation to obtain F (u, v);
s12, multiplying F (u, v) by H (u, v) to complete filtering G (u, v) × F (u, v);
Figure FDA0003509704470000021
wherein D (u, v) is the distance from the Fourier transform origin, and σ is the Gaussian curve expansion coefficient;
s13, performing Fourier inversion;
s14, taking the real part of the Fourier inverse transformation, multiplying by (-1)x+yImage post-processing is completed to obtain an enhanced image f' (x, y).
4. The real-time high-accuracy detection method for the auricle deformity of the child according to claim 3, which comprises the following steps: step S6 further includes:
step S61, taking any possible distortion point as a midpoint, intercepting an image with the size of N multiplied by M as a thin image, dividing the thin image into blocks with the size of w multiplied by w, and calculating the pixel mean value avg _ local and the mean square error local of each block;
Figure FDA0003509704470000031
Figure FDA0003509704470000032
blob (i, j) refers to a pixel value of a pixel point (i, j), wherein i is less than or equal to M, and j is less than or equal to N;
step S62, comparing the mean square deviation value local of each block with a predefined threshold value, defining k1 times of the maximum value of each block as the threshold value, if the mean square deviation value local is greater than the threshold value, taking the value 1, otherwise taking the value 0, and traversing to obtain a logic matrix, wherein k1 is a constant;
step S63, defining a closed operation coefficient and a corrosion operation coefficient, initially performing closed operation on the logic matrix, and then performing corrosion operation;
step S64, calculating the number of non-0 points in the dot product of the logic matrix after corrosion operation and the filtered image, adjusting the closing operation coefficient and the corrosion operation coefficient if the number of the non-0 points is more than or equal to a predefined threshold value, and returning to execute the step S63; when the number of the non-0 points is smaller than a predefined threshold value, recording the position of the current non-0 point and the gray value of the corresponding pixel point;
step S65, calculating the direction field of the thin image, and calculating the gradient of each pixel point (i, j) as
Figure FDA0003509704470000033
Calculating the local direction FX (i, j) of each block, and combining the local directions of all the blocks into a fine image direction field map;
Figure FDA0003509704470000034
Figure FDA0003509704470000035
Figure FDA0003509704470000041
step S66, finding out the area with the angle less than or equal to pi/2 in the fine image direction field image, and setting the value of the area position as 1, otherwise setting the value as 0, and obtaining a corresponding logic matrix;
step S67, calculating the boundary of the logic matrix corresponding to the thin image direction field map, performing point multiplication operation on the obtained matrix and the logic matrix of the pixel position in the step S64 to obtain a new logic matrix, performing point multiplication operation on the new logic matrix and the image which is in accordance with the filtering, calculating the position and the gray value of the maximum pixel point, and defining the point as an alternative shape point;
step S68, updating the angle to be pi/2 +/-g multiplied by delta angle, repeating the steps S66-S67 to obtain a candidate shape point set, wherein delta angle is any angle number, and g is an integer;
and step S69, classifying the alternative shape points according to distance, calculating the average value of the alternative shape point positions in the classification with the point number larger than the predefined threshold value, and determining the maximum alternative shape point in the class with the maximum average value as the shape point coordinate of the thin image.
CN202210148662.8A 2022-02-18 2022-02-18 Real-time high-precision children auricle deformity detection method Active CN114627056B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210148662.8A CN114627056B (en) 2022-02-18 2022-02-18 Real-time high-precision children auricle deformity detection method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210148662.8A CN114627056B (en) 2022-02-18 2022-02-18 Real-time high-precision children auricle deformity detection method

Publications (2)

Publication Number Publication Date
CN114627056A true CN114627056A (en) 2022-06-14
CN114627056B CN114627056B (en) 2024-04-02

Family

ID=81900665

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210148662.8A Active CN114627056B (en) 2022-02-18 2022-02-18 Real-time high-precision children auricle deformity detection method

Country Status (1)

Country Link
CN (1) CN114627056B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101322969A (en) * 2008-07-18 2008-12-17 中国农业大学 Test and classification method
US20210004957A1 (en) * 2018-03-26 2021-01-07 Pediametrix Inc. Systems and methods of measuring the body based on image analysis

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101322969A (en) * 2008-07-18 2008-12-17 中国农业大学 Test and classification method
US20210004957A1 (en) * 2018-03-26 2021-01-07 Pediametrix Inc. Systems and methods of measuring the body based on image analysis

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
王艳;: "儿童摔倒行为图像检测数学建模方法仿真", 计算机仿真, no. 11, 15 November 2016 (2016-11-15) *

Also Published As

Publication number Publication date
CN114627056B (en) 2024-04-02

Similar Documents

Publication Publication Date Title
CN112424828B (en) Nuclear fuzzy C-means quick clustering algorithm integrating space constraint
CN110334762B (en) Feature matching method based on quad tree combined with ORB and SIFT
CN110532946B (en) Method for identifying axle type of green-traffic vehicle based on convolutional neural network
CN109086777B (en) Saliency map refining method based on global pixel characteristics
CN110008844B (en) KCF long-term gesture tracking method fused with SLIC algorithm
CN110910421A (en) Weak and small moving object detection method based on block characterization and variable neighborhood clustering
CN111583279A (en) Super-pixel image segmentation method based on PCBA
CN110147816B (en) Method and device for acquiring color depth image and computer storage medium
CN111553348A (en) Anchor-based target detection method based on centernet
CN112232351A (en) License plate recognition system based on deep neural network
CN116030396B (en) Accurate segmentation method for video structured extraction
CN111160107B (en) Dynamic region detection method based on feature matching
JP2022166799A (en) Method for identifying edge based on deep learning
CN114612550A (en) Circle fitting algorithm based on multistage optimization
CN114492619A (en) Point cloud data set construction method and device based on statistics and concave-convex property
CN109978858B (en) Double-frame thumbnail image quality evaluation method based on foreground detection
CN110363197B (en) Video region of interest extraction method based on improved visual background extraction model
CN114627056A (en) Real-time high-accuracy detection method for auricle deformity of child
CN106295716A (en) A kind of movement of traffic objective classification method based on video information and device
CN110060290B (en) Binocular parallax calculation method based on 3D convolutional neural network
CN115331008A (en) End-to-end target detection method based on target probability density graph
CN112508168B (en) Frame regression neural network construction method based on automatic correction of prediction frame
CN113344947B (en) Super-pixel aggregation segmentation method
CN115147751A (en) Method for counting station passenger flow in real time based on video image
CN115222652A (en) Method for identifying, counting and centering end faces of bundled steel bars and memory thereof

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant