CN111539972A - Method for segmenting earthworm part in ultrasound image - Google Patents
Method for segmenting earthworm part in ultrasound image Download PDFInfo
- Publication number
- CN111539972A CN111539972A CN202010335271.8A CN202010335271A CN111539972A CN 111539972 A CN111539972 A CN 111539972A CN 202010335271 A CN202010335271 A CN 202010335271A CN 111539972 A CN111539972 A CN 111539972A
- Authority
- CN
- China
- Prior art keywords
- point
- points
- segmentation
- division
- image
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/181—Segmentation; Edge detection involving edge growing; involving edge linking
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- 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
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- 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
- G06T2207/30016—Brain
Abstract
The invention discloses a method for dividing cerebellum lumbricus in an ultrasonic image, which comprises the following steps: s1, acquiring a CT image of the lumbricus cerebellalis and converting the CT image into a gray-scale image; s2, reading the edge coordinates of the gray-scale image, and taking the center point of the gray-scale image as the center of a circle to obtain the coordinates of all pixel points on a line segment formed from the center of the circle to the edge point of the image at each angle within 0-360 degrees and the corresponding pixel values on the image; s3, finding the initial segmentation point on the ray formed by each angle with the outward center of the circle, S4, carrying out error judgment on the roughly adjusted segmentation point by adopting an error point judgment mode: searching a single-peak error point, carrying out error judgment on the division points around the single-peak error point, and S5, adjusting all points in the array to be accurately adjusted to obtain accurate division points; and S6, sequentially connecting the adjusted accurate division points to form the lumbricus cerebellar margin image.
Description
Technical Field
The invention relates to the field of medical image processing, in particular to a method for segmenting small brain earthworm parts in an ultrasonic image.
Background
With the widespread use of computer technology in medicine, earlier manually segmented images are gradually replaced by computer automated segmentation. At present, for how to detect whether the lumbricus cerebellar of a fetus develops normally, the method generally adopts the steps of manually segmenting the area of the lumbricus cerebellar in a CT image and calculating the area of the lumbricus cerebellar to detect whether the lumbricus cerebellar develops normally. How to simplify image segmentation and make the segmentation effect finer becomes a key problem. At present, for the problem of splitting the cerebellar lumbricus of the CT image, no known effective computer automatic processing method exists, and the CT image is manually split. This requires a professional knowledge base, takes much time, and the effect is not necessarily subtle, so accuracy cannot be guaranteed.
Disclosure of Invention
According to the current situation that the division of the earthworm part of cerebellum does not exist in the prior art, the invention discloses a method for dividing the earthworm part of cerebellum in an ultrasonic image, which specifically comprises the following steps:
acquiring a CT image of lumbricus cerebelli and converting the CT image into a gray-scale image;
reading the edge coordinates of the gray-scale image, taking the central point of the gray-scale image as the center of a circle, and calculating the coordinates of all pixel points on a line segment formed from each angle to the edge point of the image within 0-360 degrees of the center of the circle and the corresponding pixel values on the image;
searching an initial segmentation point on a ray formed by each angle with the center of the circle outward, performing error rough judgment on the obtained initial segmentation point and a plurality of circles of segmentation points formed by angle values at intervals in an error point judgment mode, modifying error points, and removing miscellaneous points to obtain rough adjustment segmentation points;
carrying out error judgment on the roughly adjusted segmentation points by adopting an error point judgment mode, searching single-peak error points, carrying out error judgment on segmentation points around the single-peak error points so as to re-determine segmentation points to be deleted, and storing all segmentation points to be deleted into an array to be accurately adjusted;
adjusting all points in the array to be accurately adjusted to obtain accurate segmentation points;
and connecting the adjusted accurate division points in sequence to form the lumbricus cerebellar edge image.
Further, when searching for the segmentation point on the ray: setting interval angle values, and respectively taking a line segment corresponding to each integer angle as a start to search a segmentation point on the angle line segment at intervals of the interval angle values;
the following method is adopted when the segmentation point is selected on each angle line segment: firstly, obtaining the number of all points on the angle ray, accumulating the number of background points on the ray [ length/3, length/2] area, if the number is less than length/6 and the pixel value at length/2 is greater than a set threshold value A, adopting a judgment interval starting value of length/2 to search for a division point, otherwise adopting length/3 to search for the division point; if the pixel value of one point is found to be smaller than a set threshold A in sequence from the initial position of the judgment interval to the edge point on the line segment, the next pixel is regarded as a background pixel, the point is set as an initial segmentation point, and otherwise, the point with the minimum pixel value is selected as the initial segmentation point;
and adjusting the obtained initial segmentation point in a mode of error point judgment: taking the interval angle value as an area, respectively taking the division point corresponding to each integer angle as a start, defining a circle of division points formed by every interval angle value as a division point circumference, and carrying out error judgment on the division point circumference:
if the current point (x)j,yj) To the central point O (x)0,y0) Is smaller than the distance from the previous dividing point and the next dividing point on the circumference of the dividing point to the central point, and the distance difference is larger than a set threshold B, the current point is considered (x)j,yj) Is an error point, the current point (x) is changedj,yj) Wherein the manner of change is as follows: the sum of half the distance from the previous division point to the center point and half the distance from the next division point to the center point on the circumference of the division point is taken as the current point (x)j,yj) A distance M from the center point;
if the current point (x)j,yj) To the central point O (x)0,y0) Is greater than the distance from the previous dividing point and the next dividing point on the circumference of the dividing point to the central point, and the distance difference is greater than a set threshold value B, the current point (x) is considered to bej,yj) Is an error point, the current point (x) is changedj,yj) Wherein the manner of change is as follows: the sum of half of the distance from the previous dividing point to the central point and half of the distance from the next dividing point to the central point on the circumference of the dividing point is taken as the current point (x)j,yj) The distance M to the center point, and (x) is adjusted according to the obtained distance Mj,yj) The new point coordinates of (2) are used as rough adjustment division points.
Further, the rough adjustment division points are subjected to error adjustment again, namely 360 division points on the circumference of the division points are subjected to error judgment in sequence in an error point judgment mode and the error points are modified;
and (3) carrying out unimodal error point elimination on the modified obtained segmentation points:
defining the average value of complementary angles of vector included angles formed by all three adjacent rough adjustment segmentation points as an angle threshold TθDefining Q times of the average value of the distances from all the rough adjustment division points to the central point as a chord height threshold Ts(ii) a Calculating the distance d from each point to the central point for each of 360 roughly adjusted segmentation points on the circumference of the segmentation point, if the distance from each point to the central point is larger than the chord height threshold TsAnd the angle is greater than an angle threshold TθIf yes, storing the point in a list to be deleted;
judging in the list to be deleted, and re-determining the segmentation points to be deleted on the basis of the list to be deleted: if the distance difference between a certain division point and the next division point needing to be deleted to the central point is larger than a set threshold value B, deleting the certain division point, the next division point needing to be deleted and the middle division point of the certain division point and the next division point needing to be deleted; if the distance difference between a certain division point and the previous division point needing to be deleted and the central point is larger than a set threshold value B, and the division points behind the certain division point are continuously the division points needing to be deleted, the division points are deleted at the moment, and all the division points needing to be deleted are stored in the array to be accurately adjusted.
Further, deleting repeated segmentation points in the array to be accurately adjusted, sequencing the segmentation points according to the sequence of angles if the current segmentation point (x) isr,yr) Previous phase in raw dataIf the adjacent division point is not consistent with the previous division point in the array to be accurately adjusted, the current division point (x) is setr,yr) Distance t to the center pointjChanging to the distance t from the previous dividing point to the central pointj-1According to the newly acquired distance tj-1Calculating the current segmentation point (x)r,yr) New coordinates of (c), will be the current segmentation point (x)r,yr) Removing the array to be accurately adjusted; and circularly processing the division points in the array to be accurately adjusted in the above mode until no member exists in the array.
Due to the adoption of the technical scheme, according to the method for dividing the earthworm cerebellum in the ultrasonic image, the required edge of the earthworm cerebellum can be obtained only by giving the CT image of the earthworm cerebellum to be divided by a user, and the operation that a doctor manually marks the edge in daily life is reduced, so that errors caused by manual errors are reduced, the accuracy and the efficiency are improved, the operation can be carried out without a corresponding medical technology, the occupied memory is small, and the operation time is short.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments described in the present application, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flow chart of a method implementation of the present invention;
FIG. 2 is a diagram illustrating the effect of converting an image into a gray scale image according to the present invention;
FIG. 3 is a schematic diagram illustrating an image edge point calculation according to the present invention;
FIG. 4 is a diagram illustrating single-peak error point rejection in the present invention;
FIG. 5 is a diagram illustrating adding deletion points in a deletion list according to the present invention;
fig. 6 is a diagram showing the effect of the division of the lumbricus cerebellar CT image in the present invention.
Detailed Description
In order to make the technical solutions and advantages of the present invention clearer, the following describes the technical solutions in the embodiments of the present invention clearly and completely with reference to the drawings in the embodiments of the present invention:
in the implementation process of the method for dividing the lumbricus cerebellum part in the ultrasonic image shown in fig. 1, the image is converted into an effect graph of a gray scale graph shown in fig. 2, then a circle of 360-degree edge points of the image are calculated, the division points corresponding to the angle are found between the center of the image and each edge point, then the error points are continuously adjusted to be closer to the real condition, and finally the division points are displayed. The method disclosed by the invention comprises the following specific steps:
s1: inputting a CT image of the lumbricus cerebelli, and converting the image into a gray scale image, wherein the method specifically comprises the following steps:
pixel value (r) of each pixel point of the traversal imagek,gk,bk) Num, num is the number of pixel points, and the calculated gray value is:
grayk＝0.299*rk+0.587*gk+0.114*bk
rk＝grayk，gk＝grayk，bk＝grayk
by processing each pixel point, the image can be converted into a gray image. The specific effect is shown in fig. 2.
S2: calculating the edge point coordinates of one circle of the edge of the gray scale map obtained in the step S1; and (3) with the center point of the image as the center of a circle, calculating the coordinates of all pixel points on a line segment formed from the center point of the circle to the edge point of the image at each angle of 0-360 degrees (excluding 360 degrees), and the corresponding pixel values (with 1 as a step length) on the image, and storing the two values. The method specifically adopts the following steps:
s21: obtaining the edge point coordinates of one circle of the edge of the gray-scale image:
the width of the image is width, the height is height, so the coordinate of the center point center of the image is (width/2, height/2), and then the edge point of the image corresponding to each angle with the center point as the center point and the center point of 0-360 degrees (excluding 360 degrees) is calculated.
The image is schematically shown in FIG. 3, O (x)0,y0) Defining OO 'as ray whose centre is O and angle α is 0 deg. (or 360 deg.), starting with OO', clockwise once by 1 deg., 2 deg. and rotating for one turn, and dividing image into 1, 2, 3 and 4 four regions, in the first region, α is in the range of 0 deg. - α<45°∪315°≤α<360 degrees, and the coordinate (x, y) of the edge point is x-width;
y＝y0+(width-x0)*tanα；
in the second region, the range of α is: alpha is more than or equal to 45 degrees and less than 135 degrees, and the coordinates (x, y) of the edge points are
x＝x0+(height-y0)*cotα；
y＝height；
In the third region, the range of α is: alpha is more than or equal to 135 degrees and less than 225 degrees, and the coordinates (x, y) of the edge points are
x＝0
y＝y0-x0*tanα
In the fourth region, the range of α is: alpha is more than or equal to 225 degrees and less than 315 degrees, and the coordinates (x, y) of the edge points are
x＝x0-y0*cotα；
y＝0
S22: in this way, the edge point coordinates of the image corresponding to each angle from 0 to 360 ° (excluding 360 °) of the center can be found.
S23: and obtaining a connecting line between the two points through the circle center and the obtained edge point, thereby obtaining line segments at all angles. Next, we obtain the coordinates of all the pixels on each line segment and the corresponding pixel values (with 1 as a step length) on the image.
At the center point is O (x)0,y0) And on the line segment corresponding to the angle α, when the step length is t, the coordinates (x, y) of the pixel point at this time are:
(x,y)＝(x0+t*cosα,y0+t*sinα) (1)
here, since the pixel point on the image is discrete (the coordinate value of the pixel point is an integer value), the obtained (x, y) may be a non-integer value, and in this case, there is no pixel value corresponding to the coordinate point. For this case, we use a bilinear interpolation (INTER _ LINEAR) method to find an approximate pixel value for the coordinate point. The specific mode is as follows:
firstly, edge expansion is carried out on the image, and the image is expanded up, down, left and right according to values x on four edges0,0...x0,n-1,x1,n-1...xm-1,n-1,xm-1,n-2...xm-1,0,xm-2,0...x1,0(where m and n are the column height and row width of the matrix, respectively). The expanded image data is stored as im array, and the specific expansion mode is as follows
Example (c): the original pixel matrix data is
After expansion transformation (1 row up, 2 rows down, 2 columns left, 1 column right), new image data are obtained
(wherein "1" in the third column of the second row, "2" in the fourth column of the second row, "3" in the third column of the third row, "4" in the fourth column of the third row are original data, and are extended outward by edge values of the matrix, and the rest of the values are values after the extension).
Then, an approximate pixel value (x, y) on the image is obtained from the extended image im.
(where i is an integer number of x, u is a decimal number of x, j is an integer number of y, and v is a decimal number of y)
After this step, the coordinates of all points on the line segment corresponding to each angle i are stored in a lines [ i ] array, the pixel values of all points on the line segment corresponding to each angle i are stored in a ys [ i ] array, and the serial numbers of all points on the line segment corresponding to each angle i are stored in an xs [ i ] array.
S3: at the point of the circle O (x)0,y0) Finding an initial segmentation point on a line segment formed by each outward angle i, carrying out error rough judgment on the obtained initial segmentation point by adopting an error point judgment mode according to a plurality of circles of segmentation points formed by angle values of every interval, modifying an error point, and removing miscellaneous points to obtain a rough adjustment segmentation point, wherein the following mode is specifically adopted:
s31, setting an interval angle value (the angle value can be 12 degrees), and respectively taking a line segment corresponding to each integer angle as a start to search a segmentation point on the angle line segment every interval of the interval angle value; taking 12 degrees as an interval, starting with a line segment corresponding to 0 degree for 360 obtained line segments, then searching for segmentation points on the angle line segment at intervals of 12 degrees, and then respectively taking 1 degree, 2 degrees and 3 degrees. . . And taking the corresponding line segment as the starting point, performing the operation until the line segment corresponding to 11 degrees is the starting point, and repeating the operation. Namely, the method is divided into 12 sections, and the line segments on each section are processed respectively, wherein the following mode is adopted when a dividing point is selected on each angle line segment:
s311: obtaining the number of all points on the angle ray, using the number of the points on the angle ray as a limit for judging the edge (judging and selecting the dividing point from the length to the image edge on the angle), finding that 1/3 or 1/2 of the total length has better effect on selecting the edge value for different pictures through a plurality of experiments, so accumulating the number of background points on the ray [ length/3, length/2] area, if the number is less than length/6 and the pixel value at the length/2 is greater than a threshold value A (here, the threshold value A is 60, that is, the area is not all background pixels, considering that the situation is the situation that a small part of background exists in the middle, the real dividing point is at the rear position, and the boundary is the target pixel), adopting the length/2 of the judgment interval to search the dividing point, otherwise adopting the length/3 to search the dividing point, at this point in the angle, the desired segmentation point is too close to the center point.
S312: an initial segmentation point is selected on each angular ray. Judgment conditions of the initial segmentation points: if the pixel value of a point is found to be smaller than a set threshold value A (the threshold value A can be 60) in sequence from the initial position of the judgment interval to the edge point on the line segment, the following pixel is considered as a background pixel, the point is set as an initial segmentation point, and otherwise, the point with the minimum pixel value is selected as the initial segmentation point.
The coordinates of all initial segmentation points after this step are saved in the line points array.
And S32, adjusting the obtained initial segmentation points in an error point judgment mode to enable the initial segmentation points to be more accurately positioned at the edge positions of the real lumbricus cerebellar. The method specifically adopts the following steps:
taking the interval angle value as an area, respectively taking the division point corresponding to each integer angle as a start, defining a circle of division points formed by every interval angle value as a division point circumference, and carrying out error judgment on a circle of division points formed by every interval angle value:
if the current point (x)j,yj) To the central point O (x)0,y0) Is simultaneously less than the previous division point (x) on the circumference of the division pointj-1,yj-1) And a latter dividing point (x)j+1,yj+1) The distance to the center point and the distance difference is larger than a set threshold B (where the threshold B is taken to be 8) is regarded as the current point (x)j,yj) Is an error point, the current point (x) is changedj,yj) Wherein the distance is calculated by the formula,
the position is changed as follows: half of the distance from the previous division point on the circumference of the division point to the central point andthe sum of the distance from the latter dividing point to the center point by half is taken as the current point (x)j,yj) Distance to centre point, i.e. tj＝(tj-1+tj+1)/2；
If the current point (x)j,yj) To the central point O (x)0,y0) Is greater than the distances from the previous and next division points on the division point circumference to the center point, and the distance difference is greater than a set threshold B (which may be 8), the current point (x) is considered to be the current point (x)j,yj) Is an error point, the current point (x) is changedj,yj) Wherein the manner of change is as follows: the sum of half of the distance from the previous dividing point to the central point and half of the distance from the next dividing point to the central point on the circumference of the dividing point is used as the current point (x)j,yj) The distance M to the center point is calculated by the formula (1) according to the newly obtained distancejThe coordinates of the lower points are used as rough adjustment division points. Adjusting (x) according to the obtained distance Mj,yj) As a coarse adjustment division point.
But due to previously saved ys i]In all points on the line segment corresponding to each angle i in the array, each step is an integer, but the distance t obtained at the momentjNot necessarily an integer, and so is at the value tj-4～tj+4A suitable point is taken as the rough adjustment segmentation point, and the specific selection mode is the same as the step in S312.
And S4, performing error judgment on the roughly adjusted segmentation points by adopting an error point judgment mode, searching single-peak error points, performing error judgment on segmentation points around the single-peak error points, re-determining points to be deleted, and storing all segmentation points to be deleted into an array to be accurately adjusted. The method specifically adopts the following steps:
s41, performing error adjustment again on the roughly adjusted segmentation points, namely sequentially performing error judgment on 360 segmentation points on the circumference of the segmentation points in a mode of error point judgment and modifying the error points;
and S42, performing unimodal error point elimination on the modified obtained segmentation points, as shown in the figure 4. The method comprises the following specific steps:
① setting angle threshold TθChord height threshold Ts；
② defining the average value of the supplementary angles of the vector angles formed by all the adjacent front and back coarse adjustment division points as the angle threshold value
Wherein
(xm,ym)、(xm+1,ym+1) The division points are the previous point, the current point and the next point respectively, wherein m-1 is (m-1+ 360)% 360, and m +1 is (m +1+ 360)% 360.
③ define the chord height threshold as the Q times the average (i.e., the sample standard deviation) of the distances from all the coarse adjustment segmentation points to the center point (here we take Q to be 0.75)
Multiple tests prove that the obtained filtered center point sequence is ideal in distribution.
Fourthly, calculating the distance d from each point of the 360 roughly adjusted segmentation points on the circumference of the segmentation points to the central point;
⑤ if the distance from each point to the center is greater than the chord height threshold TsAnd the angle is greater than an angle threshold TθThen the point is stored in the list to be deleted.
S43, judging in the list to be deleted, re-determining the division points to be deleted on the basis of the list to be deleted, and storing all the division points to be deleted in the array to be accurately adjusted:
<1> if the distance difference between a certain division point and its next division point to be deleted to the center point is larger than a set threshold value B, the certain division point and its next division point to be deleted and the middle division point of the two division points are deleted, because the two division points are equivalent to one division point located at the peak or the bottom of the wave, the middle division points are error points. As shown in fig. 5, the points between the division point P and the next division point Q to be deleted are added to the array to be adjusted accurately.
<2> if the distance difference between a certain division point and the center point of the previous division point needing to be deleted is larger than a set threshold value B, and the division points after the certain division point are continuously the division points needing to be deleted, the division points are deleted. Since a certain division point is a peak or a trough at this time, the following division points move and meander, but are all located in the range of the error point. (since the previous segmentation point to be deleted and the segmentation point between a segmentation point have already been processed in the first step, we only need to consider the second half). As shown in fig. 5, the division point Q is a division point, the division point P is a point that needs to be deleted before, the division points between the two division points have been deleted in the previous step, and the two division points after the division point Q are added to the set of integers to be adjusted accurately.
S5, adjusting all the division points in the array to be accurately adjusted to obtain accurate division points, and sequentially connecting the adjusted accurate division points to form a cerebellum lumbricus edge image, wherein the method specifically comprises the following steps:
s51, deleting repeated segmentation points in the array to be accurately adjusted, and sequencing the segmentation points according to the angle sequence;
s52, sequentially judging the points a in the array to be accurately adjusted:
【1】 If the current segmentation point (x)r,yr) When the previous adjacent division point in the original array which is not subjected to the error point elimination processing, namely the previous division point in the real data is different from the previous division point of the array to be accurately adjusted: it means that the current segmentation point a can be adjusted according to the value of the previous segmentation point of the real data, i.e. the distance t from the segmentation point a to the central pointjChanging to the distance t from the previous dividing point to the central pointj-1According to the newly obtained distance tj-1The point at the angle and the distance is obtained by the formula (1)As new coordinates of the division points. But due to previously saved ys i]In all points on the line segment corresponding to each angle i in the array, each step is an integer, but the distance t obtained at the momentjNot necessarily an integer, and so is at the value tj-4～tj+4A suitable point is taken as the edge point, and the specific selection mode is the same as the step in S312. Therefore, the division point a is not deleted, and the division point a is removed from the array to be accurately adjusted.
【2】 And circularly processing the array to be accurately adjusted until no member exists in the array.
S6: and displaying the adjusted division points. The image and all the segmentation points are displayed using the library function in python.
As shown in fig. 6, the effect of the segmentation of one of the lumbricus cerebellar CT images is shown. The earthworm part can be seen to be well divided, particularly the shape is fitted, error points on the edge of the divided earthworm part are few, the error range is small, and almost no error exists.
The invention relates to a method for directly and respectively obtaining the lumbricus cerebellar margins of all input images without manual marking in a CT image. The method replaces the original mode that the edge of the lumbricus cerebellum of a single image can be obtained only by manual marking. According to the method, only the CT image of the lumbricus cerebellum part needs to be input by a user, so that the operation of manually marking the edge by a doctor in daily life is reduced, errors caused by manual errors are reduced, the accuracy and the efficiency are improved, the operation can be carried out without a corresponding medical technology, the occupied memory is small, and the operation time is short.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and equivalent substitutions or changes according to the technical solution and the inventive concept of the present invention should be covered by the scope of the present invention.
Claims (4)
1. A method for dividing earthworm part of cerebellum in an ultrasonic image is characterized by comprising the following steps:
acquiring a CT image of lumbricus cerebelli and converting the CT image into a gray-scale image;
reading the edge coordinates of the gray-scale image, taking the central point of the gray-scale image as the center of a circle, and calculating the coordinates of all pixel points on a line segment formed from each angle to the edge point of the image within 0-360 degrees of the center of the circle and the corresponding pixel values of the pixel points on the image;
searching an initial segmentation point on a ray formed by each angle with the center of the circle outward, performing error rough judgment on the obtained initial segmentation point and a plurality of circles of segmentation points formed by angle values at intervals in an error point judgment mode, modifying error points, and removing miscellaneous points to obtain rough adjustment segmentation points;
carrying out error judgment on the roughly adjusted segmentation points by adopting an error point judgment mode, searching single-peak error points, carrying out error judgment on segmentation points around the single-peak error points so as to re-determine segmentation points to be deleted, and storing all segmentation points to be deleted into an array to be accurately adjusted;
adjusting all points in the array to be accurately adjusted to obtain accurate segmentation points;
and connecting the adjusted accurate division points in sequence to form the lumbricus cerebellar edge image.
2. The method for dividing lumbricus cerebellum in an ultrasound image according to claim 1, further comprising: when searching for a segmentation point on this ray: setting interval angle values, and respectively taking a line segment corresponding to each integer angle as a start to search a segmentation point on the angle line segment at intervals of the interval angle values;
the following method is adopted when the segmentation point is selected on each angle line segment: firstly, obtaining the number of all points on the angle ray, accumulating the number of background points on the ray [ length/3, length/2] area, if the number is less than length/6 and the pixel value at length/2 is greater than a set threshold value A, adopting a judgment interval starting value of length/2 to search for a division point, otherwise adopting length/3 to search for the division point; if the pixel value of one point is found to be smaller than a set threshold value A in sequence from the initial position of the judgment interval to the edge point on the line segment, the next pixel is considered as a background pixel, the point is set as an initial segmentation point, and otherwise, the point with the minimum pixel value is selected as the initial segmentation point;
and adjusting the obtained initial segmentation point in a mode of error point judgment: taking the interval angle value as an area, respectively taking the division point corresponding to each integer angle as a start, defining a circle of division points formed by every interval angle value as a division point circumference, and carrying out error judgment on the division point circumference:
if the current point (x)j,yj) To the central point O (x)0,y0) Is smaller than the distance from the previous dividing point and the next dividing point on the circumference of the dividing point to the central point, and the distance difference is larger than a set threshold B, the current point is considered (x)j,yj) Is an error point, the current point (x) is changedj,yj) Wherein the manner of change is as follows: the sum of half of the distance from the previous dividing point to the central point and half of the distance from the next dividing point to the central point on the circumference of the dividing point is used as the current point (x)j,yj) A distance M from the center point;
if the current point (x)j,yj) To the central point O (x)0,y0) Is greater than the distance from the previous dividing point and the next dividing point on the circumference of the dividing point to the central point, and the distance difference is greater than a set threshold value B, the current point (x) is considered to bej,yj) Is an error point, the current point (x) is changedj,yj) Wherein the manner of change is as follows: the sum of half of the distance from the previous dividing point to the central point and half of the distance from the next dividing point to the central point on the circumference of the dividing point is used as the current point (x)j,yj) The distance M to the center point, and (x) is adjusted according to the obtained distance Mj,yj) As a coarse adjustment segmentation point.
3. The method for dividing lumbricus cerebellum in an ultrasound image according to claim 2, further comprising: carrying out error adjustment on the roughly adjusted segmentation points again, namely sequentially carrying out error judgment on 360 segmentation points on the circumference of the segmentation points in a mode of error point judgment and modifying the error points;
and (3) carrying out unimodal error point elimination on the modified obtained segmentation points:
defining the average value of complementary angles of vector included angles formed by all three adjacent rough adjustment segmentation points as an angle threshold TθDefining Q times of the average value of the distances from all the rough adjustment division points to the central point as a chord height threshold Ts(ii) a Calculating the distance d from each point to the central point for each of 360 roughly adjusted segmentation points on the circumference of the segmentation point, if the distance from each point to the central point is larger than the chord height threshold TsAnd the angle is greater than an angle threshold TθIf yes, storing the point in a list to be deleted;
judging in the list to be deleted, and re-determining the segmentation points to be deleted on the basis of the list to be deleted: if the distance difference between a certain division point and the next division point needing to be deleted to the central point is larger than a set threshold value B, deleting the certain division point, the next division point needing to be deleted and the middle division point of the certain division point and the next division point needing to be deleted; and if the distance difference between a certain division point and the previous division point needing to be deleted to the central point is larger than a set threshold value B, and the division points behind the certain division point are continuously the division points needing to be deleted, deleting the division points at the moment, and storing all the division points needing to be deleted into the array to be accurately adjusted.
4. The method as claimed in claim 3, wherein the repeated segmentation points in the array to be precisely adjusted are deleted, the segmentation points are sorted according to the sequence of angles if the current segmentation point (x) is the current segmentation pointr,yr) If the previous adjacent division point in the original data is inconsistent with the previous division point in the array to be accurately adjusted, the current division point (x) is determinedr,yr) Distance t to the center pointjChanging to the distance t from the previous dividing point to the central pointj-1According to the newly acquired distance tj-1Calculating the current segmentation point (x)r,yr) New coordinates of (c), will be the current segmentation point (x)r,yr) Removing the array to be accurately adjusted; and circularly processing the division points in the array to be accurately adjusted in the above mode until no member exists in the array.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010335271.8A CN111539972B (en) | 2020-04-24 | 2020-04-24 | Method for segmenting cerebellar lumbricus in ultrasonic image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010335271.8A CN111539972B (en) | 2020-04-24 | 2020-04-24 | Method for segmenting cerebellar lumbricus in ultrasonic image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111539972A true CN111539972A (en) | 2020-08-14 |
| CN111539972B CN111539972B (en) | 2023-04-18 |
Family
ID=71980014
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010335271.8A Active CN111539972B (en) | 2020-04-24 | 2020-04-24 | Method for segmenting cerebellar lumbricus in ultrasonic image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111539972B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112258534A (en) * | 2020-10-26 | 2021-01-22 | 大连理工大学 | Method for positioning and segmenting small brain earthworm parts in ultrasonic image |
| CN112258535A (en) * | 2020-10-26 | 2021-01-22 | 大连理工大学 | Integrated positioning and segmentation method for corpus callosum and lumbricus in ultrasonic image |
| CN115439474A (en) * | 2022-11-07 | 2022-12-06 | 山东天意机械股份有限公司 | Rapid positioning method for power equipment fault |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101170641A (en) * | 2007-12-05 | 2008-04-30 | 北京航空航天大学 | A method for image edge detection based on threshold sectioning |
| CN101334263A (en) * | 2008-07-22 | 2008-12-31 | 东南大学 | Circular target circular center positioning method |
| US20110085218A1 (en) * | 2009-10-13 | 2011-04-14 | Xerox Corporation | Vector error diffusion with edge control |
| CN103400365A (en) * | 2013-06-26 | 2013-11-20 | 成都金盘电子科大多媒体技术有限公司 | Automatic segmentation method for lung-area CT (Computed Tomography) sequence |
| CN106920244A (en) * | 2017-01-13 | 2017-07-04 | 广州中医药大学 | A kind of method of background dot near detection image edges of regions |
| CN110415257A (en) * | 2019-07-23 | 2019-11-05 | 东南大学 | A kind of biphase gas and liquid flow overlapping bubble image partition method |
-
2020
- 2020-04-24 CN CN202010335271.8A patent/CN111539972B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101170641A (en) * | 2007-12-05 | 2008-04-30 | 北京航空航天大学 | A method for image edge detection based on threshold sectioning |
| CN101334263A (en) * | 2008-07-22 | 2008-12-31 | 东南大学 | Circular target circular center positioning method |
| US20110085218A1 (en) * | 2009-10-13 | 2011-04-14 | Xerox Corporation | Vector error diffusion with edge control |
| CN103400365A (en) * | 2013-06-26 | 2013-11-20 | 成都金盘电子科大多媒体技术有限公司 | Automatic segmentation method for lung-area CT (Computed Tomography) sequence |
| CN106920244A (en) * | 2017-01-13 | 2017-07-04 | 广州中医药大学 | A kind of method of background dot near detection image edges of regions |
| CN110415257A (en) * | 2019-07-23 | 2019-11-05 | 东南大学 | A kind of biphase gas and liquid flow overlapping bubble image partition method |
Non-Patent Citations (1)
| Title |
|---|
| 史延新;: "结合边缘检测和区域方法的医学图像分割算法" * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112258534A (en) * | 2020-10-26 | 2021-01-22 | 大连理工大学 | Method for positioning and segmenting small brain earthworm parts in ultrasonic image |
| CN112258535A (en) * | 2020-10-26 | 2021-01-22 | 大连理工大学 | Integrated positioning and segmentation method for corpus callosum and lumbricus in ultrasonic image |
| CN115439474A (en) * | 2022-11-07 | 2022-12-06 | 山东天意机械股份有限公司 | Rapid positioning method for power equipment fault |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111539972B (en) | 2023-04-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111539972B (en) | Method for segmenting cerebellar lumbricus in ultrasonic image | |
| CN111445478B (en) | Automatic intracranial aneurysm region detection system and detection method for CTA image | |
| CN111583188B (en) | Surgical navigation mark point positioning method, storage medium and computer equipment | |
| CN109035170B (en) | Self-adaptive wide-angle image correction method and device based on single grid image segmentation mapping | |
| CN111091567B (en) | Medical image registration method, medical device and storage medium | |
| CN110415250B (en) | Overlapped chromosome segmentation method and device based on deep learning | |
| CN110378913B (en) | Image segmentation method, device, equipment and storage medium | |
| JPH0638274B2 (en) | Image recognition apparatus and image recognition method | |
| JP2009169925A (en) | Image retrieval device and image retrieval method | |
| CN111127487B (en) | Real-time multi-tissue medical image segmentation method | |
| US6542639B2 (en) | Region extraction apparatus, region extraction method and computer readable recording medium | |
| CN114037720A (en) | Pathological image segmentation and classification method and device based on semi-supervised learning | |
| CN115083571A (en) | Pathological section processing method, computer device and storage medium | |
| CN112200809B (en) | Adherent chromosome separation method and device based on skeleton segmentation and key point detection | |
| Somasundaram et al. | Straightening of highly curved human chromosome for cytogenetic analysis | |
| JP3819236B2 (en) | Pattern recognition method and computer-readable storage medium storing program for performing pattern recognition | |
| JP2011002965A (en) | Image retrieval method and device | |
| CN108447066B (en) | Biliary tract image segmentation method, terminal and storage medium | |
| CN115761230A (en) | Spine segmentation method based on three-dimensional image | |
| CN114266781A (en) | Defect inspection apparatus, defect inspection method, and information recording medium | |
| CN114170440A (en) | Method and device for determining image feature points, computer equipment and storage medium | |
| CN112991395A (en) | Vision tracking method based on foreground condition probability optimization scale and angle | |
| CN112545566A (en) | Fetal neck diaphragmatic layer measuring method and device | |
| CN111968178A (en) | Sub-pixel positioning method based on particle swarm algorithm | |
| CN116228779B (en) | Blood vessel center path extraction method and device |
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 |