CN115690047A - Prostate ultrasound image segmentation method and device based on abnormal point detection - Google Patents
Prostate ultrasound image segmentation method and device based on abnormal point detection Download PDFInfo
- Publication number
- CN115690047A CN115690047A CN202211350084.2A CN202211350084A CN115690047A CN 115690047 A CN115690047 A CN 115690047A CN 202211350084 A CN202211350084 A CN 202211350084A CN 115690047 A CN115690047 A CN 115690047A
- Authority
- CN
- China
- Prior art keywords
- contour
- point
- prostate
- points
- boundary
- 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.)
- Pending
Links
Images
Landscapes
- Ultra Sonic Daignosis Equipment (AREA)
Abstract
The invention discloses a prostate ultrasonic image segmentation method and a device based on abnormal point detection, wherein the method comprises the steps of obtaining an average contour representing the shape of a prostate; deforming the average contour to make the deformed contour preliminarily coincide with the real prostate boundary on the ultrasound prostate image; adjusting each point on the contour after deformation processing to a gray value mutation position on the prostate ultrasonic image through a normal vector contrast boundary algorithm to obtain a rough segmentation contour; detecting and eliminating position abnormal points on the rough segmentation contour; and sequentially connecting all points on the roughly-segmented contour after the abnormal points in the position are removed to obtain the final segmented contour of the prostate ultrasonic image. The invention has the beneficial effects that: the prostate contour is preliminarily framed through the average contour, calculation amount required by follow-up contour correction is greatly reduced, and meanwhile, the abnormal point detection algorithm is adopted to identify the abnormal point in the position, so that the calculation amount can be further reduced.
Description
Technical Field
The invention relates to the technical field of prostate ultrasonic image processing, in particular to a prostate ultrasonic image segmentation method and device based on abnormal point detection.
Background
The prostate is the largest accessory gonad peculiar to men, is chestnut-shaped, slightly flat in front and back, and is located deep in the pelvic cavity of a human body. The prostate gland is a part of the urethra, and the prostatic fluid secreted by the prostate gland is an important component of semen, so the prostate gland has an important function on the urinary control, reproductive and sexual functions of a human body. The symptoms of prostate diseases usually include frequent micturition, urgent micturition, swelling and pain of the pudendum, etc., and the symptoms may affect male reproduction and even life health. The data of the prostate incidence in China are not optimistic every year, so that the relevant treatment of the prostate diseases has great significance for improving the health level of men in China. In clinical practice, in order to observe a diseased part and perform quantitative analysis, a physician often needs to segment a prostate, and currently, segmentation of a prostate ultrasonic image is mainly completed by experienced physicians, but the method has the disadvantages of low efficiency and different segmentation effects among people, so that an automatic, rapid and accurate segmentation method is very necessary in clinical application.
Currently, ultrasound image segmentation methods for prostate are divided into two categories, one is that image edge information such as Sobel and Robert is extracted by traditional boundary operators proposed according to specific problems, and the other is that an end-to-end deep learning method which is popular in recent years, a network model represented by a U-net structure and a derivation mode thereof. However, the former detection effect is easily affected by the image artifact, and the latter often needs a large number of data sets to support iterative training, and is low in calculation efficiency and not very cost-effective in calculation time and calculation power.
Disclosure of Invention
In view of the above, there is a need to provide a prostate ultrasound image segmentation method and apparatus based on outlier detection, so as to solve the technical problems that the existing prostate ultrasound image segmentation method is greatly affected by artifact factors and/or has low computational efficiency.
In order to achieve the above object, the present invention provides a prostate ultrasound image segmentation method based on outlier detection, comprising the following steps:
s1, acquiring an average contour representing the shape of a prostate;
s2, obtaining a prostate ultrasonic image to be segmented, displaying the prostate ultrasonic image and the average contour on the same plane, and then deforming the average contour so as to preliminarily fit the deformed contour with a real prostate boundary on the prostate ultrasonic image;
s3, adjusting each point on the deformed outline to a gray value mutation position on the prostate ultrasonic image through a normal vector contrast boundary algorithm to obtain a rough segmentation outline;
s4, detecting and eliminating position abnormal points on the rough segmentation contour;
and S5, sequentially connecting all points on the roughly-segmented contour after the abnormal points in the position are removed to obtain the final segmented contour of the prostate ultrasonic image.
In some embodiments, the step S1 of obtaining the mean contour characterizing the shape of the prostate includes the following steps:
s11, selecting a certain number of artificially outlined typical prostate boundary contours;
s12, sampling each boundary contour at equal intervals according to a certain sequence to obtain a plurality of boundary contour point sets, and obtaining an average contour representing the shape of the prostate according to each boundary contour point set.
In some embodiments, in step S12, each boundary contour is sampled at equal intervals in a certain order to obtain a plurality of boundary contour point sets, and an average contour representing the shape of the prostate is obtained according to each boundary contour point set, which specifically includes:
s121, for each boundary contour, taking contour points right above the midpoint of the boundary contour as starting points, selecting a preset number of boundary contour points at equal intervals in the counterclockwise direction to obtain a plurality of boundary contour point sets, and setting the coordinate of the jth point on the ith boundary contour as (x) ij ，y ij )；
S122, press andthe coordinates of each point on the average contour are obtained by the following formula
Wherein, the serial numbers of i boundary contours, n is the number of boundary contours, and j is the serial number of contour points.
In some embodiments, in the step S2, the mean contour is deformed, and a method of the deformation includes at least one of scaling, translation, and rotation.
In some embodiments, in step S3, adjusting each point on the contour after the deformation processing to a position where the gray value on the ultrasound image of the prostate changes suddenly by using a normal vector contrast boundary algorithm to obtain a rough segmentation contour, specifically including the following steps:
s31, carrying out average method vector on each point on the deformed outline, extracting gray values of each pixel point on the prostate ultrasonic image where the normal vector corresponding to each point passes through, and sequentially combining the gray values into a gray value list;
s32, selecting a point, calculating a first gray value sum in a preset length range at one side of the selected point in the gray value list, and calculating a second gray value sum in a preset length range at the other side of the selected point to obtain a gray value difference value at two sides of the selected point;
s33, moving the selected point along the normal vector direction, calculating the gray value difference value of the two sides of the selected point after moving in place each time, and adjusting the selected point to the point position with the maximum gray value difference value;
and S34, processing other points on the deformed contour according to the step S32 and the step S33 in sequence to obtain a rough-divided contour.
In some embodiments, in the step S4, detecting and eliminating the position outlier on the rough-segmented contour specifically includes the following steps:
s41, obtaining coordinates of a central point of the roughly-segmented contour according to the roughly-segmented contour;
s42, sequentially calculating Euclidean distances between each point on the roughly-segmented contour and the central point according to a certain sequence to obtain a functional relation between the Euclidean distances and the serial numbers of the contour points;
s43, obtaining the gradient of the function of the Euclidean distance and the serial number of the contour point at each serial number, and when the gradient meets a preset relation, determining the contour point at the serial number as a position abnormal point;
and S44, eliminating the abnormal points at the positions on the rough segmentation contour.
In some embodiments, in step S41, the coordinates of the center point of the roughly segmented contour are:
center_x＝(x min +x max )/2
center_y＝(y min +y max )/2
where center _ x is the abscissa of the center point of the roughly-divided contour, center _ y is the ordinate of the center point of the roughly-divided contour, and x is the ordinate of the center point of the roughly-divided contour max The maximum abscissa value, x, corresponding to the roughly divided profile min Is the minimum abscissa value, y, corresponding to the roughly segmented contour max For the maximum ordinate value, y, corresponding to the roughly divided contour min Is the minimum ordinate value corresponding to the roughly divided contour.
In some embodiments, in step S42, the function relationship between the euclidean distance and the number of the contour points is:
wherein l i Point _ x [ i ] is the Euclidean distance between the contour point numbered i and the center point]Point _ y [ i ] as the abscissa of the contour point numbered i]The ordinate of the contour point numbered i.
In some embodiments, in step S43, the step of obtaining a gradient of the function relating to the euclidean distance and the number of the contour points at each number, and when the gradient satisfies a preset relationship, the contour point at the number is determined as the abnormal position point, which specifically includes:
s431, calculating the gradient of a function of the Euclidean distance and the number of the contour point at each number according to the following formula:
f i ＝(l i -l i-1 )/dis
wherein l i Is the Euclidean distance between the contour point with the number i and the central point, f i The gradient at the contour point with the number of i is shown, and dis is the distance between two adjacent contour points;
s432, when the following relation is satisfied, the contour point with the number i is a position abnormal point:
wherein f is i The gradient at the contour point numbered i, M is a preset value.
The present invention also provides a prostate ultrasound image segmentation apparatus, including: a processor, a memory, and a communication bus; the memory has stored thereon a computer readable program executable by the processor; the communication bus realizes connection communication between the processor and the memory; the processor, when executing the computer readable program, implements the method for prostate ultrasound image segmentation based on outlier detection.
Compared with the prior art, the technical scheme provided by the invention has the beneficial effects that: the method comprises the steps of preliminarily framing the prostate outline on a prostate ultrasonic image by representing the average outline of the prostate shape, correcting the preliminarily framed outline by a normal vector contrast boundary algorithm to enable the corrected outline to conform to gray scale information on the prostate ultrasonic image, detecting abnormal points of the corrected outline to eliminate noise points on the corrected outline, and connecting the rest normal points to obtain a final segmented outline of the prostate ultrasonic image.
Drawings
FIG. 1 is a collection of boundary contour points obtained by sampling an artificially delineated boundary contour of a typical prostate at equal intervals;
FIG. 2 is a mean profile of the characteristic prostate shape extracted from FIG. 1;
FIG. 3 is a schematic diagram of a process of deforming the mean contour to preliminarily fit the deformed contour to the real prostate boundary on the ultrasound image of the prostate;
FIG. 4 is a schematic diagram of a normal vector contrast boundary algorithm;
FIG. 5 is a schematic illustration of the mean vector for each point on the deformed profile of FIG. 3;
FIG. 6 is a rough segmentation contour of the ultrasound image of the prostate obtained by processing FIG. 5 using the normal vector contrast boundary algorithm;
FIG. 7 is a graph showing Euclidean distances between a point on the roughly-divided contour and the center point in FIG. 6
FIG. 8 is a set of individual point-to-center point distances on the coarsely segmented contours of FIG. 7;
FIG. 9 is a gradient at each number of the Euclidean distance function of the number of contour points in FIG. 8;
FIG. 10 is a schematic diagram of a normal dot link on the coarsely segmented contour in FIG. 7;
fig. 11 is a final segmented outline of the ultrasound image of the prostate.
Detailed Description
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate preferred embodiments of the invention and together with the description, serve to explain the principles of the invention and not to limit the scope of the invention.
Referring to fig. 1, the present invention provides a prostate ultrasound image segmentation method based on outlier detection, which includes the following steps:
s1, acquiring an average contour representing the shape of a prostate;
the step S1 specifically includes the following steps:
s11, selecting a certain number of artificially outlined typical prostate boundary contours (as shown in figure 1);
s12, sampling each boundary contour at equal intervals according to a certain sequence to obtain a plurality of boundary contour point sets (as shown in figure 1), and obtaining an average contour (as shown in figure 2) representing the shape of the prostate according to each boundary contour point set.
Step S12 specifically includes:
s121, for each boundary contour, taking contour points right above the midpoint of the boundary contour as starting points, selecting a preset number of boundary contour points at equal intervals in the counterclockwise direction to obtain a plurality of boundary contour point sets, and setting the coordinate of the jth point on the ith boundary contour as (x) ij ，y ij )；
S122, calculating the coordinates of each point on the average contour according to the following formula
Wherein, the serial numbers of i boundary contours, n is the number of boundary contours, and j is the serial number of contour points.
S2, obtaining a prostate ultrasonic image to be segmented, displaying the prostate ultrasonic image and the average contour on the same plane, and then deforming the average contour so as to enable the deformed contour to be preliminarily matched with a real prostate boundary on the prostate ultrasonic image;
the method of deformation processing includes at least one of scaling, translation and rotation.
Specifically, in this embodiment, as shown in fig. 3, after the average contour in the left image is scaled and translated, the deformed contour is preliminarily matched with the real prostate boundary on the ultrasound image of the prostate (as shown in the right image of fig. 3).
Step S2 may be performed manually or may be implemented by an automated algorithm. The specific principle of the automatic algorithm is as follows; a pyramid method from coarse to fine is adopted, in which the average shape of the training model is taken as a seed contour, reduced to adapt to different resolution levels, and the contour points of the model move along the normal direction, with a transition from light to dark; when all points are moved, a partially significant outline x can be detected s . By using x s And corresponding average shape portion
The similarity transformation T can be calculated s ：
Where m is the number of salient contour points and the transformed shape is used as the initialization result for the current level. The shape obtained at the optimal level is considered as the final initialization result.
S3, adjusting each point on the deformed outline to a gray value mutation position on the prostate ultrasonic image through a normal vector contrast boundary algorithm to obtain a rough segmentation outline;
for ease of understanding, the normal vector contrast boundary algorithm is first described below:
as shown in fig. 4, the method in steps S1 and S2 is used to initially locate the target object in the ultrasound image, forming a broad profile curve, which can be represented as a discretized set of points:
S＝{(x 1 ，y 1 )，(x 2 ，y 2 )，...，(x n ，y n )}
and (4) making a normal vector of a dividing line at each point on the contour curve, setting the normal vector into a point set with the same density as the pixel points, and not specifying that the direction of the normal vector points to the outside. Then each point on the normal vector has its fixed index value, we take any point as p (x, y), and assume that the index value of that point on the normal vector is ind. Then, the gray values of m points are respectively taken along the positive and negative directions of the p points, and a point set for representing the gray of each point is defined:
f＝[f ind+m ，f ind+m-1 ，...f ind ...，f ind-m+1 ，f ind-m ]
wherein f is i Indicating the gray value of the point with the row index value i. The prostate outline border is expanded into a strip and shown in a close-up view, as shown in figure 8.
Next, the gray-scale difference of the accumulated gray-scale values of the upper half area and the lower half area at a certain point along the direction is calculated by the following formula:
ind＝argmaxC ind
(3-2)
wherein i represents the value range of the upper and lower region boundary algorithm, ind represents the row index value of a certain point, and C is all C ind A set of constructs. (3-1) the gray difference of the accumulated gray values of the upper half area and the lower half area along a point in the direction is calculated by the formula, the point is moved on the normal vector and the set C is formed ind . Equation (3-2) is for determining the optimum division point, and the row index corresponding to the maximum gray scale difference is taken as C, and it can be considered that this is the index value of the target boundary point. In this way, all points on the discretized curve are traversed once, and the rough contour boundary can be obtained.
For better understanding of the present invention, the following description will be made in detail on how the normal vector contrast boundary algorithm is applied to the prostate ultrasound image segmentation.
The step S3 specifically includes the following steps:
s31, performing uniform vector processing on each point on the deformed contour (as shown in FIG. 5), extracting gray values of pixel points of normal vectors corresponding to each point on the prostate ultrasonic image, and sequentially combining the gray values into a gray value list;
s32, selecting a point, calculating a first gray value sum located in a preset length range on one side of the selected point in a gray value list, and calculating a second gray value sum located in a preset length range on the other side of the selected point to obtain a gray value difference value on two sides of the selected point;
s33, moving the selected point along the normal vector direction, calculating the gray value difference value of the two sides of the selected point after moving in place each time, and adjusting the selected point to the point position with the maximum gray value difference value;
s34, the other points on the deformed contour are processed in sequence in steps S32 and S33, and a rough-divided contour is obtained (as shown in fig. 6).
S4, detecting and eliminating position abnormal points on the roughly segmented contour;
the step S4 specifically includes the following steps:
s41, obtaining the coordinates of the central point of the roughly-divided outline according to the roughly-divided outline;
the coordinates of the center point of the roughly segmented contour are:
center_x＝(x min +x max )/2
center_y＝(y min +y max )/2
where center _ x is the abscissa of the center point of the roughly-divided contour, center _ y is the ordinate of the center point of the roughly-divided contour, and x is the ordinate of the center point of the roughly-divided contour max Is the maximum abscissa value, x, corresponding to the roughly divided contour min Is the minimum abscissa value, y, corresponding to the roughly segmented contour max Is the maximum ordinate value, y, corresponding to the roughly divided profile min Is the minimum ordinate value corresponding to the roughly divided contour.
S42, sequentially calculating Euclidean distances between each point on the roughly-segmented contour and the central point according to a certain sequence to obtain a functional relation between the Euclidean distances and the serial numbers of the contour points;
as shown in fig. 7, in step S42, the function relationship between the euclidean distance and the number of the contour points is:
wherein l i Point _ x [ i ] is the Euclidean distance between the contour point numbered i and the center point]Point _ y [ i ] as the abscissa of the contour point numbered i]The ordinate of the contour point numbered i.
Each point on the roughly-divided contour is traversed to obtain a point set of distances from each point to the central point, and the result is shown in fig. 8.
S43, solving gradients of functions of Euclidean distances and the serial numbers of the contour points at all serial numbers, and when the gradients meet a preset relation, determining the contour points at the serial numbers as abnormal position points;
step S43 specifically includes:
s431, calculating the gradient of a function of the Euclidean distance and the number of the contour point at each number according to the following formula:
f i ＝(l i -l i-1 )/dis
wherein l i Is the Euclidean distance between the contour point with the number i and the central point, f i The gradient of the contour point with the number of i is defined, dis is the distance between two adjacent contour points, and since the contour points are uniformly sampled, the distances between any two adjacent contour points are equal, so that dis =1 for convenient calculation;
the gradient of the function of euclidean distance and the number of contour points at each number is shown in fig. 9.
S432, when the following relation is satisfied, the contour point with the number i is a position abnormal point:
wherein f is i For the gradient at the contour point with the number i, M and M are preset values, and the values of M and M are set by experience, which can be set to 10 and 3 respectively in this embodiment.
In this embodiment, all contour points are traversed by the above formula, and the numbers of the position outliers i =51, 57, 76, 78 can be detected, and these points are stored in the list, such as list = [51, 57, 76, 78]. The list is then traversed if:
f [list[j]] *f [list[j+1]] ＜0
it indicates that the points corresponding to the list [ j ] to the points corresponding to the list [ j +1] are all outliers, and the list [ j ] and the list [ j +1] are taken out and then the remaining elements of the list are continuously traversed, and after the traversal is completed, all the outliers are obtained, where the number of the outliers in this embodiment is 51, 52, 53, 54, 55, 56, 57, 76, 77, and 78.
And S44, eliminating the abnormal points at the positions on the rough segmentation contour.
And S5, sequentially connecting all points on the roughly-segmented contour after the abnormal points in the position are removed to obtain the final segmented contour of the prostate ultrasonic image.
Extracting the extract satisfying the above formula f [list[j]] *f [list[j+1]] Value j < 0 due to list [ j]Corresponding point to list [ j +1]]The corresponding point is an outlier, so list [ j ]]-1 and list [ j +1]The point corresponding to +1 is a normal point, in this embodiment, 50, 58 and 75, 79 are normal points, and each group of normal points is connected into a line segment, such as a white line in fig. 10, so that a final segmented contour of the ultrasound image of the prostate can be obtained (see fig. 11).
In summary, the implementation process of the present invention is performed in three stages.
The first stage is as follows: roughly positioning the prostate in the original picture through the step S2, and roughly dividing the target prostate through the normal vector contour algorithm in the step S3;
and a second stage: and detecting the prostate contour with the noise points obtained in the first stage to obtain abnormal position points.
And a third stage: and deleting the abnormal points and connecting the other normal points to obtain the final prostate ultrasonic image segmentation contour.
The present invention also provides a prostate ultrasound image segmentation apparatus, including: a processor, a memory, and a communication bus;
the memory has stored thereon a computer readable program executable by the processor;
the communication bus realizes connection communication between the processor and the memory;
the processor, when executing the computer readable program, implements the method for prostate ultrasound image segmentation based on outlier detection.
In summary, the invention initially frames the prostate contour on the prostate ultrasound image by representing the average contour of the prostate shape, and then corrects the initially framed contour by the normal vector contrast boundary algorithm, so that the corrected contour conforms to the gray scale information on the prostate ultrasound image, and then performs the abnormal point detection on the corrected contour to eliminate the noise points thereon, and then connects the rest normal points to obtain the final segmented contour of the prostate ultrasound image.
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 changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.
Claims (10)
1. A prostate ultrasonic image segmentation method based on abnormal point detection is characterized by comprising the following steps:
s1, acquiring an average contour representing the shape of a prostate;
s2, obtaining a prostate ultrasonic image to be segmented, displaying the prostate ultrasonic image and the average contour on the same plane, and then deforming the average contour so as to enable the deformed contour to be preliminarily matched with a real prostate boundary on the prostate ultrasonic image;
s3, adjusting each point on the deformed outline to a gray value mutation position on the prostate ultrasonic image through a normal vector contrast boundary algorithm to obtain a rough segmentation outline;
s4, detecting and eliminating position abnormal points on the rough segmentation contour;
and S5, sequentially connecting all points on the roughly-segmented contour after the abnormal points in the position are removed to obtain the final segmented contour of the prostate ultrasonic image.
2. The method for segmenting a prostate ultrasound image based on outlier detection according to claim 1, wherein in step S1, the step of obtaining an average contour representing the shape of the prostate specifically comprises the following steps:
s11, selecting a certain number of artificially outlined typical prostate boundary contours;
s12, sampling all the boundary contours at equal intervals according to a certain sequence to obtain a plurality of boundary contour point sets, and obtaining an average contour representing the shape of the prostate according to all the boundary contour point sets.
3. The method for dividing a prostate ultrasound image based on outlier detection according to claim 2, wherein in step S12, each boundary contour is sampled at equal intervals in a certain order to obtain a plurality of boundary contour point sets, and an average contour representing a prostate shape is obtained according to each boundary contour point set, specifically comprising:
s121, selecting a preset number of boundary contour points at equal intervals in a counterclockwise direction by taking contour points right above the middle point of each boundary contour as starting points,obtaining a plurality of boundary contour point sets, and setting the coordinate of the jth point on the ith boundary contour as (x) ij ，y ij )；
S122, calculating the coordinates of each point on the average contour according to the following formula
Wherein, the serial numbers of i boundary contours, n is the number of the boundary contours, and j is the serial number of the contour point.
4. The method for segmenting a prostate ultrasound image based on outlier detection according to claim 1, wherein in the step S2, the mean contour is deformed by at least one of scaling, translation and rotation.
5. The method for segmenting a prostate ultrasonic image based on outlier detection according to claim 1, wherein in step S3, each point on the contour after the deformation processing is adjusted to a gray value mutation on the prostate ultrasonic image by using a normal vector contrast boundary algorithm, so as to obtain a rough segmented contour, specifically comprising the following steps:
s31, carrying out average method vector on each point on the deformed outline, extracting gray values of each pixel point on the prostate ultrasonic image where the normal vector corresponding to each point passes through, and sequentially combining the gray values into a gray value list;
s32, selecting a point, calculating a first gray value sum in a preset length range at one side of the selected point in the gray value list, and calculating a second gray value sum in a preset length range at the other side of the selected point to obtain a gray value difference value at two sides of the selected point;
s33, moving the selected point along the normal vector direction, calculating the gray value difference value of the two sides of the selected point after moving in place each time, and adjusting the selected point to the point position with the maximum gray value difference value;
and S34, processing other points on the deformed contour according to the step S32 and the step S33 in sequence to obtain a rough-divided contour.
6. The method for segmenting a prostate ultrasound image based on outlier detection according to claim 1, wherein the step S4 of detecting and removing the location outlier on the rough segmentation contour includes the following steps:
s41, obtaining the coordinates of the central point of the roughly-divided outline according to the roughly-divided outline;
s42, sequentially calculating Euclidean distances between each point on the roughly-segmented contour and the central point according to a certain sequence to obtain a functional relation between the Euclidean distances and the serial numbers of the contour points;
s43, obtaining the gradient of the function of the Euclidean distance and the serial number of the contour point at each serial number, and when the gradient meets a preset relation, determining the contour point at the serial number as a position abnormal point;
and S44, eliminating the abnormal points at the positions on the rough segmentation contour.
7. The method for segmenting a prostate ultrasound image based on outlier detection according to claim 6, wherein in said step S41, the coordinates of the central point of the roughly segmented contour are:
center_x＝(x min +x max )/2
center_y＝(y min +y max )/2
where center _ x is the abscissa of the center point of the roughly-divided contour, center _ y is the ordinate of the center point of the roughly-divided contour, and x is the ordinate of the center point of the roughly-divided contour max Is the maximum abscissa value corresponding to the roughly divided contour,
x min is the minimum abscissa value, y, corresponding to the roughly divided contour max For the maximum ordinate value, y, corresponding to the roughly divided contour min Is the minimum ordinate value corresponding to the roughly divided contour.
8. The method for segmenting prostate ultrasound images based on outlier detection as claimed in claim 7, wherein said step S42, the Euclidean distance is related to the number of contour points as a function of:
wherein l i Point-x [ i ] is the Euclidean distance between the contour point numbered i and the center point]Point _ y [ i ] as the abscissa of the contour point numbered i]The ordinate of the contour point numbered i.
9. The method of claim 8, wherein the step S43 of obtaining the gradient of the function of Euclidean distance and the number of contour points at each number, and if the gradient satisfies a predetermined relationship, the contour point at the number is determined as the abnormal position point, and the method specifically comprises:
s431, calculating the gradient of a function of the Euclidean distance and the number of the contour points at each number according to the following formula:
f i ＝(l i -l i-1 )/dis
wherein l i Euclidean distance, f, of the contour point with number i from the center point i The gradient at the contour point with the number of i is shown, and dis is the distance between two adjacent contour points;
s432, when the following relation is satisfied, the contour point with the number i is a position abnormal point:
wherein, f i The gradients at the contour points numbered i, M, are preset values.
10. An apparatus for segmenting an ultrasound image of a prostate, comprising: a processor, a memory, and a communication bus;
the memory has stored thereon a computer readable program executable by the processor;
the communication bus realizes the connection communication between the processor and the memory;
the processor, when executing the computer readable program, implements the method of prostate ultrasound image segmentation based on outlier detection of any of claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211350084.2A CN115690047A (en) | 2022-10-31 | 2022-10-31 | Prostate ultrasound image segmentation method and device based on abnormal point detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211350084.2A CN115690047A (en) | 2022-10-31 | 2022-10-31 | Prostate ultrasound image segmentation method and device based on abnormal point detection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115690047A true CN115690047A (en) | 2023-02-03 |
Family
ID=85046798
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211350084.2A Pending CN115690047A (en) | 2022-10-31 | 2022-10-31 | Prostate ultrasound image segmentation method and device based on abnormal point detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115690047A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117911792A (en) * | 2024-03-15 | 2024-04-19 | 垣矽技术(青岛)有限公司 | Pin detecting system for voltage reference source chip production |
-
2022
- 2022-10-31 CN CN202211350084.2A patent/CN115690047A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117911792A (en) * | 2024-03-15 | 2024-04-19 | 垣矽技术(青岛)有限公司 | Pin detecting system for voltage reference source chip production |
| CN117911792B (en) * | 2024-03-15 | 2024-06-04 | 垣矽技术(青岛)有限公司 | Pin detecting system for voltage reference source chip production |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111798462B (en) | Automatic delineation method of nasopharyngeal carcinoma radiotherapy target area based on CT image | |
| CN110930416B (en) | MRI image prostate segmentation method based on U-shaped network | |
| CN112950651B (en) | Automatic delineation method of mediastinal lymph drainage area based on deep learning network | |
| CN112257704A (en) | Cervical fluid-based cell digital image classification method based on deep learning detection model | |
| CN110599528A (en) | Unsupervised three-dimensional medical image registration method and system based on neural network | |
| EP4420598A1 (en) | Method and system for measuring lesion features of hypertensive retinopathy | |
| CN108876769B (en) | Left auricle CT image segmentation method | |
| WO2021136368A1 (en) | Method and apparatus for automatically detecting pectoralis major region in molybdenum target image | |
| CN109767459A (en) | Novel ocular base map method for registering | |
| CN102831614B (en) | Sequential medical image quick segmentation method based on interactive dictionary migration | |
| CN113066093A (en) | Automatic segmentation and parting and thickness measuring method based on endometrial ultrasound image | |
| CN113643354B (en) | Measuring device of vascular caliber based on fundus image with enhanced resolution | |
| CN115690047A (en) | Prostate ultrasound image segmentation method and device based on abnormal point detection | |
| US20240005494A1 (en) | Methods and systems for image quality assessment | |
| CN113782184A (en) | Cerebral apoplexy auxiliary evaluation system based on facial key point and feature pre-learning | |
| CN101667297A (en) | Method for extracting breast region in breast molybdenum target X-ray image | |
| CN103345741A (en) | Non-rigid multimode medical image precise registering method | |
| CN111080556A (en) | Method, system, equipment and medium for strengthening trachea wall of CT image | |
| CN116725563B (en) | Eyeball salience measuring device | |
| CN109816665B (en) | Rapid segmentation method and device for optical coherence tomography image | |
| CN116309388A (en) | Bone age intelligent assessment method based on multi-region combination | |
| CN114627136B (en) | Tongue image segmentation and alignment method based on feature pyramid network | |
| CN113706563B (en) | X-ray chest radiography lung field segmentation method for automatically initializing Snake model | |
| CN114359308A (en) | Aortic dissection method based on edge response and nonlinear loss | |
| CN113269783A (en) | Pulmonary nodule segmentation method and device based on three-dimensional attention mechanism |
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 |