CN116385436A - Cholelithiasis auxiliary detection system based on CT image - Google Patents
Cholelithiasis auxiliary detection system based on CT image Download PDFInfo
- Publication number
- CN116385436A CN116385436A CN202310645058.0A CN202310645058A CN116385436A CN 116385436 A CN116385436 A CN 116385436A CN 202310645058 A CN202310645058 A CN 202310645058A CN 116385436 A CN116385436 A CN 116385436A
- Authority
- CN
- China
- Prior art keywords
- gray
- representative
- pixel
- super
- target
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 19
- 201000001883 cholelithiasis Diseases 0.000 title claims abstract description 16
- 208000001130 gallstones Diseases 0.000 claims abstract description 49
- 230000011218 segmentation Effects 0.000 claims abstract description 24
- 210000000232 gallbladder Anatomy 0.000 claims abstract description 22
- 238000000034 method Methods 0.000 claims abstract description 10
- 238000010899 nucleation Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 6
- 230000008901 benefit Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 239000003086 colorant Substances 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 208000020694 gallbladder disease Diseases 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 208000004845 Cholecystolithiasis Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000739 chaotic effect Effects 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 238000013170 computed tomography imaging Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20021—Dividing image into blocks, subimages or windows
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20076—Probabilistic image processing
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Medical Informatics (AREA)
- Quality & Reliability (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Image Processing (AREA)
Abstract
The invention relates to the field of image processing, in particular to a cholelithiasis auxiliary detection system based on CT images, which comprises the following components: the gall bladder CT image acquisition module, the super-pixel segmentation module, the similarity calculation module, the super-pixel block merging module and the gall bladder stone identification module acquire gall bladder CT images, further obtain super-pixel segmentation images, obtain representative pixel points of each super-pixel block, obtain a representative gray sequence and a representative gray curve according to the representative pixel points, further obtain local fluctuation degree of the representative gray curve, obtain a smooth gray sequence according to the representative gray sequence, obtain integral fluctuation degree of the representative gray curve according to the weight of each gray value in the smooth gray sequence, obtain attention weight of the local fluctuation degree, obtain the similarity of two adjacent super-pixel blocks according to the local fluctuation degree and the integral fluctuation degree, merge the super-pixel blocks, and identify a gall bladder stone area according to the merged images. The gall-stone region identified by the method is more accurate.
Description
Technical Field
The invention relates to the field of image processing, in particular to a cholelithiasis auxiliary detection system based on CT images.
Background
In the field of medical imaging, CT imaging technology has been widely used in clinical diagnosis and treatment with its unique advantage, so that the clinician can observe the lesions of each part in the patient more clearly, and the judgment rate and accuracy are higher.
Since the incidence of gallbladder diseases increases year by year, cholecystolithiasis is the most common gallbladder disease, and ultrasonic images are processed by an image processing technology so that whether the gallstone exists or not is easier to judge.
In the prior art, the super-pixel segmentation is used for segmenting an image, when the super-pixel block is too small, the super-segmentation is caused to the gall-stone region, the gall-stone region is divided into a plurality of super-pixels, and when the super-pixel block is too large, the under-segmentation is caused to the gall-stone region, so that the recognition of the gall-stone is not facilitated.
Disclosure of Invention
In order to solve the above problems, the present invention provides a cholelithiasis auxiliary detection system based on CT images, the system comprising:
the gallbladder CT image acquisition module acquires a gallbladder CT image;
the super-pixel segmentation module is used for performing super-pixel segmentation on the gallbladder CT image to obtain a super-pixel segmented image;
the similarity calculation module is used for obtaining the mass center of each super pixel block in the super pixel segmentation image, and taking the mass center as a target pixel point of each super pixel block; obtaining the probability of the target pixel point as the representative pixel point, and obtaining the representative pixel point of each super pixel block according to the probability of the target pixel point as the representative pixel point;
taking any two adjacent super-pixel blocks as a target super-pixel block pair, and acquiring a representative gray sequence and a representative gray curve according to representative pixel points of the two super-pixel blocks in the target super-pixel block pair; obtaining the local fluctuation degree of the representative gray scale curve according to the representative gray scale sequence; acquiring a smooth gray sequence according to the representative gray sequence; acquiring the weight of each gray value in the smooth gray sequence, and acquiring the overall fluctuation degree of the representative gray curve according to the weight; obtaining the attention weight of the local fluctuation degree according to the representative gray sequence, and obtaining the similarity between two super-pixel blocks in the target super-pixel block pair according to the local fluctuation degree, the whole fluctuation degree and the attention weight of the local fluctuation degree;
obtaining the similarity between all adjacent two super pixel blocks in the super pixel segmentation image;
the super pixel block merging module merges the super pixel blocks according to the similarity to obtain a merged image;
and the gall-stone identification module is used for identifying a gall-stone area according to the combined image.
Preferably, the obtaining the probability that the target pixel point is a representative pixel point, and obtaining the representative pixel point of each super pixel block according to the probability that the target pixel point is the representative pixel point, includes the steps of:
obtaining the probability that the target pixel point is the representative pixel point:
wherein ,
representing the probability that the target pixel point is the representative pixel point;
a gray value representing a target pixel;
representing the eighth neighborhood of the target pixel point
Gray values of the individual pixels;
is an exponential function with a natural constant as a base;
is an absolute value symbol;
setting an empty candidate representative pixel point set; when the probability that the target pixel point is the representative pixel point is larger than the representative threshold value, adding the target pixel point into a candidate representative pixel point set, and taking the target pixel point with the maximum probability of being the representative pixel point in the candidate representative pixel point set as the representative pixel point of the super pixel point to which the target pixel point belongs; when the probability of the target pixel point being the representative pixel point is smaller than or equal to a representative threshold value, each pixel point in the eight adjacent areas of the target pixel point is respectively used as a new target pixel point, the probability of each target pixel point being the representative pixel point is obtained, all the target pixel points with the probability of being the representative pixel point being larger than the representative threshold value are added into a candidate representative pixel point set, and when the candidate representative pixel point set is not empty, the target pixel point with the maximum probability of being the representative pixel point in the candidate representative pixel point set is used as the representative pixel point of the super pixel point to which the target pixel point belongs; and when the probability that all the target pixel points are the representative pixel points is smaller than or equal to the representative threshold value, repeatedly acquiring new target pixel points, and the like, and stopping iteration until the representative pixel points of the super pixel blocks are acquired.
Preferably, the step of obtaining the representative gray scale sequence and the representative gray scale curve according to the representative pixel points of the two super pixel blocks in the target super pixel block pair includes the following steps:
connecting the representative pixel points of two super pixel blocks in the target super pixel block pair, obtaining the gray values of all pixel points on the connecting line between the representative pixel points of the two super pixel blocks in the target super pixel block pair, and forming a representative gray sequence of the target super pixel block pair; and drawing a representative gray scale curve of the target super-pixel block pair by taking an index representing each gray scale value in the gray scale sequence as an abscissa and taking each gray scale value in the gray scale sequence as an ordinate.
Preferably, the step of obtaining the local fluctuation degree of the representative gray scale curve according to the representative gray scale sequence includes the steps of:
counting the frequency of each gray value appearing in the representative gray sequence of the target super-pixel block pair, and acquiring the local fluctuation degree of the representative gray curve of the target super-pixel block pair according to the frequency of each gray value in the representative gray sequence:
wherein 
Representing gray scale curves for pairs of target superpixel blocksIs a local degree of fluctuation;
the first of the representative gray scale sequences for the target superpixel block pair
The frequency of the seed gray value;
the number of classes of gray values in the representative gray sequence for the target superpixel block pair.
Preferably, the step of obtaining the smoothed gray scale sequence from the representative gray scale sequence includes the steps of:
and calculating the average value of all gray values in each window in the sliding process of the sliding window, rounding the average value to be used as the average gray value of the window, and forming a one-dimensional sequence of the average gray values of all windows in the sliding process of the sliding window according to the sequence to be used as the smooth gray sequence of the target super-pixel block pair.
Preferably, the step of obtaining the weight of each gray value in the smoothed gray sequence includes the steps of:
acquiring absolute values of differences between each gray value and two adjacent gray values in the smooth gray sequence, and respectively serving as a first difference value and a second difference value of each gray value; taking the average value of the first difference value and the second difference value of each gray value as the sequential difference value of each gray value;
combining the sequential difference value of each gray value in the smooth gray sequence of the target super-pixel block pair to obtain the weight of each gray value in the smooth gray sequence:
wherein ,
the smooth gray sequence for the target super-pixel block pair
A weight for seeding gray values;
the smooth gray sequence for the target super-pixel block pair
The seed gray value is in the smooth gray sequence
Sequential differences in the next occurrence;
the smooth gray sequence for the target super-pixel block pair
The number of times the seed gray value occurs in the smoothed gray sequence;
the number of classes of gray values in the smoothed gray sequence for the target superpixel block pair.
Preferably, the step of obtaining the overall fluctuation degree of the representative gray scale curve according to the weight includes the steps of:
counting the frequency of each gray value in the smooth gray sequence, and acquiring the overall fluctuation degree of the representative gray curve of the target super-pixel block pair according to the frequency and the weight of each gray value in the smooth gray sequence:
wherein 
The overall fluctuation degree of the representative gray level curve of the target super pixel block pair;
the smooth gray sequence for the target super-pixel block pair
The frequency of the seed gray value;
the smooth gray sequence for the target super-pixel block pair
A weight for seeding gray values;
the number of classes of gray values in the smoothed gray sequence for the target superpixel block pair.
Preferably, the method for obtaining the attention weight of the local fluctuation degree according to the representative gray sequence comprises the following steps:
the gray value belonging to the first super pixel block in the target super pixel block pair in the representative gray sequence of the target super pixel block pair is called a first gray value, the gray value belonging to the second super pixel block is called a second gray value, and the nearest to the second gray value is obtained
The average value of the first gray values is used as the first boundary gray to obtain the nearest gray value
The average value of the second gray values is used as the second boundary gray, the absolute value of the difference value between the first boundary gray and the second boundary gray is obtained, the range of the representative gray sequence is obtained, and the ratio of the absolute value of the difference value between the first boundary gray and the second boundary gray to the range is used as the attention weight of the local fluctuation degree of the representative gray curve of the target super-pixel block pair, wherein
Is the preset boundary number.
Preferably, the step of obtaining the similarity between two superpixel blocks in the target superpixel block pair according to the local fluctuation degree, the overall fluctuation degree and the attention weight of the local fluctuation degree includes the steps of:
wherein ,
for the similarity between two superpixel blocks in the target superpixel block pair:
a focus weight representing the local fluctuation degree of the gray scale curve for the target super pixel block pair;
the overall fluctuation degree of the representative gray level curve of the target super pixel block pair;
is the local fluctuation degree of the representative gray scale curve of the target super pixel block pair.
Preferably, the step of identifying the gall stone region according to the combined image includes the steps of:
and carrying out ellipse fitting on the edge of each super pixel block, and obtaining a gall-stone region according to the fitting error and the area of each super pixel block.
The invention has the following beneficial effects: according to the method, the representative pixel point of each super pixel block is obtained according to the centroid of each super pixel block, the representative gray sequence and the representative gray curve are obtained according to the connecting line between the representative pixel points of the adjacent super pixel blocks, the adjacent super pixel blocks are combined through the fluctuation condition of the representative gray curve. According to the invention, the super pixel blocks are combined, so that each super pixel block can represent an independent image characteristic in the gall-bladder CT image, and each gall-stone region identified according to the super pixel blocks in the combined image is more complete and accurate.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a system block diagram of a cholelithiasis auxiliary detection system based on a CT image according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the present invention to achieve the preset purpose, the following detailed description refers to specific embodiments, structures, features and effects of a cholelithiasis auxiliary detection system based on CT images according to the present invention with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following specifically describes a specific scheme of the cholelithiasis auxiliary detection system based on CT images.
Referring to fig. 1, a cholelithiasis auxiliary detection system based on CT images according to an embodiment of the present invention is shown, and the system includes the following modules:
the gallbladder CT image acquisition module S101 acquires a gallbladder CT image.
The CT machine is used for scanning the gallbladder part of the patient, and CT images of the gallbladder part are acquired and recorded as gallbladder CT images.
So far, a gallbladder CT image is acquired.
The super-pixel segmentation module S102 performs super-pixel segmentation on the gallbladder CT image.
The super-pixel segmentation is to form a super-pixel block with a certain visual meaning by using adjacent pixel points with the same texture, color, brightness and other characteristics, and the super-pixel segmentation can be used for segmenting the gall-stone region to a certain extent because the gray values of the pixels of the gall-stone region are uniform.
In the embodiment of the invention, the gall bladder CT image is subjected to super-pixel segmentation to obtain a super-pixel segmented image. It should be noted that, in the embodiment of the present invention, when performing superpixel segmentation on a gallbladder CT image, the size of the superpixel block is set to be
In other embodiments, the practitioner may set the size of the superpixel block according to the actual implementation, but note that, in order to ensure that the gall-stone region and other regions are not divided into the same superpixel block, to affect the recognition of the subsequent gall-stone, the size of the superpixel block needs to be smaller than that of the following superpixel block
。
So far, the super-pixel segmentation of the gallbladder CT image is realized, and the super-pixel segmented image is obtained.
The similarity calculation module S103 obtains the similarity between the adjacent super pixel blocks.
In order to avoid that the gall-stone region and other regions are segmented into the same super-pixel block and influence the recognition of gall-stone, the super-pixel segmentation module is provided with a smaller super-pixel block size to segment the gall-stone CT image. Because the size of the super-pixel blocks is smaller, the gall-stone region is possibly divided into a plurality of super-pixel blocks, so that the similarity among different super-pixel blocks needs to be analyzed, and the super-pixel blocks are combined to obtain the complete gall-stone region. The super pixel segmentation is to form a super pixel block with a certain visual meaning by adjacent pixel points with the same texture, color, brightness and other characteristics, so that the colors and textures of all the pixel points in each super pixel block are similar, and the colors and textures of different super pixel blocks can be distinguished. In order to obtain the similarity between different super-pixel blocks, first, a representative pixel point capable of representing the characteristics of each super-pixel block needs to be obtained.
In the embodiment of the invention, an empty candidate representative pixel point set is set.
And acquiring the mass center of each super pixel block in the super pixel segmentation image, and taking the mass center as a target pixel point of each super pixel block. Obtaining the probability that the target pixel point is the representative pixel point:
wherein ,
representing the probability that the target pixel point is the representative pixel point;
a gray value representing a target pixel;
representing the eighth neighborhood of the target pixel point
Gray values of the individual pixels;
is an exponential function with a natural constant as a base;
is an absolute value symbol;
representing the difference between the target pixel point and the gray average value in the eight neighborhood of the target pixel point, when the difference is larger, the target pixel point is more likely to be a noise point, the target pixel point cannot represent the characteristics of the super pixel block, and the probability that the target pixel point is a representative pixel point is smaller; on the contrary, when the difference is smaller, the gray value of the target pixel point is more similar to the gray value of the neighborhood pixel point, the target pixel point is less likely to be a noise point, and at the moment, the target pixel point can represent the characteristic of the super pixel block, and the probability that the target pixel point is a representative pixel point is larger.
Presetting a representative threshold value
Wherein the embodiment of the invention is as follows
For example, =0.98, the embodiment of the invention is not particularly limited, wherein
Depending on the particular implementation. When the probability of the target pixel point being the representative pixel point is larger than the representative threshold value
When the target pixel point is added into the candidate representative pixel point set, the target pixel point with the highest probability of being the representative pixel point in the candidate representative pixel point set is taken as the representative pixel point of the super pixel point to which the target pixel point belongs; when the probability of the target pixel point being the representative pixel point is smaller than or equal to the representative threshold value
When the pixel points in the eight adjacent areas of the target pixel point are respectively used as new target pixel points, the probability that each target pixel point is a representative pixel point is obtained, and the probability that all the representative pixel points are larger than a representative threshold value
Adding the target pixel points of (1) into the candidate representative pixel point set, whenAnd when the candidate representative pixel point set is not empty, taking the target pixel point with the highest probability of being the representative pixel point in the candidate representative pixel point set as the representative pixel point of the super pixel point to which the target pixel point belongs. When the probability of all the target pixel points being the representative pixel points is smaller than or equal to the representative threshold value
And repeatedly acquiring a new target pixel point, and so on until the representative pixel point of the super pixel block is acquired.
Thus, a representative pixel point of each super pixel block is obtained.
It should be noted that, the gray values of the gall-stone region are relatively uniform, and the gray values of the gall-stone region and the surrounding regions are different, so that the similarity of the adjacent super-pixel blocks can be obtained according to the gray value fluctuation condition of the pixel points between the representative pixel points of the adjacent super-pixel blocks, and the super-pixel blocks can be combined according to the similarity to obtain the gall-stone region.
In the embodiment of the invention, two adjacent super-pixel blocks are used as a super-pixel block pair, any one super-pixel block pair is used as a target super-pixel block pair, representative pixel points of two super-pixel blocks in the target super-pixel block pair are connected, gray values of all pixel points on a connecting line between the representative pixel points of the two super-pixel blocks in the target super-pixel block pair are obtained, and a representative gray sequence of the target super-pixel block pair is formed. And drawing a representative gray scale curve of the target super-pixel block pair by taking an index representing each gray scale value in the gray scale sequence as an abscissa and taking each gray scale value in the gray scale sequence as an ordinate.
It should be noted that, because the gray values of the gall-stone regions are uniform, if two super-pixel blocks in the target super-pixel block pair belong to the same gall-stone region, the change trend of the representative gray curve of the target super-pixel block pair is stable, so that the local fluctuation degree of the representative gray curve of the target super-pixel block pair can be obtained to reflect the change trend of the representative gray curve of the target super-pixel block pair.
In the embodiment of the invention, the frequency of each gray value appearing in the representative gray sequence of the target super-pixel block pair is counted, and the local fluctuation degree of the representative gray curve of the target super-pixel block pair is obtained according to the frequency of each gray value in the representative gray sequence:
wherein 
Local fluctuation degree of the representative gray scale curve of the target super pixel block pair;
the first of the representative gray scale sequences for the target superpixel block pair
The frequency of the seed gray value;
the number of classes of gray values in the representative gray sequence for the target super pixel block pair;
the entropy of all gray values in the representative gray sequence of the target super pixel block pair is larger when a plurality of gray values exist in the representative gray sequence and the frequency distribution of each gray value is more consistent, the representative gray sequence is more chaotic, and the local fluctuation degree of the representative gray curve is larger; when the frequency of one gray value in the representative gray sequence is larger and the frequency of the rest gray values is smaller, the representative gray sequence mainly takes the gray value with larger frequency as the main part, and at the moment, the entropy is smaller and the local fluctuation degree of the representative gray curve is smaller.
It should be noted that, the local fluctuation degree of the representative gray scale curve may reflect the variation trend of the representative gray scale curve, but the representative gray scale curve has a locally smaller fluctuation condition due to the influence of the image noise, and the local fluctuation degree is influenced by the locally smaller fluctuation condition of the representative gray scale curve, resulting in a possibly larger local fluctuation degree. In order to avoid the influence of image noise, the representative gray scale curve can be smoothed to obtain a smoothed gray scale sequence, and the whole fluctuation degree of the representative gray scale curve is obtained by combining the smoothed gray scale sequence.
In the embodiment of the invention, the representative gray sequence of the target super pixel block pair is constructed to be as large as
A sliding window of step size 1, in an embodiment of the present invention,
in other embodiments, the practitioner can set according to the actual implementation
Is a value of (2). In the sliding process of the sliding window, calculating the average value of all gray values in each window, rounding and rounding the average value to be used as the average gray value of the window, and forming a one-dimensional sequence of the average gray values of all windows in the sliding process of the sliding window according to the sequence to be used as a smooth gray sequence of the target super-pixel block pair.
It should be noted that, the smoothed gray scale sequence after smoothing the representative gray scale curve of the target super pixel block pair eliminates the local fluctuation caused by the noise point, and can reflect the overall fluctuation condition of the representative gray scale curve to a certain extent. The entropy may reflect the degree of confusion of the data, and the entropy of the smoothed gray sequence may reflect the degree of confusion of gray values occurring in the sequence, but may not reflect the sequential relationship of the gray values occurring. The sequence relation of gray values can reflect the overall fluctuation condition of the representative gray curve to a certain extent. Therefore, the embodiment of the invention obtains the weight of each gray value according to the local position relation of the gray value in the smooth gray sequence, and reflects the local sequence relation of the gray value by using the weight, so that the entropy of the smooth gray sequence is corrected according to the weight of each gray value, and the integral fluctuation degree of the representative gray curve is obtained.
In the embodiment of the invention, the absolute value of the difference value between each gray value and two adjacent gray values in the smooth gray sequence is obtained and is respectively used as a first difference value and a second difference value of each gray value. The average value of the first difference value and the second difference value of each gray value is taken as the sequential difference value of each gray value. It should be noted that, the first gray value in the smoothed gray sequence has no first difference, the second difference of the gray values is used as the sequential difference of the gray values, the last gray value in the smoothed gray sequence has no second difference, and the first difference of the gray values is used as the sequential difference of the gray values.
Combining the sequential difference value of each gray value in the smooth gray sequence of the target super-pixel block pair to obtain the weight of each gray value in the smooth gray sequence:
wherein ,
the smooth gray sequence for the target super-pixel block pair
A weight for seeding gray values;
the smooth gray sequence for the target super-pixel block pair
The seed gray value is in the smooth gray sequence
Sequential differences in the next occurrence;
the smooth gray sequence for the target super-pixel block pair
The number of times the seed gray value occurs in the smoothed gray sequence;
the number of classes of gray values in the smooth gray sequence for the target super-pixel block pair; when (when)First, the
The description of the first gray level is given when the sequential difference of each occurrence of the gray level values in the smoothed gray level sequence is large
The gray value of each occurrence in the smooth gray sequence is greatly different from the adjacent gray value, namely the first
The seed gray value can cause larger data fluctuation when each occurrence in the smooth gray sequence, and is the first time
The gray values are provided with larger weights, so that the gray values which can cause larger data fluctuation are more focused when the overall fluctuation degree of the representative gray curve is acquired later, and the obtained overall fluctuation degree is more accurate.
Counting the frequency of each gray value in the smooth gray sequence, and acquiring the overall fluctuation degree of the representative gray curve of the target super-pixel block pair according to the frequency and the weight of each gray value in the smooth gray sequence:
wherein 
The overall fluctuation degree of the representative gray level curve of the target super pixel block pair;
the smooth gray sequence for the target super-pixel block pair
The frequency of the seed gray value;
the smooth gray sequence for the target super-pixel block pair
A weight for seeding gray values;
the number of classes of gray values in the smooth gray sequence for the target super-pixel block pair; when the first is
The greater the weight of the seed gray value, the more attention is paid to the first
The frequency of the seed gray value, as
The smaller the weight of the seed gray value is, the less attention is paid to the first
The frequency of the seed gray value; when a plurality of gray values exist in the smooth gray sequence, the frequency distribution of each gray value is more consistent, and the weight of each gray value is larger, the overall fluctuation degree of the representative gray curve is larger; when the frequency of one gray value in the smooth gray sequence is larger, the weight of the gray value is smaller, and the frequencies of the rest gray values are smaller, the smooth gray sequence mainly takes the gray value with larger frequency as the main part, and the gray values with larger frequency are intensively distributed, so that the integral fluctuation degree of the representative gray curve is smaller.
It should be noted that, the overall fluctuation degree of the representative gray level curve is obtained according to a smooth gray level sequence, and the smooth gray level sequence eliminates the local fluctuation condition caused by noise in the representative gray level curve, but if two super pixel blocks in the target super pixel pair belong to different gall stones, the connection line of the representative pixel points of the two super pixel blocks may also include other background pixel points between the two gall stones, that is, a part of background gray level still exists on the representative gray level curve of the target super pixel pair. The smoothed gray sequence also eliminates local fluctuations caused by the background gray in the representative gray curve. Therefore, the embodiment of the invention combines the local fluctuation degree and the overall fluctuation degree of the representative gray level curve to reflect the similarity between the two super-pixel blocks of the target super-pixel pair. In order to avoid the excessive influence of noise in the local fluctuation degree, a concern weight is required to be set for the local fluctuation degree by combining the gray level distribution characteristics representing the junction of two super-pixel blocks in the gray level curve, when the gray level difference of the junction is large, the local fluctuation degree is more concerned, and when the gray level difference of the junction is small, the whole fluctuation degree is more concerned.
In the embodiment of the invention, a numerical value is preset
And is denoted as the number of junctions, in embodiments of the present invention,
in other embodiments, the practitioner can set according to the actual implementation
Is a value of (2).
The gray value belonging to the first super pixel block in the target super pixel block pair in the representative gray sequence of the target super pixel block pair is called a first gray value, the gray value belonging to the second super pixel block is called a second gray value, and the nearest to the second gray value is obtained
The average value of the first gray values is used as the first boundary gray to obtain the nearest gray value
And taking the average value of the second gray values as a second boundary gray, acquiring the absolute value of the difference value between the first boundary gray and the second boundary gray, acquiring the range of the representative gray sequence, and taking the ratio of the absolute value of the difference value between the first boundary gray and the second boundary gray to the range as the attention weight of the local fluctuation degree of the representative gray curve of the target super pixel block pair.
Obtaining the similarity between two super pixel blocks in the target super pixel block pair:
wherein ,
for the similarity between two superpixel blocks in the target superpixel block pair:
a focus weight representing the local fluctuation degree of the gray scale curve for the target super pixel block pair;
the overall fluctuation degree of the representative gray level curve of the target super pixel block pair;
local fluctuation degree of the representative gray scale curve of the target super pixel block pair; when the attention weight of the local fluctuation degree is larger, the gray scale difference of the adjacent parts of the two super pixel blocks of the target super pixel block pair is larger, and the local fluctuation degree representing the gray scale curve is more concerned at the moment; when the attention weight of the local fluctuation degree is smaller, the gray value difference of the adjacent parts of the two super pixel blocks of the target super pixel block pair is smaller, and the attention is paid to the whole fluctuation degree representing the gray curve; when the attention weight of the local fluctuation degree is smaller and the whole fluctuation degree is smaller, the whole representative gray scale curve is more stable, and the two super-pixel blocks in the target super-pixel block pair are more similar; when the degree of overall fluctuation is larger, or the degree of attention of the degree of local fluctuation is larger and the degree of local fluctuation is larger, the representative gradation curve fluctuation is larger, and the two super-pixel blocks in the target super-pixel block pair are more dissimilar.
Thus, the similarity between the two super pixel blocks in the target super pixel block pair is obtained. And similarly, obtaining the similarity between all two adjacent super pixel blocks in the super pixel segmentation image.
And the super pixel block merging module S104 merges the super pixel blocks to obtain a merged image.
It should be noted that, the more similar two super-pixel blocks may contain the same image features, and the less similar two super-pixel blocks may contain different image features. To obtain a complete gall-stone region, the super-pixel blocks need to be combined according to the similarity of adjacent super-pixel blocks.
In the embodiment of the invention, a similarity threshold value is preset
Wherein the embodiment of the invention is as follows
For example, =0.74, the embodiment of the invention is not particularly limited, wherein
Depending on the particular implementation. When the similarity between two adjacent super-pixel blocks is greater than the similarity threshold
When the two super pixel blocks are combined into one super pixel block, otherwise, when the similarity between two adjacent super pixel blocks is smaller than or equal to the similarity threshold value
When the two super pixel blocks are not merged. The image after the merging of the super pixel blocks is noted as a merged image.
Thus, a combined image is acquired.
The gall-stone recognition module S105 recognizes a gall-stone region according to the combined image.
The gall stone is similar to an ellipse and has a smaller area.
In the embodiment of the invention, the edge of each super pixel block is subjected to ellipse fitting by using a least square method to obtain a fitting result and a fitting error, when the fitting error is larger, the shape of the corresponding super pixel block is larger than the ellipse, and when the fitting error is smaller, the shape of the corresponding super pixel block is basically ellipse, and in the embodiment of the invention, a fitting error threshold value is preset
An area threshold
Wherein the embodiment of the invention is as follows
=10、
The embodiment of the present invention is not particularly limited, and is described by taking 1000 as an example, wherein
、
Depending on the particular implementation. Fitting error is less than
And an area smaller than
The super pixel block is the gall-stone area.
So far, the gall stone region is obtained.
In summary, the system comprises a gallbladder CT image acquisition module, a superpixel segmentation module, a similarity calculation module, a superpixel block merging module and a gall stone identification module, the invention performs superpixel segmentation on a gallbladder CT image, acquires the representative pixel point of each superpixel block according to the mass center of each superpixel block, acquires a representative gray scale sequence and a representative gray scale curve according to the connecting line between the representative pixel points of adjacent superpixel blocks, merges the adjacent superpixel blocks through the fluctuation condition of the representative gray scale curve, acquires the local fluctuation degree of the representative gray scale curve, acquires a smooth gray scale sequence according to the representative gray scale sequence, eliminates the local fluctuation caused by noise, acquires the integral fluctuation degree of the representative gray scale curve according to the weight of each gray scale value in the smooth gray scale sequence, acquires the attention weight of the local fluctuation degree according to the distribution condition of the gray scale value at the junction on the connecting line between the representative pixel points of the adjacent superpixel blocks, acquires the similarity of the adjacent superpixel blocks according to the local fluctuation degree, avoids the background gray scale between two adjacent superpixel blocks when the influence of noise is eliminated, and carries out gall stone merging according to the similarity of the super-pixel blocks. According to the invention, the super pixel blocks are combined, so that each super pixel block can represent an independent image characteristic in the gall-bladder CT image, and each gall-stone region identified according to the super pixel blocks in the combined image is more complete and accurate.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (10)
1. A cholelithiasis auxiliary detection system based on CT images, the system comprising:
the gallbladder CT image acquisition module acquires a gallbladder CT image;
the super-pixel segmentation module is used for performing super-pixel segmentation on the gallbladder CT image to obtain a super-pixel segmented image;
the similarity calculation module is used for obtaining the mass center of each super pixel block in the super pixel segmentation image, and taking the mass center as a target pixel point of each super pixel block; obtaining the probability of the target pixel point as the representative pixel point, and obtaining the representative pixel point of each super pixel block according to the probability of the target pixel point as the representative pixel point;
taking any two adjacent super-pixel blocks as a target super-pixel block pair, and acquiring a representative gray sequence and a representative gray curve according to representative pixel points of the two super-pixel blocks in the target super-pixel block pair; obtaining the local fluctuation degree of the representative gray scale curve according to the representative gray scale sequence; acquiring a smooth gray sequence according to the representative gray sequence; acquiring the weight of each gray value in the smooth gray sequence, and acquiring the overall fluctuation degree of the representative gray curve according to the weight; obtaining the attention weight of the local fluctuation degree according to the representative gray sequence, and obtaining the similarity between two super-pixel blocks in the target super-pixel block pair according to the local fluctuation degree, the whole fluctuation degree and the attention weight of the local fluctuation degree;
obtaining the similarity between all adjacent two super pixel blocks in the super pixel segmentation image;
the super pixel block merging module merges the super pixel blocks according to the similarity to obtain a merged image;
and the gall-stone identification module is used for identifying a gall-stone area according to the combined image.
2. The CT image-based gallstone auxiliary detection system as claimed in claim 1, wherein the obtaining the probability of the target pixel being the representative pixel, obtaining the representative pixel of each super pixel block according to the probability of the target pixel being the representative pixel, comprises the steps of:
obtaining the probability that the target pixel point is the representative pixel point:
wherein ,
representing the probability that the target pixel point is the representative pixel point; />
A gray value representing a target pixel; />
Represents the eighth +.>
Gray values of the individual pixels; />
Is an exponential function with a natural constant as a base;/>
is an absolute value symbol;
setting an empty candidate representative pixel point set; when the probability that the target pixel point is the representative pixel point is larger than the representative threshold value, adding the target pixel point into a candidate representative pixel point set, and taking the target pixel point with the maximum probability of being the representative pixel point in the candidate representative pixel point set as the representative pixel point of the super pixel point to which the target pixel point belongs; when the probability of the target pixel point being the representative pixel point is smaller than or equal to a representative threshold value, each pixel point in the eight adjacent areas of the target pixel point is respectively used as a new target pixel point, the probability of each target pixel point being the representative pixel point is obtained, all the target pixel points with the probability of being the representative pixel point being larger than the representative threshold value are added into a candidate representative pixel point set, and when the candidate representative pixel point set is not empty, the target pixel point with the maximum probability of being the representative pixel point in the candidate representative pixel point set is used as the representative pixel point of the super pixel point to which the target pixel point belongs; and when the probability that all the target pixel points are the representative pixel points is smaller than or equal to the representative threshold value, repeatedly acquiring new target pixel points, and the like, and stopping iteration until the representative pixel points of the super pixel blocks are acquired.
3. The auxiliary detection system for gall-stone based on CT image as set forth in claim 1, wherein the steps of obtaining the representative gray scale sequence and the representative gray scale curve according to the representative pixel points of two super-pixel blocks in the target super-pixel block pair include:
connecting the representative pixel points of two super pixel blocks in the target super pixel block pair, obtaining the gray values of all pixel points on the connecting line between the representative pixel points of the two super pixel blocks in the target super pixel block pair, and forming a representative gray sequence of the target super pixel block pair; and drawing a representative gray scale curve of the target super-pixel block pair by taking an index representing each gray scale value in the gray scale sequence as an abscissa and taking each gray scale value in the gray scale sequence as an ordinate.
4. The CT image-based gallstone auxiliary detection system as claimed in claim 1, wherein the step of obtaining the local fluctuation degree of the representative gray scale curve according to the representative gray scale sequence comprises the steps of:
counting the frequency of each gray value appearing in the representative gray sequence of the target super-pixel block pair, and acquiring the local fluctuation degree of the representative gray curve of the target super-pixel block pair according to the frequency of each gray value in the representative gray sequence:
wherein 
Local fluctuation degree of the representative gray scale curve of the target super pixel block pair; />
For the representative gray sequence of the target super-pixel block pair +.>
The frequency of the seed gray value; />
The number of classes of gray values in the representative gray sequence for the target superpixel block pair.
5. The CT image-based gallstone auxiliary detection system as claimed in claim 1, wherein the step of obtaining a smoothed gray scale sequence from a representative gray scale sequence comprises the steps of:
and calculating the average value of all gray values in each window in the sliding process of the sliding window, rounding the average value to be used as the average gray value of the window, and forming a one-dimensional sequence of the average gray values of all windows in the sliding process of the sliding window according to the sequence to be used as the smooth gray sequence of the target super-pixel block pair.
6. The CT image-based gallstone auxiliary detection system as claimed in claim 1, wherein the step of obtaining the weight of each gray value in the smoothed gray sequence comprises the steps of:
acquiring absolute values of differences between each gray value and two adjacent gray values in the smooth gray sequence, and respectively serving as a first difference value and a second difference value of each gray value; taking the average value of the first difference value and the second difference value of each gray value as the sequential difference value of each gray value;
combining the sequential difference value of each gray value in the smooth gray sequence of the target super-pixel block pair to obtain the weight of each gray value in the smooth gray sequence:
wherein ,
the smooth gray sequence for the target super-pixel block pair +.>
A weight for seeding gray values; />
The smooth gray sequence for the target super-pixel block pair +.>
Seed gray value +.>
Sequential differences in the next occurrence; />
For pairs of target superpixel blocksThe (th) in the smoothed gray sequence>
The number of times the seed gray value occurs in the smoothed gray sequence; />
The number of classes of gray values in the smoothed gray sequence for the target superpixel block pair.
7. The auxiliary detection system for gall stones based on CT images according to claim 1, wherein the step of obtaining the overall fluctuation degree of the representative gray scale curve according to the weight comprises the steps of:
counting the frequency of each gray value in the smooth gray sequence, and acquiring the overall fluctuation degree of the representative gray curve of the target super-pixel block pair according to the frequency and the weight of each gray value in the smooth gray sequence:
wherein 
The overall fluctuation degree of the representative gray level curve of the target super pixel block pair; />
The smooth gray sequence for the target super-pixel block pair +.>
The frequency of the seed gray value; />
The smooth gray sequence for the target super-pixel block pair +.>
A weight for seeding gray values; />
The number of classes of gray values in the smoothed gray sequence for the target superpixel block pair.
8. The CT image-based gallstone auxiliary detection system as claimed in claim 1, wherein the step of obtaining the attention weight of the local fluctuation degree according to the representative gray sequence comprises the steps of:
the gray value belonging to the first super pixel block in the target super pixel block pair in the representative gray sequence of the target super pixel block pair is called a first gray value, the gray value belonging to the second super pixel block is called a second gray value, and the nearest to the second gray value is obtained
The average value of the first gray values is used as the first boundary gray value to obtain the nearest +.>
The average value of the second gray values is used as the second boundary gray, the absolute value of the difference value between the first boundary gray and the second boundary gray is obtained, the range of the representative gray sequence is obtained, the ratio of the absolute value of the difference value between the first boundary gray and the second boundary gray to the range is used as the attention weight of the local fluctuation degree of the representative gray curve of the target super pixel block pair, wherein>
Is the preset boundary number.
9. The CT image-based gallstone auxiliary detection system as claimed in claim 1, wherein the step of obtaining the similarity between two super-pixel blocks in the target super-pixel block pair according to the local fluctuation degree, the overall fluctuation degree and the attention weight of the local fluctuation degree comprises the steps of:
wherein ,
for the similarity between two superpixel blocks in the target superpixel block pair: />
A focus weight representing the local fluctuation degree of the gray scale curve for the target super pixel block pair; />
The overall fluctuation degree of the representative gray level curve of the target super pixel block pair; />
Is the local fluctuation degree of the representative gray scale curve of the target super pixel block pair.
10. The CT image-based gallstone auxiliary detection system as claimed in claim 1, wherein the step of identifying the gallstone region from the combined image comprises the steps of:
and carrying out ellipse fitting on the edge of each super pixel block, and obtaining a gall-stone region according to the fitting error and the area of each super pixel block.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310645058.0A CN116385436B (en) | 2023-06-02 | 2023-06-02 | Cholelithiasis auxiliary detection system based on CT image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310645058.0A CN116385436B (en) | 2023-06-02 | 2023-06-02 | Cholelithiasis auxiliary detection system based on CT image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116385436A true CN116385436A (en) | 2023-07-04 |
| CN116385436B CN116385436B (en) | 2023-08-15 |
Family
ID=86971485
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310645058.0A Active CN116385436B (en) | 2023-06-02 | 2023-06-02 | Cholelithiasis auxiliary detection system based on CT image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116385436B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117830223A (en) * | 2023-12-04 | 2024-04-05 | 华南师范大学 | Kidney stone detection and assessment method and device based on CT flat scanning image |
| CN117911406A (en) * | 2024-03-19 | 2024-04-19 | 中国人民解放军空军军医大学 | Neck radiological image lesion area feature extraction method |
| CN118038064A (en) * | 2024-04-11 | 2024-05-14 | 青岛宝迈得生物科技有限公司 | Image segmentation method for hepatic duct and biliary tract calculus |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090226058A1 (en) * | 2008-03-05 | 2009-09-10 | Shenzhen Mindray Bio-Medical Electronics Co., Ltd. | Method and apparatus for tissue border detection using ultrasonic diagnostic images |
| US20180365808A1 (en) * | 2017-06-14 | 2018-12-20 | Shenzhen United Imaging Healthcare Co., Ltd. | System and method for image processing |
| CN109685814A (en) * | 2019-01-02 | 2019-04-26 | 兰州交通大学 | Cholecystolithiasis ultrasound image full-automatic partition method based on MSPCNN |
| CN111292307A (en) * | 2020-02-10 | 2020-06-16 | 刘肖 | Digestive system gallstone recognition method and positioning method |
| CN112233777A (en) * | 2020-11-19 | 2021-01-15 | 中国石油大学(华东) | Gallstone automatic identification and segmentation system based on deep learning, computer equipment and storage medium |
| CN115294338A (en) * | 2022-09-29 | 2022-11-04 | 中威泵业(江苏)有限公司 | Impeller surface defect identification method |
| CN115457031A (en) * | 2022-10-27 | 2022-12-09 | 江苏集宿智能装备有限公司 | Method for identifying internal defects of integrated box based on X-ray |
| CN115661017A (en) * | 2022-12-29 | 2023-01-31 | 华东交通大学 | Infrared thermal imaging super-pixel segmentation and fusion method, system, medium and computer |
| CN115861135A (en) * | 2023-03-01 | 2023-03-28 | 铜牛能源科技(山东)有限公司 | Image enhancement and identification method applied to box panoramic detection |
-
2023
- 2023-06-02 CN CN202310645058.0A patent/CN116385436B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090226058A1 (en) * | 2008-03-05 | 2009-09-10 | Shenzhen Mindray Bio-Medical Electronics Co., Ltd. | Method and apparatus for tissue border detection using ultrasonic diagnostic images |
| US20180365808A1 (en) * | 2017-06-14 | 2018-12-20 | Shenzhen United Imaging Healthcare Co., Ltd. | System and method for image processing |
| CN109685814A (en) * | 2019-01-02 | 2019-04-26 | 兰州交通大学 | Cholecystolithiasis ultrasound image full-automatic partition method based on MSPCNN |
| CN111292307A (en) * | 2020-02-10 | 2020-06-16 | 刘肖 | Digestive system gallstone recognition method and positioning method |
| CN112233777A (en) * | 2020-11-19 | 2021-01-15 | 中国石油大学(华东) | Gallstone automatic identification and segmentation system based on deep learning, computer equipment and storage medium |
| CN115294338A (en) * | 2022-09-29 | 2022-11-04 | 中威泵业(江苏)有限公司 | Impeller surface defect identification method |
| CN115457031A (en) * | 2022-10-27 | 2022-12-09 | 江苏集宿智能装备有限公司 | Method for identifying internal defects of integrated box based on X-ray |
| CN115661017A (en) * | 2022-12-29 | 2023-01-31 | 华东交通大学 | Infrared thermal imaging super-pixel segmentation and fusion method, system, medium and computer |
| CN115861135A (en) * | 2023-03-01 | 2023-03-28 | 铜牛能源科技(山东)有限公司 | Image enhancement and identification method applied to box panoramic detection |
Non-Patent Citations (2)
| Title |
|---|
| 廖苗;李阳;赵于前;刘毅志;: "一种新的图像超像素分割方法", 电子与信息学报, no. 02 * |
| 李盼;卜起荣;欧立奇;刘瀚;: "基于区域显著性的腹部CT图像目标检测", 计算机仿真, no. 05 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117830223A (en) * | 2023-12-04 | 2024-04-05 | 华南师范大学 | Kidney stone detection and assessment method and device based on CT flat scanning image |
| CN117911406A (en) * | 2024-03-19 | 2024-04-19 | 中国人民解放军空军军医大学 | Neck radiological image lesion area feature extraction method |
| CN117911406B (en) * | 2024-03-19 | 2024-06-04 | 中国人民解放军空军军医大学 | Neck radiological image lesion area feature extraction method |
| CN118038064A (en) * | 2024-04-11 | 2024-05-14 | 青岛宝迈得生物科技有限公司 | Image segmentation method for hepatic duct and biliary tract calculus |
| CN118038064B (en) * | 2024-04-11 | 2024-06-11 | 青岛宝迈得生物科技有限公司 | Image segmentation method for hepatic duct and biliary tract calculus |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116385436B (en) | 2023-08-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116385436B (en) | Cholelithiasis auxiliary detection system based on CT image | |
| CN109636808B (en) | Lung lobe segmentation method based on full convolution neural network | |
| WO2018120942A1 (en) | System and method for automatically detecting lesions in medical image by means of multi-model fusion | |
| US6154560A (en) | System and method for staging regional lymph nodes using quantitative analysis of endoscopic ultrasound images | |
| JP4184842B2 (en) | Image discrimination device, method and program | |
| JP2004032684A (en) | Automated method and apparatus for detecting mass or substantial tissue deformation in medical image using computer | |
| CN117649357B (en) | Ultrasonic image processing method based on image enhancement | |
| CN113643353B (en) | Measurement method for enhancing resolution of vascular caliber of fundus image | |
| CN117474823B (en) | CT data processing system for pediatric infectious inflammation detection assistance | |
| CN116402824B (en) | Endocrine abnormality detection method based on children bone age X-ray film | |
| CN110706241A (en) | Three-dimensional focus area extraction method and device | |
| CN112270667A (en) | TI-RADS-based integrated deep learning multi-tag identification method | |
| CN117877691A (en) | Intelligent wound information acquisition system based on image recognition | |
| CN116503426B (en) | Ultrasonic image segmentation method based on image processing | |
| CN116228767B (en) | X-ray lung tumor image processing method based on computer vision | |
| CN113658106A (en) | Liver focus automatic diagnosis system based on abdomen enhanced CT | |
| US20160100789A1 (en) | Computer-aided diagnosis system and computer-aided diagnosis method | |
| CN116128774B (en) | Stomach spiral CT data enhancement processing method | |
| Tong et al. | Computer-aided lung nodule detection based on CT images | |
| CN116071337A (en) | Endoscopic image quality evaluation method based on super-pixel segmentation | |
| JP2005034211A (en) | Image discrimination device, method using it and program | |
| CN115187592A (en) | Image segmentation method suitable for nuclear magnetic resonance image | |
| CN114240893A (en) | Method for measuring and calculating spinal Cobb angle in external image | |
| CN108921860B (en) | Full-automatic segmentation method for prostate magnetic resonance image | |
| CN112927225A (en) | Wisdom tooth growth state auxiliary detection system based on artificial intelligence |
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 |