CN108537785A - A kind of uterine ultrasound micro-creep method for processing video frequency - Google Patents
A kind of uterine ultrasound micro-creep method for processing video frequency Download PDFInfo
- Publication number
- CN108537785A CN108537785A CN201810295212.5A CN201810295212A CN108537785A CN 108537785 A CN108537785 A CN 108537785A CN 201810295212 A CN201810295212 A CN 201810295212A CN 108537785 A CN108537785 A CN 108537785A
- Authority
- CN
- China
- Prior art keywords
- micro
- creep
- uterus
- frequency domain
- video
- 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
- 238000002604 ultrasonography Methods 0.000 title claims abstract description 37
- 238000000034 method Methods 0.000 title claims abstract description 22
- 238000012545 processing Methods 0.000 title claims abstract description 12
- 210000004291 uterus Anatomy 0.000 claims abstract description 71
- 230000003321 amplification Effects 0.000 claims abstract description 7
- 238000003199 nucleic acid amplification method Methods 0.000 claims abstract description 7
- 230000002123 temporal effect Effects 0.000 claims abstract description 3
- 238000001914 filtration Methods 0.000 claims description 18
- 239000000284 extract Substances 0.000 claims description 12
- 229910002056 binary alloy Inorganic materials 0.000 claims description 10
- 238000001514 detection method Methods 0.000 claims description 9
- 230000000007 visual effect Effects 0.000 claims description 8
- 238000007781 pre-processing Methods 0.000 claims description 7
- 230000011218 segmentation Effects 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 claims description 5
- 238000010801 machine learning Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000003709 image segmentation Methods 0.000 claims description 2
- 238000011426 transformation method Methods 0.000 claims description 2
- 238000000354 decomposition reaction Methods 0.000 claims 1
- 210000004696 endometrium Anatomy 0.000 description 8
- 210000003205 muscle Anatomy 0.000 description 4
- 230000002572 peristaltic effect Effects 0.000 description 4
- 239000000523 sample Substances 0.000 description 4
- 238000012216 screening Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 239000000203 mixture Substances 0.000 description 3
- 210000002569 neuron Anatomy 0.000 description 3
- 238000003384 imaging method Methods 0.000 description 2
- 230000008855 peristalsis Effects 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000010009 beating Methods 0.000 description 1
- 210000004204 blood vessel Anatomy 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 230000002357 endometrial effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 230000008570 general process Effects 0.000 description 1
- 210000000936 intestine Anatomy 0.000 description 1
- 210000000982 limb bud Anatomy 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 210000004914 menses Anatomy 0.000 description 1
- 230000035935 pregnancy Effects 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 238000013442 quality metrics Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
- G06T9/007—Transform coding, e.g. discrete cosine transform
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10132—Ultrasound image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Discrete Mathematics (AREA)
- Multimedia (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of uterine ultrasound micro-creep method for processing video frequency.First, ultrasonic video is pre-processed；Then, the uterine region in ultrasonic video is decomposed by Multi scale and phase amplification carries out micro-creep and highlights, obtained uterus micro-creep and highlight video；In turn, the standard frequency domain characteristics dictionary based on preset 81 kinds of temporal and spatial orientations, statistics uterus micro-creep highlight the frequency disribution for meeting preset standard frequency domain character in video, and using frequency disribution as the micro-creep pattern in uterus；Uterus micro-creep is finally highlighted video and micro-creep pattern by display module output to be showed.This method the sightless uterus micro-creep of naked eyes can be become naked eyes as it can be seen that and the pattern of uterus micro-creep can be extracted, realize the feature qualitative reference to uterus micro-creep pattern.
Description
Technical field
It is the present invention relates to field of video processing, more particularly to a kind of that video is carried out to uterus micro-creep based on ultrasonic video
The method of processing.
Background technology
Endometrium shrinks caused uneven temper with muscle layer under endometrium and the asynchronous of uterine leio muscle layer
Internal pressure of uterine cavity power, and occur like the same mechanical movement of intestines peristalsis wave, referred to as Endometrial wavelike movements, i.e. endometrium is compacted
Dynamic wave.Endometrium is wriggled and micro-creep can reflect endometrium receptivity.The direction of endometrium peristaltic wave is a variety of more
Sample, have uterine neck to palace bottom, have palace bottom to uterine neck, also have wriggle at random etc..The transport of endometrium peristaltic wave and sperm,
Menses discharge, Embryonic limb bud cell and maintenance gestation are related, and the direction of peristaltic wave and frequency can also change with the variation of time.It removes
Except this, the asynchronization of muscle layer and uterine leio muscle layer is shunk under endometrium and the small beating of blood vessel all may cause
Existing uterus micro-creep.
It is currently understood that uterus peristalsis wave and the method for micro-creep are visually observed by ultrasonic video.Due to wriggling
Change faint, randomness is strong, generally requires waiting 5~20 minutes in the uterus video of conventional ultrasound equipment acquisition, could pass through
Carefully peristaltic wave is aware of in comparison, and for the micro-creep in uterus, often more it is not easy in ultrasonic video to see by naked eyes
It observes.
Invention content
In order to allow uterus micro-creep to become visualization and feature qualitative reference, this hair can be carried out to the pattern of uterus micro-creep
The bright uterine region by ultrasonic video is decomposed by Multi scale and phase amplification carries out micro-creep and highlights, and it is micro- to obtain uterus
Wriggling highlights video；In turn, the standard frequency domain characteristics dictionary based on preset 81 kinds of temporal and spatial orientations counts uterus micro-creep
The frequency disribution for meeting preset standard frequency domain character in video is highlighted, and using frequency disribution as the micro-creep pattern in uterus；Most
Uterus micro-creep highlights video and micro-creep pattern and is showed by display module output at last.Technical scheme is as follows：
A kind of uterine ultrasound micro-creep method for processing video frequency, as shown in Figures 1 and 2, by being formed with lower module：
1) ultrasound acquisition module, acquisition raw ultrasound video data V1；
2) preprocessing module carries out video denoising filtering to raw ultrasound video data and visual standardsization pre-processes, and
By pretreated ultrasound video data V2It is sent to uterine region detection module；
3) uterine region detection module, to V2In each frame image carry out sliding window region recognition, will be deemed as uterus
Region is integrated into uterus video data V in order3, and by V3It is sent to uterus micro-creep and highlights module；
4) uterus micro-creep highlights module, for V3Uterus micro-creep is carried out to highlight to obtain V4；
5) micro-creep pattern automatically extracts module, is responsible for extraction V4In micro-creep pattern Y；
6) display module is responsible for V4It is presented to user with micro-creep pattern Y.
Preferably, in order to acquire ultrasound video data, the acquisition mode of the ultrasound acquisition module is Type B ultrasound
Or harmonic imaging ultrasound, acquisition position are uterus, collect the raw ultrasound video data V comprising uterus1。
Preferably, described in order to meet the requirement for the nyquist sampling law analyzed uterus micro-creep frequency
Ultrasound acquisition module video data sample frequency be not less than 5Hz.
Preferably, in order to remove influence of the noise to video quality, the preprocessing module is by raw ultrasound video counts
According to V1Video filtering denoising is carried out, the method for video filtering denoising can be gaussian filtering, pass through discretization window sliding window convolution
To realize gaussian filtering.
Preferably, in order to ensure the image consistency of different video, the preprocessing module is by raw ultrasound video counts
According to V1Visual standards processing is carried out, visual standardsization processing can be histogram equalization.
Preferably, in order to reduce phase noise of the non-uterine region to uterine region, the uterine region detection module
For V2Each frame carry out uterine region segmentation, dividing method can be manual image segmentation or be based on machine learning method
The identification model that training obtains, using the segmentation result of each frame as the uterine region of the frame, by the video of only uterine region
Data are extracted as the uterus video data V detected3。
Preferably, uterus micro-creep is highlighted in order to realize, that is, the sightless uterus micro-creep of naked eyes is allowed to become naked eyes
It is made of change of scale group, filter, phase amplifier, inverse change of scale group as it can be seen that the uterus micro-creep highlights module；
Change of scale group is by uterus video data V3The video data of different scale is converted to, scale transformation method can be that small echo becomes
It changes, gaussian pyramid decomposes, can manipulation direction pyramid transform；Traffic filter receives the video data of different scale, and base
It is sent to phase amplifier after preset bandpass range carries out signal filtering to the video data of different scale；Phase amplifier
The signal filter result is received, and inverse change of scale group is sent to after carrying out the amplification of multiplying property to its phase；Inverse change of scale
Group receives the phase amplification of the different scale as a result, and carrying out obtaining uterus micro-creep after scale is inversely rebuild to it highlighting
Video data V4, uterus micro-creep is finally highlighted into video data V4It is sent to micro-creep pattern and automatically extracts module.
Preferably, in order to extract the time-space domain feature mode of uterus micro-creep, the micro-creep pattern automatically extracts mould
Block is made of frequency domain character encoder B, standard frequency domain characteristics dictionary D, Frequency statistics device；Uterus micro-creep highlights video data V4
Each frame image frequency domain data F is all converted to by Fourier transformi(i=1,2 ..., K), K is totalframes；All frames
Frequency domain data F obtains Q kind standard frequency domains feature (0 by frequency domain character encoder B and feature degeneracy method J<Q<=81, and Q is
Integer)；Frequency statistics device counts the frequency point of above-mentioned Q kinds standard frequency domain feature according to putting in order for standard frequency domain characteristics dictionary
Frequency disribution is sent to display module by cloth.
Preferably, in order to which the space-time characteristic to frequency domain is described, the frequency domain character encoder B is by the frequency domain of reception
Data F is divided into a series of identical subspace of sizes；Subspace size is a × b × c, and the range of a is 0.05~0.5M, b's
Range is 0.05~0.5N, and the range of c is 0.05~0.5K, and a, b, c are the positive integer after rounding up；Video V4Picture
Size is M × N, and totalframes K, M, N, K are positive integer；Maximum real number and maximum imaginary number per sub-spaces subspace thus
Maximum real number in interior all the points and maximum imaginary number；Own in minimum real number and minimum imaginary number thus subspace per sub-spaces
Minimum real number in point and minimum imaginary number；Intermediate value real number and intermediate value imaginary number per sub-spaces is thus in subspace in all the points
It is worth real number and intermediate value imaginary number；Arbitrary subspace constitutes 3 × 3 × 3 three-dimensional matrice with its neighborhood subspace, each position in matrix
It includes 3 plural numbers to set all, is that maximum real number and maximum imaginary number, minimum real number and minimum imaginary number, intermediate value real number and intermediate value are empty respectively
Number；Frequency domain character encoder by all domain complex of arbitrary subspace and its neighborhood subspace be restructured as in order one 1 ×
The one-dimensional vector of 162 (162=3 × 3 × 3 × 3 × 2), real number are closely aligned with corresponding imaginary number；To this 1 × 162 it is one-dimensional to
Amount each numerical value, be marked as 1 more than 0,0 be marked as less than or equal to 0, to obtain 1 × 162 containing only 1 and 0
One-dimensional vector, this one-dimensional vector be binary system frequency domain character coding.
Illustrated with three-dimensional 3 × 3 × 3 matrixes, number in order respectively as 1 to 27 sub-spaces, 2,3 ..., 26,27, most
Intermediate subspace is exactly the 14th sub-spaces, and neighborhood subspace is exactly remaining 26 son skies for surrounding most intermediate subspace
Between.
Preferably, in order to 2162A binary system frequency domain character coding is simplified, and the feature degeneracy method J is by 162
The binary system frequency domain character coding of position regards end to end annulus as, each binary system from counter clockwise direction traversal annulus
Number will occur from 0 to 1 or 1 to 0 saltus step total degree be denoted as Z；By the binary system frequency domain character degeneracy of the code with identical Z values
For same standard frequency domain feature, the wherein value range of Z is 0~80；All 81 kinds of frequency domain characters form standard frequency domain feature
Dictionary.
Illustrate so that 4 binary system frequency domain characters encode as an example.Such as 1000,0110,1010,0101 be all two into
Frequency domain character coding processed；1000,0110 all only has 2 saltus steps from 0 to 1 or from 1 to 0, they are just degenerated into and are classified as same class
Standard frequency domain feature is set as standard frequency domain feature having the same, number 2；1010,0101 all only have 4 times from 0 to 1 or
Saltus step from 1 to 0, they are just classified as another standard frequency domain feature, number 4；And so on.Why set in this way, is
Because of the saltus step having the same from 0 to 1 or from 1 to 0 of similar frequency domain character.In order to remove the redundancy of frequency domain character as possible
Property, and for reducing calculation amount and memory space, need to carry out degeneracy to frequency domain character.
The present invention has the advantages that：
1) the sightless uterine ultrasound micro-creep of naked eyes is become visually visible by the present invention；
2) present invention can extract the pattern of uterus micro-creep, provide feature for the pattern analysis of uterus micro-creep and quantify
With reference to.
Description of the drawings
Fig. 1 is the general process flow figure of present invention extraction uterus micro-creep pattern；
Fig. 2 is the detailed process figure of present invention extraction uterus micro-creep pattern；
Fig. 3 is that uterus micro-creep highlights front and back Spatial-temporal slice comparison diagram in the embodiment of the present invention；
Fig. 4 is the uterus micro-creep ideograph that everywoman is extracted in the embodiment of the present invention, and horizontal axis is frequency domain character serial number,
The longitudinal axis is the frequency of corresponding frequency domain character.
Specific implementation mode
The present invention will be further described by the following examples, to more fully understand technical scheme of the present invention, but
The present invention is not limited thereto.Ultrasonic device as used in the following examples, couplant etc., unless otherwise specified, can be from quotient
Industry approach obtains.
1. ultrasonic video acquires
1 everywoman of screening is included in the present embodiment after signing informed consent form.Using ordinary ultrasonic diagnostic equipment with
Transvaginal probe acquires the uterine ultrasound video of everywoman.Acquisition mode is harmonic imaging ultrasound, and acquisition position is uterus, acquisition
Frame per second is 30Hz, and center probe frequency is 10MHz.
2. pretreatment
Video denoising filtering and visual standardsization pretreatment are carried out to raw ultrasound video data.Preprocessing module is to original
Ultrasonic video carries out video filtering denoising, and the method choice gaussian filtering of video filtering denoising is rolled up by discretization window sliding window
It accumulates to realize that gaussian filtering, sliding window size are selected as 5 × 5 matrix.Preprocessing module carries out video filter to raw ultrasound video
Visual standards processing is carried out after wave denoising again, visual standards processing method selects histogram equalization.
3. uterine region detects
Ultrasonic video after the pre-treatment detects uterine region and obtains only including the video of uterine region.For pretreatment
Each frame of ultrasonic video afterwards carries out uterine region segmentation, the detection mould that dividing method is trained based on machine learning method
The video data of only uterine region is extracted conduct by type using the segmentation result of each frame as the uterine region of the frame
The uterus video data detected.The uterus detection model that this is trained based on machine learning method is by sliding window core group, uterus
Grader, cluster screening washer composition；Sliding window core group is made of 8 various sizes of sliding window cores, from every frame figure in above-mentioned video
A various sizes of small images are respectively cut as in as region of interest area image, until having traversed all image datas；Son
Palace grader is an artificial neural network, is made of 15-30 convolutional layer, each convolutional layer is by 100-250000 neuron
Composition, bottom neuron read multigroup region of interest area image for transmitting of sliding window core group, and top layer neuron are exported multigroup
Identification data are sent to cluster screening washer；It clusters screening washer to cluster for Mean-shift, Mean-shift search radius is 6 pictures
Element clusters the manifold classification data, using the cluster centre of each frame as the uterus center of the frame, will finally gather
The pericentral uterus video extraction of class is out as the uterus video detected.
4. uterus micro-creep highlights
Uterus micro-creep highlights module and is made of change of scale group, filter, phase amplifier, inverse change of scale group；Ruler
Degree transformation group uses wavelet transformation by uterus Video Quality Metric for the video of different scale；Traffic filter receives regarding for different scale
Frequency evidence, and sent after carrying out signal filtering to the video data of different scale based on preset 0.1~14.9Hz of bandpass range
To phase amplifier；Phase amplifier receives the signal filter result, and is sent out after carrying out 10 times of multiplying property amplifications to its phase
Give inverse change of scale group；The phase that inverse change of scale group receives the different scale is amplified as a result, and carrying out scale to it
Uterus micro-creep is obtained after reverse reconstruction and highlights video, and is sent to display module.
Fig. 3 is that uterus micro-creep highlights front and back Spatial-temporal slice comparison diagram in the embodiment of the present invention.Fig. 3 .A are a typical cases
The original uterine ultrasound video of women, dotted line are the single-row position in longitudinal direction of Spatial-temporal slice acquisition；Fig. 3 .B are that corresponding uterus is micro- compacted
Dynamic to highlight video, dotted line is the single-row position in longitudinal direction of the identical Spatial-temporal slice acquisition with Fig. 3 .A；Take in video one it is longitudinal sectional
As soon as pixel be unfolded and be superposed to width figure on a timeline, Spatial-temporal slice figure can be obtained.The horizontal axis of Spatial-temporal slice figure is video
The 1st, 2 ..., k-th frame, the longitudinal axis is corresponding 1,2 ..., the single column of pixels of the same position of k-th frame, wherein K is video
Totalframes.Fig. 3 .C are the Spatial-temporal slice figure of the uterine ultrasound video of original uterine ultrasound video；Fig. 3 .D are that corresponding uterus is micro- compacted
The dynamic Spatial-temporal slice figure for highlighting video.
5. micro-creep pattern automatically extracts
Micro-creep pattern automatically extracts module by frequency domain character encoder, standard frequency domain characteristics dictionary, Frequency statistics device structure
At；Uterus micro-creep highlights video data V4Each frame image frequency domain data F is all converted to by Fourier transformi(i=1,
2 ..., 300), 300 be video totalframes；The frequency domain data F of all frames passes through frequency domain character encoder B and feature degeneracy method J
Obtain 40 kinds of standard frequency domain features；Frequency statistics device counts above-mentioned 40 kinds of standards according to putting in order for standard frequency domain characteristics dictionary
Frequency disribution is sent to display module by the frequency disribution of frequency domain character.
Frequency domain character encoder receives the frequency domain data F, and is divided into the subspace of a series of same size；Depending on
Frequency picture size is 200 × 200, totalframes 300；Subspace size is 20 × 20 × 30.Maximum real number per sub-spaces
With the maximum real number and maximum imaginary number in all the points in maximum imaginary number thus subspace；Minimum real number and minimum per sub-spaces
The imaginary number minimum real number in subspace in all the points and minimum imaginary number thus；Intermediate value real number and intermediate value imaginary number per sub-spaces are
The intermediate value real number Yu intermediate value imaginary number of all the points in this subspace；Arbitrary subspace and the three of its neighborhood subspace composition 3 × 3 × 3
Matrix is tieed up, each position includes 3 plural numbers in matrix, is maximum real number and maximum imaginary number, minimum real number and minimum respectively
Imaginary number, intermediate value real number and intermediate value imaginary number.
All domain complex of arbitrary subspace and its neighborhood subspace are restructured as one by frequency domain character encoder in order
The one-dimensional vector of a 1 × 162 (162=3 × 3 × 3 × 3 × 2), real number are closely aligned with corresponding imaginary number；To this 1 × 162
Each numerical value of one-dimensional vector, 1 is marked as more than 0, and 0 is marked as less than or equal to 0, to obtain 1 × 162
Containing 1 and 0 one-dimensional vector, this one-dimensional vector is binary system frequency domain character coding.
1 everywoman of the present embodiment pair carries out uterine ultrasound video acquisition, and processing method according to the present invention is to super
After sound video data is handled, result as shown in Figure 3 and Figure 4 is obtained.Fig. 3 is that uterus micro-creep is convex in the embodiment of the present invention
Front and back Spatial-temporal slice comparison diagram is shown, shows that the present invention can effectively highlight uterus micro-creep.Fig. 4 is implementation of the present invention
Example in extract everywoman uterus micro-creep ideograph, show that the present invention can effectively extract uterus micro-creep pattern, for into
The comparison of row micro-creep pattern provides quantitative characteristic reference.
Claims (8)
1. a kind of uterine ultrasound micro-creep method for processing video frequency, by being formed with lower module：
Ultrasound acquisition module (1), acquisition raw ultrasound video data V1；
Preprocessing module (2) carries out video denoising filtering to raw ultrasound video data and visual standardsization pre-processes, and will be pre-
Treated ultrasound video data V2It is sent to uterine region detection module (3)；
Uterine region detection module (3), to V2In each frame image carry out sliding window region recognition, will be deemed as the region in uterus
It is integrated into uterus video data V in order3, and by V3It is sent to uterus micro-creep and highlights module (4)；
Uterus micro-creep highlights module (4), for V3Uterus micro-creep is carried out to highlight to obtain V4；
Micro-creep pattern automatically extracts module (5), is responsible for extraction V4In micro-creep pattern Y；
Display module (6) is responsible for V4It is presented to user with micro-creep pattern Y；
It is characterized in that, the uterine region in ultrasonic video is convex by Multi scale decomposition and phase amplification progress micro-creep
It is aobvious, it obtains uterus micro-creep and highlights video V4；In turn, the standard frequency domain characteristics dictionary D based on default 81 kinds of temporal and spatial orientations,
Count V4In meet the frequency disribution of preset standard frequency domain character, and using frequency disribution as the micro-creep pattern Y in uterus；Finally
By V4Showed by display module output with Y.
2. ultrasound acquisition module (1) according to claim 1, which is characterized in that acquisition mode be Type B ultrasound or harmonic wave at
As ultrasound, acquisition position is uterus, collects the raw ultrasound video data V comprising uterus1, the sample frequency of video data
Not less than 5Hz.
3. preprocessing module (2) according to claim 1, which is characterized in that by raw ultrasound video data V1Carry out video
Filtering and noise reduction and visual standardsization processing, by treated ultrasound video data V2It is sent to uterine region detection module
(3)；Wherein, the method for video filtering denoising can be gaussian filtering G, and visual standardsization processing can be histogram equalization H.
4. uterine region detection module (3) according to claim 1, which is characterized in that be directed to V2Each frame carry out uterus
Region segmentation, dividing method can be manual image segmentation, markov random file segmentation or be trained based on machine learning method
Obtained identification model, using the segmentation result of each frame as the uterine region of the frame, by the video data of only uterine region
It extracts as the uterus video data V detected3。
5. uterus micro-creep according to claim 1 highlights module (4), which is characterized in that by change of scale group, filtering
Device, phase amplifier, inverse change of scale group are constituted；Change of scale group is by uterus video data V3Be converted to the video of different scale
Data, scale transformation method can be wavelet transformation, gaussian pyramid decompose, can manipulation direction pyramid transform；Signal filters
Device receives the video data of different scale, and carries out signal filtering to the video data of different scale based on preset bandpass range
After be sent to phase amplifier；Phase amplifier receives the signal filter result, and after carrying out the amplification of multiplying property to its phase
It is sent to inverse change of scale group；The phase that inverse change of scale group receives the different scale is amplified as a result, and carrying out ruler to it
Uterus micro-creep, which is obtained, after the reverse reconstruction of degree highlights video data V4, uterus micro-creep is finally highlighted into video data V4It is sent to
Micro-creep pattern automatically extracts module (5).
6. micro-creep pattern according to claim 1 automatically extracts module (5), which is characterized in that encoded by frequency domain character
Device B, standard frequency domain characteristics dictionary D, Frequency statistics device are constituted；Uterus micro-creep highlights video data V4Each frame image it is all logical
It crosses Fourier transform and is converted to frequency domain data Fi(i=1,2 ..., K), K is totalframes；The frequency domain data F of all frames passes through frequency domain
Feature coding device B and feature degeneracy method J obtain Q kind standard frequency domains feature (0<Q<=81, and Q is integer)；Frequency statistics device
The frequency disribution of above-mentioned Q kinds standard frequency domain feature is counted according to putting in order for standard frequency domain characteristics dictionary, frequency disribution is made
It is sent to display module (6) for micro-creep pattern Y.
7. frequency domain character encoder according to claim 6, which is characterized in that the frequency domain data F of reception is divided into one
The identical subspace of serial size；Subspace size is a × b × c, and the range of a is 0.05~0.5M, the range of b is 0.05~
The range of 0.5N, c are 0.05~0.5K, and a, b, c are the positive integer after rounding up；Video V4Picture size be M × N,
Totalframes is K, and M, N, K are positive integer；Maximum real number and maximum imaginary number per sub-spaces are thus in subspace in all the points
Maximum real number and maximum imaginary number；Minimum real number per sub-spaces and the minimum in all the points in minimum imaginary number thus subspace
Real number and minimum imaginary number；Intermediate value real number and intermediate value imaginary number per sub-spaces thus in subspace the intermediate value real number of all the points in
It is worth imaginary number；Arbitrary subspace constitutes 3 × 3 × 3 three-dimensional matrice with its neighborhood subspace, each position includes 3 in matrix
A plural number is maximum real number and maximum imaginary number, minimum real number and minimum imaginary number, intermediate value real number and intermediate value imaginary number respectively；Frequency domain is special
All domain complex of arbitrary subspace and its neighborhood subspace are restructured as 1 × 162 (162=3 by sign encoder in order
× 3 × 3 × 3 × 2) one-dimensional vector, real number are closely aligned with corresponding imaginary number；To each of this 1 × 162 one-dimensional vector
Numerical value, 1 is marked as more than 0, and 0 is marked as less than or equal to 0, to obtain 1 × 162 containing only 1 and 0 it is one-dimensional to
Amount, this one-dimensional vector are binary system frequency domain character coding.
8. feature degeneracy method according to claim 6, which is characterized in that encode 162 binary system frequency domain characters
Regard end to end annulus as, each binary number from counter clockwise direction traversal annulus will occur from 0 to 1 or 1 to 0
Saltus step total degree is denoted as Z；It is same standard frequency domain feature by the binary system frequency domain character degeneracy of the code with identical Z values,
The value range of middle Z is 0~80；All 81 kinds of frequency domain characters form standard frequency domain characteristics dictionary.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810295212.5A CN108537785B (en) | 2018-04-04 | 2018-04-04 | Uterine ultrasound micro-peristalsis video processing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810295212.5A CN108537785B (en) | 2018-04-04 | 2018-04-04 | Uterine ultrasound micro-peristalsis video processing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108537785A true CN108537785A (en) | 2018-09-14 |
| CN108537785B CN108537785B (en) | 2021-04-30 |
Family
ID=63482633
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810295212.5A Expired - Fee Related CN108537785B (en) | 2018-04-04 | 2018-04-04 | Uterine ultrasound micro-peristalsis video processing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108537785B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020070504A1 (en) * | 2018-10-03 | 2020-04-09 | Cmr Surgical Limited | Indicator system |
| CN112587089A (en) * | 2020-11-19 | 2021-04-02 | 新希望六和股份有限公司 | Pregnancy detecting method, apparatus, computer device and medium based on artificial intelligence |
| CN112704517A (en) * | 2021-02-02 | 2021-04-27 | 深圳华声医疗技术股份有限公司 | Method, system, equipment and storage medium for processing endometrium peristalsis ultrasonic image |
| WO2021087765A1 (en) * | 2019-11-05 | 2021-05-14 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic imaging device and method for detecting endometrial peristalsis |
| CN114431893A (en) * | 2020-10-30 | 2022-05-06 | 浙江省人民医院 | Parameter measuring method of peristaltic wave and ultrasonic measuring system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102526065A (en) * | 2011-12-26 | 2012-07-04 | 西安力邦制药有限公司 | Compound injection preparation for treating cardiovascular and cerebrovascular diseases and preparation method thereof |
| US20150004214A1 (en) * | 2011-12-29 | 2015-01-01 | Universidad De Chile | Vaginal ring comprising dhea or dhea sulfate and optionally a release-modulating agent of the active principle, useful to increase the ovarian reserve in women and to relieve symptoms associated with menopause |
| US9326722B2 (en) * | 2011-03-22 | 2016-05-03 | University Of Vermont And State Agricultural College | Methods of predicting and monitoring labor |
| CN107028657A (en) * | 2011-11-04 | 2017-08-11 | 波士顿科学医疗有限公司 | Tissue extraction apparatus and method |
| CN107209934A (en) * | 2014-12-03 | 2017-09-26 | 文塔纳医疗系统公司 | Mthods, systems and devices for quantitative analysis heterogeneous biological marker profile |
| CN107492117A (en) * | 2017-07-27 | 2017-12-19 | 北京大学 | A kind of micro- vibration function based on video is as imaging method |
-
2018
- 2018-04-04 CN CN201810295212.5A patent/CN108537785B/en not_active Expired - Fee Related
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9326722B2 (en) * | 2011-03-22 | 2016-05-03 | University Of Vermont And State Agricultural College | Methods of predicting and monitoring labor |
| CN107028657A (en) * | 2011-11-04 | 2017-08-11 | 波士顿科学医疗有限公司 | Tissue extraction apparatus and method |
| CN102526065A (en) * | 2011-12-26 | 2012-07-04 | 西安力邦制药有限公司 | Compound injection preparation for treating cardiovascular and cerebrovascular diseases and preparation method thereof |
| US20150004214A1 (en) * | 2011-12-29 | 2015-01-01 | Universidad De Chile | Vaginal ring comprising dhea or dhea sulfate and optionally a release-modulating agent of the active principle, useful to increase the ovarian reserve in women and to relieve symptoms associated with menopause |
| CN107209934A (en) * | 2014-12-03 | 2017-09-26 | 文塔纳医疗系统公司 | Mthods, systems and devices for quantitative analysis heterogeneous biological marker profile |
| CN107492117A (en) * | 2017-07-27 | 2017-12-19 | 北京大学 | A kind of micro- vibration function based on video is as imaging method |
Non-Patent Citations (1)
| Title |
|---|
| 刘珊珊: "电影MRI对子宫内膜蠕动的研究进展", 《中国医学影像技术》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2022511604A (en) * | 2018-10-03 | 2022-02-01 | シーエムアール・サージカル・リミテッド | Indicator system |
| WO2020070504A1 (en) * | 2018-10-03 | 2020-04-09 | Cmr Surgical Limited | Indicator system |
| JP7500667B2 (en) | 2018-10-03 | 2024-06-17 | シーエムアール・サージカル・リミテッド | Indicator System |
| CN112804958A (en) * | 2018-10-03 | 2021-05-14 | Cmr外科有限公司 | Indicator system |
| JP7308936B2 (en) | 2018-10-03 | 2023-07-14 | シーエムアール・サージカル・リミテッド | indicator system |
| JP2022145863A (en) * | 2018-10-03 | 2022-10-04 | シーエムアール・サージカル・リミテッド | indicator system |
| WO2021087765A1 (en) * | 2019-11-05 | 2021-05-14 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic imaging device and method for detecting endometrial peristalsis |
| CN114025672A (en) * | 2019-11-05 | 2022-02-08 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic imaging equipment and method for detecting endometrial peristalsis |
| CN114025672B (en) * | 2019-11-05 | 2024-05-14 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic imaging equipment and detection method for endometrial peristalsis |
| CN114431893A (en) * | 2020-10-30 | 2022-05-06 | 浙江省人民医院 | Parameter measuring method of peristaltic wave and ultrasonic measuring system |
| CN112587089A (en) * | 2020-11-19 | 2021-04-02 | 新希望六和股份有限公司 | Pregnancy detecting method, apparatus, computer device and medium based on artificial intelligence |
| CN112587089B (en) * | 2020-11-19 | 2023-04-21 | 新希望六和股份有限公司 | Pregnancy detection method, device, computer equipment and medium based on artificial intelligence |
| CN112704517B (en) * | 2021-02-02 | 2021-08-03 | 深圳华声医疗技术股份有限公司 | Method, system, equipment and storage medium for processing endometrium peristalsis ultrasonic image |
| CN112704517A (en) * | 2021-02-02 | 2021-04-27 | 深圳华声医疗技术股份有限公司 | Method, system, equipment and storage medium for processing endometrium peristalsis ultrasonic image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108537785B (en) | 2021-04-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108537785A (en) | A kind of uterine ultrasound micro-creep method for processing video frequency | |
| CN109389585B (en) | Brain tissue extraction method based on full convolution neural network | |
| CN109784242A (en) | EEG Noise Cancellation based on one-dimensional residual error convolutional neural networks | |
| CN108717717A (en) | The method rebuild based on the sparse MRI that convolutional neural networks and alternative manner are combined | |
| CN109191457A (en) | A kind of pathological image quality validation recognition methods | |
| WO2021184805A1 (en) | Blood pressure prediction method and device using multiple data sources | |
| CN107424145A (en) | The dividing method of nuclear magnetic resonance image based on three-dimensional full convolutional neural networks | |
| CN113516601A (en) | Image restoration technology based on deep convolutional neural network and compressed sensing | |
| CN111080531A (en) | Super-resolution reconstruction method, system and device for underwater fish image | |
| CN111832431A (en) | Emotional electroencephalogram classification method based on CNN | |
| Wang et al. | Reduction of Gibbs artifacts in magnetic resonance imaging based on Convolutional Neural Network | |
| CN104102839A (en) | Alzheimer's disease cortex auto-classification method based on multi-scale mesh surface form features | |
| Mina et al. | Breast abnormality detection in mammograms using Artificial Neural Network | |
| CN112967185A (en) | Image super-resolution algorithm based on frequency domain loss function | |
| CN115661459A (en) | 2D mean teacher model using difference information | |
| CN113392937A (en) | 3D point cloud data classification method and related device thereof | |
| CN112233199A (en) | fMRI visual reconstruction method based on discrete characterization and conditional autoregression | |
| CN115527657A (en) | Image and image multi-mode reconstruction, imaging and labeling based on medical digital imaging and communication | |
| CN116758336A (en) | Medical image intelligent analysis system based on artificial intelligence | |
| CN112037304A (en) | Two-stage edge enhancement QSM reconstruction method based on SWI phase image | |
| CN109965905B (en) | Contrast region detection imaging method based on deep learning | |
| CN112515653A (en) | Brain network construction method based on nuclear magnetic resonance image | |
| Zeng et al. | Perceptual evaluation of image denoising algorithms | |
| CN111047537A (en) | System for recovering details in image denoising | |
| CN116468818A (en) | MRI image reconstruction method based on sparse representation and Bayesian estimation |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210430 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |