CN108537785B - Uterine ultrasound micro-peristalsis video processing method - Google Patents
Uterine ultrasound micro-peristalsis video processing method Download PDFInfo
- Publication number
- CN108537785B CN108537785B CN201810295212.5A CN201810295212A CN108537785B CN 108537785 B CN108537785 B CN 108537785B CN 201810295212 A CN201810295212 A CN 201810295212A CN 108537785 B CN108537785 B CN 108537785B
- Authority
- CN
- China
- Prior art keywords
- micro
- peristalsis
- uterine
- 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.)
- Expired - Fee Related
Links
- 238000002604 ultrasonography Methods 0.000 title claims abstract description 37
- 238000003672 processing method Methods 0.000 title claims abstract description 14
- 210000004291 uterus Anatomy 0.000 claims abstract description 37
- 238000000034 method Methods 0.000 claims abstract description 19
- 238000007781 pre-processing Methods 0.000 claims abstract description 13
- 238000009826 distribution Methods 0.000 claims abstract description 11
- 230000003321 amplification Effects 0.000 claims abstract description 10
- 238000003199 nucleic acid amplification method Methods 0.000 claims abstract description 10
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 5
- 238000001914 filtration Methods 0.000 claims description 24
- 230000009466 transformation Effects 0.000 claims description 18
- 238000000605 extraction Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 10
- 230000000007 visual effect Effects 0.000 claims description 10
- 238000001514 detection method Methods 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 8
- 230000002572 peristaltic effect Effects 0.000 claims description 6
- 238000010801 machine learning Methods 0.000 claims description 4
- 230000011218 segmentation Effects 0.000 claims description 4
- 238000012549 training Methods 0.000 claims description 4
- 238000003384 imaging method Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 230000009191 jumping Effects 0.000 claims description 2
- 238000011426 transformation method Methods 0.000 claims description 2
- 241000638935 Senecio crassissimus Species 0.000 claims 1
- 238000003709 image segmentation Methods 0.000 claims 1
- 230000008855 peristalsis Effects 0.000 description 8
- 230000002357 endometrial effect Effects 0.000 description 4
- 210000004696 endometrium Anatomy 0.000 description 4
- 210000002569 neuron Anatomy 0.000 description 3
- 206010034647 Peristalsis visible Diseases 0.000 description 2
- 230000008602 contraction Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 239000000523 sample Substances 0.000 description 2
- 210000002460 smooth muscle Anatomy 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000008280 blood Substances 0.000 description 1
- 210000004369 blood Anatomy 0.000 description 1
- 210000004204 blood vessel Anatomy 0.000 description 1
- 239000007822 coupling agent Substances 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 230000007850 degeneration Effects 0.000 description 1
- 230000032692 embryo implantation Effects 0.000 description 1
- 210000001035 gastrointestinal tract Anatomy 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000002175 menstrual effect Effects 0.000 description 1
- 210000003205 muscle Anatomy 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000035935 pregnancy Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000010349 pulsation Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 230000007704 transition 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
- 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
-
- G06T5/70—
-
- 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
Abstract
The invention discloses a uterine ultrasound micro-peristalsis video processing method. Firstly, preprocessing an ultrasonic video; then, carrying out micro-peristalsis highlighting on the uterine region in the ultrasonic video through space multi-scale decomposition and phase amplification to obtain a uterine micro-peristalsis highlighting video; further, counting frequency distribution meeting the preset standard frequency domain characteristics in the uterine micro-peristalsis highlighting video based on a preset 81 standard frequency domain characteristic dictionaries with space-time dynamic changes, and taking the frequency distribution as a uterine micro-peristalsis mode; finally, the uterus micro-peristalsis highlighting video and the micro-peristalsis mode are output and displayed by the display module. The method can change the uterine micro-peristalsis invisible to the naked eye into the visible state, and can extract the uterine micro-peristalsis mode, thereby realizing the quantitative reference of the characteristics of the uterine micro-peristalsis mode.
Description
Technical Field
The invention relates to the field of video processing, in particular to a method for carrying out video processing on uterine micro-peristalsis based on ultrasonic video.
Background
The mechanical motion of the endometrium, which is called the endometrium wave motion, is similar to the intestinal tract peristaltic wave along with the non-uniform intrauterine pressure caused by the asynchronous contraction of the endometrium muscle layer and the uterine smooth muscle layer. Endometrial peristalsis and microtomosis can reflect endometrial receptivity. The direction of the endometrial peristaltic waves is diverse, including cervical to fundus, fundus to cervical, and random peristaltic, among others. The direction and frequency of the endometrial peristaltic waves are also changed with time, in relation to sperm transport, menstrual blood discharge, embryo implantation and maintenance of pregnancy. In addition, the microsclerical movement of the uterus may be caused by asynchronous contraction of the endometrium and the uterine smooth muscle layers and by the minute pulsation of the blood vessels.
The current method of understanding uterine peristalsis waves and micromotion is by visual observation through ultrasound video. Because the peristalsis changes weakly and has strong randomness, the uterine video acquired by the traditional ultrasonic equipment usually needs to wait for 5-20 minutes, so that the peristalsis waves can be observed through careful comparison, and the micro peristalsis of the uterus is usually not easy to observe through naked eyes in the ultrasonic video.
Disclosure of Invention
In order to enable the uterine micro-peristalsis to become visual and carry out characteristic quantitative reference on the uterine micro-peristalsis mode, the method carries out micro-peristalsis highlighting on a uterine region in an ultrasonic video through spatial multi-scale decomposition and phase amplification to obtain a uterine micro-peristalsis highlighting video; further, counting frequency distribution meeting the preset standard frequency domain characteristics in the uterine micro-peristalsis highlighting video based on a preset 81 standard frequency domain characteristic dictionaries with space-time dynamic changes, and taking the frequency distribution as a uterine micro-peristalsis mode; finally, the uterus micro-peristalsis highlighting video and the micro-peristalsis mode are output and displayed by the display module. The technical scheme of the invention is as follows:
a uterine ultrasound micro-peristalsis video processing method, as shown in fig. 1 and fig. 2, comprising the following modules:
1) an ultrasound acquisition module for acquiring original ultrasound video data V1；
2) Preparation ofA processing module for performing video denoising filtering and visual standardization preprocessing on the original ultrasonic video data and preprocessing the preprocessed ultrasonic video data V2Sending the data to a uterus region detection module;
3) uterine region detection Module, for V2Performing sliding window area identification on each frame of image, and integrating the areas judged as the uterus into uterus video data V in sequence3And will V3Sending the data to a uterus micro-peristalsis highlighting module;
4) uterus micro-peristalsis highlighting module for V3The uterus micro-peristalsis is performed to obtain V4；
5) A micro-peristalsis mode automatic extraction module for extracting V4Micro-peristalsis mode Y;
6) a display module responsible for displaying V4And a micro-peristalsis pattern Y is presented to the user.
Preferably, in order to acquire the ultrasound video data, the acquisition mode of the ultrasound acquisition module is B-mode ultrasound or harmonic imaging ultrasound, the acquisition site is uterus, and the acquisition module acquires original ultrasound video data V including the uterus1。
Preferably, in order to meet the requirement of the nyquist sampling law for analyzing the uterine micro-peristalsis frequency, the sampling frequency of the video data of the ultrasonic acquisition module is not less than 5 Hz.
Preferably, in order to remove the influence of noise on the video quality, the preprocessing module is used for processing the original ultrasonic video data V1And (3) carrying out video filtering and denoising, wherein the video filtering and denoising method can be Gaussian filtering, and the Gaussian filtering is realized by discretizing window sliding window convolution.
Preferably, in order to ensure the image consistency of different videos, the preprocessing module is used for processing the original ultrasonic video data V1A visual normalization process is performed, which may be histogram equalization.
Preferably, in order to reduce phase noise of non-uterine regions to uterine regions, the uterine region detection module is used for V2Each frame of the image is divided into uterus regions by a manual image division methodCutting or training based on machine learning method to obtain identification model, using the cutting result of each frame as the uterus region of the frame, extracting the video data of only uterus region as the detected uterus video data V3。
Preferably, in order to realize the highlighting of the uterine micro peristalsis, namely, the uterine micro peristalsis invisible to the naked eye is changed into the uterine micro peristalsis visible to the naked eye, the uterine micro peristalsis highlighting module consists of a scale transformation group, a filter, a phase amplifier and an inverse scale transformation group; scale transformation set transforms uterine video data V3Converting into video data with different scales, wherein the scale transformation method can be wavelet transformation, Gaussian pyramid decomposition and steerable direction pyramid transformation; the signal filter receives video data of different scales, performs signal filtering on the video data of different scales based on a preset band-pass range, and sends the video data of different scales to the phase amplifier; the phase amplifier receives the signal filtering result, multiplicatively amplifies the phase of the signal filtering result and sends the amplified signal to the inverse scale transformation group; the inverse scale transformation group receives the phase amplification results of different scales and carries out scale inverse reconstruction on the phase amplification results to obtain uterine micro-peristalsis highlighted video data V4Finally, the uterine micro-peristalsis is highlighted to display the video data V4And sending the data to a micro-peristalsis mode automatic extraction module.
Preferably, in order to extract a time-space domain characteristic mode of uterine micro-peristalsis, the automatic micro-peristalsis mode extraction module consists of a frequency domain characteristic encoder B, a standard frequency domain characteristic dictionary D and a frequency statistics device; uterine micro-peristalsis highlighting video data V4Each frame of image is converted into frequency domain data F by fourier transformi(i ═ 1,2, …, K), K being the total frame number; obtaining Q standard frequency domain characteristics (0) from the frequency domain data F of all frames through a frequency domain characteristic encoder B and a characteristic degeneracy method J<Q<81 and Q is an integer); the frequency counting device counts the frequency distribution of the Q standard frequency domain characteristics according to the arrangement sequence of the standard frequency domain characteristic dictionary, and sends the frequency distribution to the display module as a micro-creeping mode Y.
Preferably, the frequency-domain feature encoder B divides the received frequency-domain data F to describe the time-space features of the frequency domainA series of subspaces of the same size; the subspace has a size of axbxc, a ranges from 0.05M to 0.5M, b ranges from 0.05N to 0.5N, c ranges from 0.05K to 0.5K, and a, b and c are rounded positive integers; video V4The picture size of (1) is M multiplied by N, the total frame number is K, and M, N, K are positive integers; the maximum real number and the maximum imaginary number of each subspace are the maximum real number and the maximum imaginary number of all points in the subspace; the minimum real number and the minimum imaginary number of each subspace are the minimum real number and the minimum imaginary number of all points in the subspace; the median real number and median imaginary number of each subspace are the median real number and median imaginary number of all points in the subspace; any subspace and a neighborhood subspace form a 3 multiplied by 3 three-dimensional matrix, each position in the matrix comprises 3 complex numbers which are respectively a maximum real number and a maximum imaginary number, a minimum real number and a minimum imaginary number, and a median real number and a median imaginary number; a frequency domain feature encoder sequentially reforms all frequency domain complex numbers of any subspace and a neighborhood subspace thereof into a one-dimensional vector of 1 × 162(162 is 3 × 3 × 3 × 3 × 2), and real numbers and corresponding imaginary numbers are arranged closely; for each value of the 1 × 162 one-dimensional vector, a value greater than 0 is marked as 1, and a value less than or equal to 0 is marked as 0, so as to obtain a 1 × 162 one-dimensional vector containing only 1 and 0, which is the binary frequency domain feature code.
Taking a three-dimensional 3 × 3 × 3 matrix as an example, 27 subspaces are numbered 1,2, 3, … …, 26, and 27 in sequence, the most intermediate subspace is the 14 th subspace, and the neighboring subspace is the remaining 26 subspaces surrounding the most intermediate subspace.
Preferably, in order to pair 2162Simplifying the binary frequency domain feature codes, wherein the feature degeneration method J takes 162-bit binary frequency domain feature codes as end-to-end circular rings, traverses each binary number on the circular rings from the counterclockwise direction, and records the total number of jumping from 0 to 1 or 1 to 0 as Z; degenerating binary frequency domain feature codes with the same Z value into the same standard frequency domain feature, wherein the value range of Z is 0-80; all 81 frequency domain features constitute a standard frequency domain feature dictionary.
The 4-bit binary frequency domain feature coding is taken as an example for explanation. For example, 1000, 0110, 1010, 0101 are all binary frequency domain feature codes; 1000. 0110 there are only 2 jumps from 0 to 1 or from 1 to 0, they are degenerated to the same standard frequency domain feature, and set to have the same standard frequency domain feature, number 2; 1010. 0101, they are classified as another standard frequency domain feature, numbered 4, with only 4 jumps from 0 to 1 or from 1 to 0; and so on. This is so because similar frequency domain features have the same transition from 0 to 1 or 1 to 0. In order to remove redundancy of the frequency domain features as much as possible and reduce the amount of computation and storage space, degeneracy of the frequency domain features is required.
The invention has the following beneficial effects:
1) the uterine ultrasound micro-peristalsis invisible to naked eyes is changed into the uterine ultrasound micro-peristalsis visible to the naked eyes;
2) the invention can extract the uterine micro-peristalsis mode and provides a characteristic quantitative reference for the analysis of the uterine micro-peristalsis mode.
Drawings
FIG. 1 is a flow chart of a generalized process for extracting uterine micro-peristalsis patterns according to the present invention;
FIG. 2 is a flow chart of a detailed process for extracting a uterine micro-peristalsis pattern according to the present invention;
FIG. 3 is a comparison of spatiotemporal slices before and after the uterine micro-peristalsis is highlighted in an example of the invention;
fig. 4 is a diagram of a uterine micro-peristalsis pattern extracted from a typical female in the embodiment of the present invention, in which the horizontal axis is the frequency domain feature serial number, and the vertical axis is the frequency of the corresponding frequency domain feature.
Detailed Description
The present invention is further illustrated by the following examples to better understand the technical solutions of the present invention, but the present invention is not limited thereto. The ultrasonic equipment, coupling agent and the like used in the following examples are commercially available unless otherwise specified.
1. Ultrasound video acquisition
Screening 1 typical woman, after signing an informed consent, was included in this example. The common ultrasonic diagnostic apparatus and the vaginal probe are adopted to acquire the uterine ultrasonic video of a typical female. The acquisition mode is harmonic imaging ultrasound, the acquisition part is uterus, the acquisition frame rate is 30Hz, and the center frequency of the probe is 10 MHz.
2. Pretreatment of
And carrying out video denoising filtering and visual standardization preprocessing on the original ultrasonic video data. The preprocessing module carries out video filtering and denoising on an original ultrasonic video, Gaussian filtering is selected by a video filtering and denoising method, the Gaussian filtering is realized by discretizing window sliding window convolution, and the size of a sliding window is selected to be a matrix of 5 multiplied by 5. The preprocessing module carries out video filtering and denoising on an original ultrasonic video and then carries out visual standardization processing, and a histogram equalization is selected as a visual standardization processing method.
3. Uterine region detection
The uterine region is detected in the pre-processed ultrasound video and a video containing only the uterine region is obtained. The method comprises the steps of carrying out uterus region segmentation on each frame of a preprocessed ultrasonic video, taking a segmentation result of each frame as a uterus region of the frame based on a detection model obtained by machine learning method training, and extracting video data of only the uterus region as detected uterus video data. The uterus detection model obtained by training based on the machine learning method consists of a sliding window kernel group, a uterus classifier and a clustering filter; the sliding window kernel group consists of 8 sliding window kernels with different sizes, and a small image with different sizes is respectively cut from each frame of image in the video to be used as an interested area image until all image data are traversed; the uterus classifier is an artificial neural network and consists of 15-30 convolutional layers, each convolutional layer consists of 100-250000 neurons, the neurons on the bottom layer read a plurality of groups of region-of-interest images transmitted from the sliding window kernel group, and a plurality of groups of identification data output by the neurons on the top layer are sent to the cluster screener; the cluster filter is Mean-shift clustering, the Mean-shift search radius is 6 pixels, the multi-component data are clustered, the cluster center of each frame is used as the uterus center of the frame, and finally, the uterus video around the cluster center is extracted to be used as the detected uterus video.
4. Prominent micro-peristalsis of uterus
The uterus micro-peristalsis highlighting module consists of a scale transformation group, a filter, a phase amplifier and an inverse scale transformation group; the scale transformation group adopts wavelet transformation to convert the uterus video into videos with different scales; the signal filter receives video data of different scales, performs signal filtering on the video data of different scales based on a preset band-pass range of 0.1-14.9 Hz, and sends the video data to the phase amplifier; the phase amplifier receives the signal filtering result, performs 10 times multiplication amplification on the phase of the signal filtering result and then sends the signal to the inverse scale transformation group; and the inverse scale transformation group receives the phase amplification results of different scales, performs scale inverse reconstruction on the phase amplification results to obtain a uterine micro-peristalsis highlighted video and sends the uterine micro-peristalsis highlighted video to the display module.
FIG. 3 is a comparison of spatiotemporal slices before and after the uterine micro-peristalsis is highlighted in an example of the invention. FIG. 3.A is a raw uterine ultrasound video of a typical female, with dashed lines representing the longitudinal single column positions of the spatiotemporal slice acquisition; FIG. 3.B is the corresponding uterine micro-peristalsis highlighting video with dashed lines at the same longitudinal single column positions of the spatiotemporal slice acquisition as in FIG. 3. A; a longitudinal cut pixel in the video is spread on a time axis and is superposed into a picture, and a space-time slice picture can be obtained. The horizontal axis of the spatio-temporal slice is the 1 st, 2 nd, … … th and Kth frames of the video, and the vertical axis is the single-column pixels at the same position of the corresponding 1 st, 2 nd, … … th and Kth frames, wherein K is the total frame number of the video. FIG. 3.C is a spatiotemporal slice of a uterine ultrasound video of an original uterine ultrasound video; FIG. 3.D is a spatiotemporal slice of the corresponding uterine micro-peristalsis highlighted video.
5. Automatic extraction in micro-peristalsis mode
The automatic extraction module of the micro-creeping mode consists of a frequency domain characteristic encoder, a standard frequency domain characteristic dictionary and a frequency counter; uterine micro-peristalsis highlighting video data V4Each frame of image is converted into frequency domain data F by fourier transformi(i ═ 1,2, …,300),300 is the total number of frames in the video; obtaining 40 standard frequency domain characteristics from the frequency domain data F of all frames through a frequency domain characteristic encoder B and a characteristic degeneracy method J; the frequency counting device counts the frequency distribution of the 40 standard frequency domain characteristics according to the arrangement sequence of the standard frequency domain characteristic dictionary, and the frequency distribution is used as the micro-peristalsisThe mode Y is sent to the display module.
The frequency domain characteristic encoder receives the frequency domain data F and divides the frequency domain data F into a series of subspaces with the same size; the video picture size is 200 x 200, and the total frame number is 300; the subspace size is 20 × 20 × 30. The maximum real number and the maximum imaginary number of each subspace are the maximum real number and the maximum imaginary number of all points in the subspace; the minimum real number and the minimum imaginary number of each subspace are the minimum real number and the minimum imaginary number of all points in the subspace; the median real number and median imaginary number of each subspace are the median real number and median imaginary number of all points in the subspace; any subspace and its neighborhood subspace form a 3 × 3 × 3 three-dimensional matrix, and each position in the matrix contains 3 complex numbers, which are respectively the maximum real number and the maximum imaginary number, the minimum real number and the minimum imaginary number, and the median real number and the median imaginary number.
A frequency domain feature encoder sequentially reforms all frequency domain complex numbers of any subspace and a neighborhood subspace thereof into a one-dimensional vector of 1 × 162(162 is 3 × 3 × 3 × 3 × 2), and real numbers and corresponding imaginary numbers are arranged closely; for each value of the 1 × 162 one-dimensional vector, a value greater than 0 is marked as 1, and a value less than or equal to 0 is marked as 0, so as to obtain a 1 × 162 one-dimensional vector containing only 1 and 0, which is the binary frequency domain feature code.
This example shows the results shown in fig. 3 and 4 after the ultrasound video acquisition of the uterus of 1 typical female and the processing of the ultrasound video data according to the processing method of the present invention. FIG. 3 is a comparison of spatiotemporal slices before and after the highlighting of uterine micro-peristalsis in an embodiment of the invention, showing that the invention can effectively highlight uterine micro-peristalsis. Fig. 4 is a diagram of extracting a uterine micro-peristalsis pattern of a typical female in the embodiment of the invention, which shows that the invention can effectively extract the uterine micro-peristalsis pattern and provide quantitative characteristic reference for comparison of the micro-peristalsis pattern.
Claims (8)
1. A uterine ultrasound micro-peristalsis video processing method comprises the following modules:
an ultrasound acquisition module (1) for acquiring original ultrasound video data V1；
The preprocessing module (2) is used for carrying out video denoising filtering and visual standardization preprocessing on the original ultrasonic video data and carrying out preprocessing on the ultrasonic video data V after preprocessing2Sending the data to a uterus region detection module (3);
uterine region detection module (3) for V2Performing sliding window area identification on each frame of image, and integrating the areas judged as the uterus into uterus video data V in sequence3And will V3Sending the data to a uterus micro-peristalsis highlighting module (4);
a uterus micro-peristalsis highlighting module (4) for V3The uterus micro-peristalsis is performed to obtain V4；
A micro-peristalsis mode automatic extraction module (5) for extracting V4Micro-peristalsis mode Y;
a display module (6) responsible for displaying V4And a micro-peristalsis mode Y is presented to the user;
the method is characterized in that the uterine region in the ultrasonic video is subjected to micro-peristalsis highlighting through space multi-scale decomposition and phase amplification to obtain a uterine micro-peristalsis highlighting video V4(ii) a Further, V is counted based on a standard frequency domain feature dictionary D with 81 preset time-space dynamic changes4The frequency distribution which meets the preset standard frequency domain characteristics is taken as a micro-peristalsis mode Y of the uterus; finally V is set4And Y is output and displayed by the display module.
2. The uterine ultrasound micro-peristalsis video processing method according to claim 1, wherein the acquisition mode of the ultrasound acquisition module (1) is B-mode ultrasound or harmonic imaging ultrasound, the acquisition part is uterus, and the acquisition obtains raw ultrasound video data V containing the uterus1The sampling frequency of the video data is not less than 5 Hz.
3. The uterine ultrasound micro-peristalsis video processing method according to claim 1, characterized in that the preprocessing module (2) converts the raw ultrasound video data V1Carrying out video filtering denoising and visual standardization processing, and processing the processed ultrasonic video data V2Sending to uterine region detectionA module (3); the video filtering and denoising method can be Gaussian filtering G, and the visual standardization processing can be histogram equalization H.
4. The uterine ultrasound micro-peristalsis video processing method according to claim 1, characterized in that the uterine region detection module (3) is directed to V2The method for segmenting the uterine region of the uterine body comprises the steps of segmenting the uterine region of each frame by a recognition model obtained by manual image segmentation, Markov random field segmentation or training based on a machine learning method, taking the segmentation result of each frame as the uterine region of the frame, and extracting the video data of only the uterine region as the detected uterine video data V3。
5. The uterine ultrasound micro-peristalsis video processing method according to claim 1, wherein the uterine micro-peristalsis highlighting module (4) is composed of a scaling group, a filter, a phase amplifier and an inverse scaling group; scale transformation set transforms uterine video data V3Converting into video data with different scales, wherein the scale transformation method can be wavelet transformation, Gaussian pyramid decomposition and steerable direction pyramid transformation; the signal filter receives video data of different scales, performs signal filtering on the video data of different scales based on a preset band-pass range, and sends the video data of different scales to the phase amplifier; the phase amplifier receives the signal filtering result, multiplicatively amplifies the phase of the signal filtering result and sends the amplified signal to the inverse scale transformation group; the inverse scale transformation group receives the phase amplification results of different scales and carries out scale inverse reconstruction on the phase amplification results to obtain uterine micro-peristalsis highlighted video data V4Finally, the uterine micro-peristalsis is highlighted to display the video data V4Sending the data to a micro-peristalsis mode automatic extraction module (5).
6. The uterine ultrasound micro-peristalsis video processing method according to claim 1, wherein the micro-peristalsis mode automatic extraction module (5) is composed of a frequency domain feature encoder B, a standard frequency domain feature dictionary D and a frequency statistics device; uterine micro-peristalsis highlighting video data V4Each frame image of the image passes through the Fourier transformTransformation of vertical leaf into frequency domain data Fi(i ═ 1,2, …, K), K being the total frame number; obtaining Q standard frequency domain characteristics (0) from the frequency domain data F of all frames through a frequency domain characteristic encoder B and a characteristic degeneracy method J<Q<81 and Q is an integer); the frequency counting device counts the frequency distribution of the Q standard frequency domain characteristics according to the arrangement sequence of the standard frequency domain characteristic dictionary, and sends the frequency distribution to the display module (6) as a micro-creeping mode Y.
7. The uterine ultrasound micro-peristalsis video processing method according to claim 1, wherein the frequency domain feature encoder B of the micro-peristalsis mode automatic extraction module (5) divides the received frequency domain data F into a series of subspaces of the same size; the subspace has a size of axbxc, a ranges from 0.05M to 0.5M, b ranges from 0.05N to 0.5N, c ranges from 0.05K to 0.5K, and a, b and c are rounded positive integers; video V4The picture size of (1) is M multiplied by N, the total frame number is K, and M, N, K are positive integers; the maximum real number and the maximum imaginary number of each subspace are the maximum real number and the maximum imaginary number of all points in the subspace; the minimum real number and the minimum imaginary number of each subspace are the minimum real number and the minimum imaginary number of all points in the subspace; the median real number and median imaginary number of each subspace are the median real number and median imaginary number of all points in the subspace; any subspace and a neighborhood subspace form a 3 multiplied by 3 three-dimensional matrix, each position in the matrix comprises 3 complex numbers which are respectively a maximum real number and a maximum imaginary number, a minimum real number and a minimum imaginary number, and a median real number and a median imaginary number; the frequency domain feature encoder B sequentially reforms all frequency domain complex numbers of any subspace and a neighborhood subspace thereof into a one-dimensional vector of 1 × 162(162 is 3 × 3 × 3 × 3 × 2), and real numbers and corresponding imaginary numbers are arranged closely; for each value of the 1 × 162 one-dimensional vector, a value greater than 0 is marked as 1, and a value less than or equal to 0 is marked as 0, so as to obtain a 1 × 162 one-dimensional vector containing only 1 and 0, which is the binary frequency domain feature code.
8. The uterine ultrasound micro peristaltic video processing method according to claim 1, wherein the characteristic degeneracy method J of the micro peristaltic mode automatic extraction module (5) treats 162-bit binary frequency domain characteristic codes as end-to-end circular rings, traverses each binary number on the circular rings from the counterclockwise direction, and records the total number of times of jumping from 0 to 1 or 1 to 0 as Z; the characteristic degenerating method J degenerates binary frequency domain characteristic codes with the same Z value into the same standard frequency domain characteristic, wherein the value range of Z is 0-80; all 81 frequency domain features constitute a standard frequency domain feature 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 CN108537785A (en) | 2018-09-14 |
| CN108537785B true 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) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB2577718B (en) * | 2018-10-03 | 2022-08-24 | Cmr Surgical Ltd | Feature identification |
| 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 |
| 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 |
| 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 |
Citations (5)
| 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 |
| 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 |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| BR112014014308A2 (en) * | 2011-12-29 | 2017-06-13 | Laboratorios Andrómaco S A | vaginal ring that includes dhea or dhea sulfate and, optionally, an agent for modulating the release of the active ingredient, which can be used to increase ovarian reserve in women and to relieve symptoms associated with menopause. |
-
2018
- 2018-04-04 CN CN201810295212.5A patent/CN108537785B/en not_active Expired - Fee Related
Patent Citations (5)
| 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 |
| 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 |
|---|
| 电影MRI对子宫内膜蠕动的研究进展;刘珊珊;《中国医学影像技术》;20161231;632-634页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108537785A (en) | 2018-09-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108537785B (en) | Uterine ultrasound micro-peristalsis video processing method | |
| Loizou et al. | Comparative evaluation of despeckle filtering in ultrasound imaging of the carotid artery | |
| WO2021184805A1 (en) | Blood pressure prediction method and device using multiple data sources | |
| CN106709967B (en) | Endoscopic imaging algorithm and control system | |
| CN109671094A (en) | A kind of eye fundus image blood vessel segmentation method based on frequency domain classification | |
| CN106127769A (en) | A kind of brain Forecasting Methodology in age connecting network based on brain | |
| Cai et al. | Identifying architectural distortion in mammogram images via a se-densenet model and twice transfer learning | |
| CN111523617B (en) | Epilepsy detection system based on white matter fusion characteristic diagram and residual error attention network | |
| Gajera et al. | CT-scan denoising using a charbonnier loss generative adversarial network | |
| Mina et al. | Breast abnormality detection in mammograms using Artificial Neural Network | |
| CN110827283B (en) | Head and neck blood vessel segmentation method and device based on convolutional neural network | |
| Ali et al. | Detection and segmentation of hemorrhage stroke using textural analysis on brain CT images | |
| CN109965905B (en) | Contrast region detection imaging method based on deep learning | |
| Du et al. | Transformer and GAN-Based Super-Resolution Reconstruction Network for Medical Images | |
| Ziani | Contribution to Single-Channel Fetal Electrocardiogram Identification. | |
| CN111340760B (en) | Knee joint positioning method based on multitask two-stage convolution neural network | |
| Lou et al. | Segmentation of infrared breast images using multiresunet neural networks | |
| CN115100182A (en) | Method for typing vertebral column vertebral body fracture | |
| Selvathi et al. | Detection of retinal blood vessels using curvelet transform | |
| Sciortino et al. | A non-supervised approach to locate and to measure the nuchal translucency by means of wavelet analysis and neural networks | |
| Bozek et al. | Bilateral asymmetry detection in digital mammography using B-spline interpolation | |
| Sabeenian et al. | Texture-based medical image classification of computed tomography images using MRCSF | |
| Rizzi et al. | A fully automatic system for detection of breast microcalcification clusters | |
| Bhalla et al. | Automatic fetus head segmentation in ultrasound images by attention based encoder decoder network | |
| Ahmed et al. | Automated Glaucoma Detection based on LBP Representation and GLRLM Feature Extraction Method |
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 |