Ultrasonic doppler blood flow imaging method
By performing eigenvalue decomposition and signal-to-noise ratio calculation on ultrasound images, combined with template processing and connected domain area calculation, the problems of incoherence and instability of blood flow signals in traditional ultrasonic color Doppler technology are solved, and the non-invasive high quality is achieved Blood flow imaging.
Patent Information
- Application Number
- PCT/CN2023/094662
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-16
- Filing Date
- 2023-05-17
- Publication Date
- 2024-11-21
AI Technical Summary
Traditional ultrasonic color Doppler technology has limitations when observing small blood vessels and slow blood flow, especially in parenchymal organs, where information of tissue and blood flow superimposed, resulting in incoherent and unstable blood flow signals, conventional wall filters It is impossible to separate effectively, and contrast agent is required to inject into the blood vessels to enhance detection accuracy.
By decomposing the eigenvalue of ultrasound images, it is divided into high-energy, medium-energy and low-energy characteristic layers. Combined with signal-to-noise ratio calculation, high-confidence blood flow image data is extracted, and the 'X' type template opening and closing function is processed Breakpoints and remove background noise, and finally remove isolated points through the calculation of the area of the connection domain to obtain a coherent and clear blood flow image.
It achieves a non-invasive acquisition of accurate, complete, clear and stable blood flow images, which can better extract tiny blood flow signals and avoids the trauma of injecting contrast agent into the human body.
Smart Images

Figure CN2023094662_21112024_PF_FP_ABST
Abstract
Description
A method for ultrasonic Doppler blood flow imaging Technical Field
[0001] The present invention relates to the technical field of ultrasonic image processing, and in particular to an ultrasonic Doppler blood flow imaging method. Background Art
[0002] Traditional ultrasound color Doppler technology has limitations in observing small blood vessels and slow blood flow, especially in solid organs where tissue and blood flow information are superimposed. In conventional blood flow imaging processing, a wall filter is used to filter out tissue signals. The wall filter is generally a high-pass filter. When the blood flow velocity is low, the frequencies of tissue and blood flow are close, and conventional wall filters cannot completely separate the blood flow signal. The resulting blood flow signal is incoherent and unstable. To solve this problem, contrast agents are usually injected into the blood vessels to enhance reflections to improve the accuracy of blood flow detection, but this is an invasive method. Summary of the Invention
[0003] The object of the present invention is to provide an ultrasonic Doppler blood flow imaging method, specifically to provide a non-invasive ultrasonic Doppler imaging method that can obtain accurate, complete, clear and stable blood flow images.
[0004] To achieve the above object, the present invention adopts the following technical solution: an ultrasonic Doppler blood flow imaging method, comprising the following steps:
[0005] S01. Perform eigenvalue decomposition on the ultrasound image obtained by ultrasound scanning to obtain N feature layers, take the envelope of the data of the N feature layers to obtain the energy of each feature layer, and divide each feature layer into three feature layers according to the energy of each feature layer: high energy, medium energy and low energy; among which, the high energy feature layer is selected from the N feature layers, and the N / 5 feature layers with the highest energy values are used as the high energy feature layer; the low energy feature layer is selected from the N feature layers, and the N / 10 feature layers with the lowest energy values are used as the low energy feature layer, and the rest are medium energy feature layers.
[0006] S02. Based on the data of the N characteristic layers obtained in step S01, determine the distribution of each point of the ultrasound image in each characteristic layer, and retain the points mainly distributed in the medium energy characteristic layer.
[0007] S03. The low-energy feature layer in step S01 is used as the noise signal, and the signal-to-noise ratio of each point in the ultrasound image is calculated. Points with a signal-to-noise ratio lower than a set value are considered to be low-credibility points, and points with a signal-to-noise ratio higher than a set value are considered to be high-credibility points. The low-credibility points are eliminated, and the high-credibility points are retained.
[0008] S04. Combining the results of step S02 and step S03, extracting points that meet both the requirements of being mainly distributed in the medium energy characteristic layer in step S02 and the requirements of high credibility in step S03, thereby obtaining high-credibility blood flow image data.
[0009] Specifically, the method further includes step S05, wherein the high-reliability blood flow image data is processed by using an "X"-shaped template opening and closing function to connect breakpoints and suppress background noise, thereby obtaining coherent and clear blood flow image data.
[0010] Specifically, the method further includes step S06, calculating the connected domain area of the coherent and clear blood flow image data obtained in step S05, and considering the points whose connected domain area is smaller than the set value as isolated points. After removing the isolated points, coherent, clear and noise-free blood flow image data is obtained.
[0011] The beneficial effects of the present invention are as follows: by performing eigenvalue decomposition on ultrasound images, tissue signals and noise signals can be filtered out in the feature dimension so that blood flow data can be better extracted, thereby achieving complete retention of blood flow data, and stable and clear imaging of tiny blood flows without the need to inject contrast agents to cause trauma to the human body. BRIEF DESCRIPTION OF THE DRAWINGS
[0012] FIG1 is a flow chart of the ultrasonic Doppler blood flow imaging method in an embodiment. Implementation Method
[0013] Example 1, referring to FIG1 , an ultrasonic Doppler blood flow imaging method comprises the following steps:
[0014] S01. Perform eigenvalue decomposition on the ultrasound image obtained by ultrasound scanning to obtain N feature layers. Take the envelope of the data of the N feature layers to obtain the energy of each feature layer. According to the energy of each feature layer, each feature layer is divided into three feature layers: high energy, medium energy and low energy. Among them, the high energy feature layer is the N / 5 feature layers with the highest energy value selected from the N feature layers as the high energy feature layer, the low energy feature layer is the N / 10 feature layers with the lowest energy value selected from the N feature layers as the low energy feature layer, and the rest are medium energy feature layers. Among the three feature layers of high energy, medium energy and low energy, the high energy feature layer represents the tissue signal, the medium energy feature layer represents the blood flow signal, and the low energy feature layer represents the noise signal. It should be noted that in the above selection of the high energy feature layer and the low energy feature layer, N / 5 and N / 10 should both be integers. If they are not integers, they are calculated by rounding down.
[0015] S02. Based on the data of the N characteristic layers obtained in step S01, determine the distribution of each point of the ultrasound image in each characteristic layer, and retain the points mainly distributed in the medium energy characteristic layer.
[0016] S03. The low-energy feature layer in step S01 is used as the noise signal, and the signal-to-noise ratio of each point in the ultrasound image is calculated. Points with a signal-to-noise ratio lower than a set value are considered to be low-credibility points, and points with a signal-to-noise ratio higher than a set value are considered to be high-credibility points. The low-credibility points are eliminated, and the high-credibility points are retained.
[0017] S04. Combining the results of step S02 and step S03, extracting points that meet both the requirements of being mainly distributed in the medium energy characteristic layer in step S02 and the requirements of high credibility in step S03, thereby obtaining high-credibility blood flow image data.
[0018] In addition, due to the low contrast of tiny blood flows, the method further includes step S05, which uses an "X"-shaped template opening and closing function to connect the breakpoints of the high-reliability blood flow image data and suppress background noise to obtain coherent and clear blood flow image data.
[0019] At the same time, in order to remove isolated noise points in the image, step S06 is also included, in which the connected domain area of the coherent and clear blood flow image data obtained in step S05 is calculated. Points with a connected domain area smaller than a set value are considered to be isolated points. After removing the isolated points, coherent, clear and noise-free blood flow image data is obtained.
[0020] Of course, the above are only preferred embodiments of the present invention and are not intended to limit the scope of use of the present invention. Therefore, any equivalent changes based on the principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. An ultrasonic Doppler blood flow imaging method, characterized in that: The steps include: S01. Perform eigenvalue decomposition on the ultrasound image obtained by ultrasound scanning to obtain N feature layers, take the envelope of the data of the N feature layers to obtain the energy of each feature layer, and divide each feature layer into three feature layers of high energy, medium energy and low energy according to the energy of each feature layer; wherein, the high energy feature layer is selected from the N feature layers, N / 5 feature layers with the highest energy value are used as the high energy feature layer, the low energy feature layer is selected from the N feature layers, N / 10 feature layers with the lowest energy value are used as the low energy feature layer, and the rest are medium energy feature layers; S02, judging the distribution of each point of the ultrasound image in each characteristic layer according to the data of the N characteristic layers obtained in step S01, and retaining the points mainly distributed in the medium energy characteristic layer; S03, taking the low energy characteristic layer in step S01 as the noise signal, and calculating the signal-to-noise ratio of each point in the ultrasound image, considering the points with a signal-to-noise ratio lower than the set value as low-credibility points, and considering the points with a signal-to-noise ratio higher than the set value as high-credibility points, removing the low-credibility points and retaining the high-credibility points; S04. Combining the results of step S02 and step S03, extracting points that satisfy both the points mainly distributed in the medium energy characteristic layer in step S02 and the points with high credibility in step S03, thereby obtaining blood flow image data with high credibility.
2. The method for ultrasonic Doppler blood flow imaging according to claim 1, characterized in that: The method further includes step S05, wherein the high-reliability blood flow image data is processed by using an "X"-shaped template opening and closing function to connect breakpoints and suppress background noise, so as to obtain coherent and clear blood flow image data.
3. The method for ultrasonic Doppler blood flow imaging according to claim 2, characterized in that: The method further includes step S06, calculating the connected domain area of the coherent and clear blood flow image data obtained in step S05, and considering points whose connected domain area is smaller than a set value as isolated points. After removing the isolated points, coherent, clear and noise-free blood flow image data is obtained.
Citation Information
Patent Citations
Ultrasonic diagnostic apparatus
CN103445809A
Method and device for calculating maximum frequency enveloping curve
CN103494615A
Ultrasonic color Doppler image post-processing method
CN105590315A
Ultrasonic Doppler blood flow imaging method
CN116269490A
Doppler energy-related parameters in an ultrasound imaging system
US5871447A
