WO2023082304A1

WO2023082304A1 - Large-field-of-view adaptive wall filtering method and system based on hierarchical search

Info

Publication number: WO2023082304A1
Application number: PCT/CN2021/131481
Authority: WO
Inventors: 徐依雯; 崔崤峣; 焦阳; 李昕泽; 唐雨嘉
Original assignee: 苏州国科昂卓医疗科技有限公司
Priority date: 2021-11-11
Filing date: 2021-11-18
Publication date: 2023-05-19
Also published as: CN114170239A

Abstract

A large-field-of-view adaptive wall filtering method and system based on hierarchical search. The method comprises: dividing a whole image into a plurality of regions by using a large window; filtering, and evaluating the blood flow information content of each region; extracting regions with the blood flow information content higher than a threshold value, and performing local filtering on the regions by using refined small windows; processing each small window by using optimal specific filtering parameters according to the noise level of the corresponding region; and finally, performing image reconstruction, and outputting a complete ultrasonic blood flow image subjected to wall filtering. The large-field-of-view adaptive wall filtering method based on hierarchical search is an efficient wall filtering method suitable for large capacity data, and the method improves the operation efficiency, and can also suppress clutters well, thereby significantly improving the filtering quality.

Description

Large Field of View Adaptive Wall Filtering Method and System Based on Hierarchical Search

technical field

The invention relates to the field of medical image processing, in particular to a hierarchical search-based large field of view adaptive wall filtering method and system.

Background technique

Ultrafast ultrasonic plane wave imaging technology can provide large field of view synchronous echo data with high frame rate, and provide more effective original information for large field of view blood flow images. Common filtering algorithms have many shortcomings in processing large field of view data.

The existing wall filtering methods can be roughly divided into three types: (1) The traditional time-domain filter can only process single-point time-domain information for each filter, and two-dimensional images are pieced together by point-by-point filtering, which has low efficiency in large-field image processing. Problems such as poor anti-interference ability, slow speed, and poor filtering effect. (2) The new Singular Value Decomposition filter (SVD filter, Singular Value Decompostion) can input all the original data at one time, which improves the processing efficiency, but adopts a "one size fits all" filter cut-off setting for the entire image, facing a large field of view The effective information in the non-uniform distribution in the image has local over- or under-processing. (3) The improved sub-area SVD filter divides the whole picture into many small areas for specific processing, which improves the filtering quality, but sacrifices the advantage of high efficiency of the SVD filter.

Therefore, a more reliable solution is now needed.

Contents of the invention

The technical problem to be solved by the present invention is to provide a large field of view adaptive wall filtering method and system based on hierarchical search in view of the above-mentioned deficiencies in the prior art. The invention can combine the advantages of high efficiency of global SVD and high filtering quality of subregional SVD, and can quickly acquire high-quality blood flow images for the entire observation region.

In order to achieve the above object, the technical solution adopted by the present invention is: a large field of view adaptive wall filtering method based on hierarchical search, which is used to perform wall filtering processing on ultrasonic images. ) Divide the entire image into several regions, perform filtering and evaluate the blood flow information content of each region, extract the region with blood flow information content higher than the threshold, and then perform local filtering processing with a thinned small window (secondary window) , each small window is processed with the optimized specific filter parameters according to the noise level of the corresponding area, and finally the image is reconstructed to output a complete ultrasonic blood flow image after wall filtering.

Preferably, the method comprises the steps of:

1) Obtain multi-frame time-continuous two-dimensional plane wave ultrasonic echo signals, and obtain large-field continuous frame two-dimensional plane wave demodulation data through IQ demodulation;

2) First-level window evaluation: divide the whole image into several regions with large windows, evaluate the blood flow information content of each region simultaneously, and extract the regions whose blood flow information content is higher than the threshold;

3) Second-level window refinement: calculate and set a suitable small window, and divide the area whose blood flow information content is higher than the threshold value extracted in step 2) into small windows that overlap each other between adjacent areas;

4) Each small window independently performs adaptive wall filtering;

5) Image reconstruction: superimpose the output results of independent adaptive filtering of each small window on the corresponding position of the whole image, obtain a normalized image through statistics and weighted average, and output the reconstructed complete blood flow image.

Preferably, the steps of performing adaptive wall filtering for a single small window are: performing time-space domain reconstruction of small window data to obtain a decomposition matrix, extracting eigenvalues through singular value decomposition, drawing a characteristic curve, and using the characteristic curve to calculate the optimal filtering Parameters, use the optimized filtering parameters to filter the small window data, reconstruct the image and output the result.

The present invention also provides a large field of view adaptive wall filtering system based on hierarchical search, which includes:

IQ demodulation module, it carries out IQ demodulation to the acquired two-dimensional plane wave ultrasonic echo signal of multi-frame time continuous, obtains two-dimensional plane wave demodulation data of continuous frame of large field of view;

The first-level window evaluation module, which divides the whole image into several regions with large windows, simultaneously evaluates the blood flow information content of each region, and extracts the regions whose blood flow information content is higher than the threshold;

The second-level window refinement module, which calculates and sets a suitable small window, and divides the area whose blood flow information content is higher than the threshold value extracted by the first-level window evaluation module into small windows that overlap each other between adjacent areas;

An adaptive wall filtering processing module, which independently performs adaptive wall filtering processing on each small window;

And the image reconstruction module, which superimposes the output results of each small window independent adaptive filtering obtained by the adaptive wall filtering processing module to the corresponding position of the whole image, obtains a normalized image through statistics and weighted average, and outputs the complete reconstruction Blood flow image.

The present invention also provides a storage medium, on which a computer program is stored, and the program is used to implement the above method when executed.

The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and operable on the processor, and the processor implements the above-mentioned method when executing the computer program.

The beneficial effects of the present invention are: the hierarchical search-based large field of view adaptive wall filtering method provided by the present invention can be applied to the high-efficiency wall filtering method of large-capacity data. The calculation power is allocated to the parts, which improves the calculation efficiency; on the other hand, although the small filtering windows in the present invention overlap, there is no iteration between the filtering processes, so multiple independent small windows can be calculated synchronously through multiple threads, which greatly saves calculation time;

The wall filtering method of the present invention can suppress clutter very well. On the one hand, the difficulty of filtering is reduced by subdividing the filtering area. The reduction of the single filtering area improves the similarity of spatial domain information, and the effective information is more concentrated, making it easier to filter the clutter. The signal is eliminated from it; on the other hand, the optimal filtering parameters are designed according to the data characteristics of each independent area, which greatly improves the filtering quality.

Description of drawings

FIG. 1 is a flow chart of the hierarchical search-based large field of view adaptive wall filtering method of the present invention.

Detailed ways

The present invention will be further described in detail below in conjunction with the embodiments, so that those skilled in the art can implement it with reference to the description.

It should be understood that terms such as "having", "comprising" and "including" used herein do not exclude the presence or addition of one or more other elements or combinations thereof.

Example 1

A large field of view adaptive wall filtering method based on hierarchical search in this embodiment, the method is used to perform wall filtering processing on ultrasonic images, the method first uses a large window (first-level window) to divide the entire image into several regions , perform filtering and evaluate the blood flow information content of each area, extract the area (rich effective information, complex image) with blood flow information content higher than the threshold, and then perform local filtering with a thinned small window (secondary window), Each small window is processed with optimized specific filter parameters according to the noise level of the corresponding area, and finally the image is reconstructed to output a complete ultrasonic blood flow image after wall filtering.

With reference to Fig. 1, the large-field-of-view adaptive wall filtering method based on hierarchical search in the present embodiment includes the following steps:

1) Acquire multi-frame time-continuous two-dimensional plane wave ultrasonic echo signals through array ultrasonic probes (including convex array and linear array), and obtain large-field continuous frame two-dimensional plane wave demodulation data through IQ demodulation;

4) Each small window independently performs adaptive wall filtering processing:

Reconstruct the small window data in time and space domain to obtain the decomposition matrix, extract the characteristic value through singular value decomposition, draw the characteristic curve, use the characteristic curve to calculate the optimized filtering parameters, and use the optimized filtering parameters to filter the small window data. Constructs an image and outputs the result. The data of each small window is filtered independently. Since each window does not interfere with each other, the process can be carried out synchronously to improve the processing speed.

The invention is suitable for processing large-capacity data brought by imaging with a large field of view, and can quickly provide high-quality blood flow images. The development of plane wave technology and the upgrading of hardware equipment have made the acquisition of two-dimensional synchronous ultrasonic signals with high frame rate and large field of view a reality. A large amount of time information, to deal with such a large amount of data, requires a wall filtering method that takes both speed and filtering quality into consideration. The algorithm provided by the present invention uses sub-area independent self-adaptive filtering to ensure the filtering quality, saves computing time through synchronous operation because there is no iterative process in each area, and can quickly process large field of view data to provide high-quality blood flow images.

The inhomogeneous distribution of imaging targets in the large field of view image leads to the core area with high effective information content (in the blood flow map, it is represented by a rich blood vessel area), and the edge area with almost no effective information (in the blood flow map). It is shown as no blood flow area), and the indiscriminate small window search algorithm will waste a lot of computing power in the edge area. The present invention adopts a hierarchical search strategy, determines the core area by rough search, and then extracts effective information more accurately by precise search, which can greatly improve the operation efficiency under the premise of ensuring the filtering quality.

In the present invention, the data of each small window is independently processed, the energy distribution is analyzed according to the characteristic curve, and the optimal filtering parameters are calculated to ensure that each small window is filtered with the optimal parameters, thereby improving the filtering quality.

Example 2

This embodiment provides a hierarchical search-based large-field-of-view adaptive wall filtering system, which utilizes the method in Embodiment 1 to perform wall filtering processing on ultrasonic images, and the system includes:

This embodiment also provides a storage medium on which a computer program is stored, and the program is used to implement the method in Embodiment 1 when executed.

This embodiment also provides a computer device, including a memory, a processor, and a computer program stored on the memory and operable on the processor, and implementing the method in Embodiment 1 when the processor executes the computer program.

Although the embodiment of the present invention has been disclosed as above, it is not limited to the use listed in the specification and implementation, it can be applied to various fields suitable for the present invention, and it can be easily understood by those skilled in the art Therefore, the invention is not limited to the specific details without departing from the general concept defined by the claims and their equivalents.

Claims

A large field of view adaptive wall filtering method based on hierarchical search, characterized in that the method is used for wall filtering processing of ultrasonic images, the method first uses a large window to divide the entire image into several regions, performs filtering and evaluation For the blood flow information content of each area, the area with blood flow information content higher than the threshold is extracted and then local filtering is performed with a thinned small window. Each small window adopts the optimized specific filtering parameters according to the noise level of the corresponding area processing, and finally image reconstruction to output a complete ultrasonic blood flow image after wall filtering.
The large field of view adaptive wall filtering method based on hierarchical search according to claim 1, wherein the method comprises the following steps:

1) Obtain multi-frame time-continuous two-dimensional plane wave ultrasonic echo signals, and obtain large-field continuous frame two-dimensional plane wave demodulation data through IQ demodulation;

2) First-level window evaluation: divide the whole image into several regions with large windows, evaluate the blood flow information content of each region simultaneously, and extract the regions whose blood flow information content is higher than the threshold;

3) Second-level window refinement: calculate and set a suitable small window, and divide the area whose blood flow information content is higher than the threshold value extracted in step 2) into small windows that overlap each other between adjacent areas;

4) Each small window independently performs adaptive wall filtering;

5) Image reconstruction: superimpose the output results of independent adaptive filtering of each small window on the corresponding position of the whole image, obtain a normalized image through statistics and weighted average, and output the reconstructed complete blood flow image.
According to claim 2, the adaptive wall filtering method for large field of view based on hierarchical search is characterized in that the step of performing adaptive wall filtering on a single small window is as follows: performing time-space domain reconstruction of small window data to obtain a decomposition matrix, Extract the eigenvalues by singular value decomposition, draw the characteristic curve, use the characteristic curve to calculate the optimized filter parameters, use the optimized filter parameters to filter the small window data, reconstruct the image and output the result.
A large field of view adaptive wall filtering system based on hierarchical search, characterized in that it includes:

IQ demodulation module, it carries out IQ demodulation to the acquired two-dimensional plane wave ultrasonic echo signal of multi-frame time continuous, obtains two-dimensional plane wave demodulation data of continuous frame of large field of view;

The first-level window evaluation module, which divides the whole image into several regions with large windows, simultaneously evaluates the blood flow information content of each region, and extracts the regions whose blood flow information content is higher than the threshold;

The second-level window refinement module, which calculates and sets a suitable small window, and divides the area whose blood flow information content is higher than the threshold value extracted by the first-level window evaluation module into small windows that overlap each other between adjacent areas;

An adaptive wall filtering processing module, which independently performs adaptive wall filtering processing on each small window;

And the image reconstruction module, which superimposes the output results of each small window independent adaptive filtering obtained by the adaptive wall filtering processing module to the corresponding position of the whole image, obtains a normalized image through statistics and weighted average, and outputs the complete reconstruction Blood flow image.
A storage medium on which a computer program is stored, wherein the program is used to implement the method according to any one of claims 1-3 when executed.
A computer device, comprising a memory, a processor, and a computer program stored on the memory and operable on the processor, characterized in that, when the processor executes the computer program, the computer program according to claims 1-3 is implemented. any one of the methods described.