Ultrasonic imaging device, signal processing method, and signal processing program

Abstract

The present invention provides an ultrasonic imaging device, a signal processing method and a signal processing program that can extract the blood flow of fine blood vessels in a short time with high precision. A plurality of vibrators are used to receive ultrasound waves reflected by a subject, thereby generating N frames of data. A correlation matrix is ​​generated by arranging the data of the corresponding positions of the frame data by a vector of N frames, and the specific values ​​and specific vectors of each level of N levels are calculated. The specific vectors above the threshold level k are multiplied by the multiple frame data respectively, and the first filter element is calculated based on the variance between the data at the position zx corresponding to the obtained multiple blood flow component frame data. The specific vector of level 1 is multiplied by the frame data to obtain the tissue component frame data, and the second filter element is calculated based on this. The frame data is weighted by the first and / or second filter elements to generate a clutter suppressed image.

Description

Technical Field

[0001] The present invention relates to an ultrasonic imaging device, and more particularly to a technique for suppressing clutter components from tissues in a technique for acquiring a blood flow image in a living body using singular value decomposition. Background Art

[0002] Ultrasonic imaging devices are widely used as medical inspection devices that present information in the body that cannot be seen visually in the form of numerical values ​​or images. As a basic imaging method, an ultrasonic signal (echo signal) is sent to the imaging object in the body, and the amplitude information of the received signal obtained by reflection is used to display the morphology of the tissue. On the other hand, the technology of imaging the position and speed of blood flow by measuring the phase change based on the Doppler effect is also widely known.

[0003] In recent years, as in Patent Documents 1 and 2, a blood flow imaging technology based on principal component analysis has been developed, and the technology of displaying a fine vascular network as image information has attracted attention. Principal component analysis is a statistical analysis method based on analytical methods such as special value decomposition and intrinsic value decomposition. When principal component analysis is applied to ultrasonic signals, information that better reproduces the original image (for example, components with relatively high brightness such as tissue boundaries and substance) is classified as a higher principal component, and conversely, information with a lower information dominance (for example, dynamic blood flow components with lower reflectivity than tissue) is classified as a lower principal component. A scheme has been proposed to utilize this characteristic to specifically extract blood flow components from ultrasonic signals and perform imaging.

[0004] The technology disclosed in Patent Document 1 is a method for performing principal component analysis on the acquired ultrasonic signal (Doppler signal) to suppress the component (clutter component) of tissue movement caused by changes in body position, breathing, heart pulsation, etc., and improve the quality of the blood flow image. For example, the main filter matrix is ​​calculated so that the first to third principal components corresponding to the clutter components are removed, and the fourth to sixth principal components corresponding to the blood flow components are maintained. By applying the main filter matrix to the Doppler signal, an image with suppressed clutter components is generated. Furthermore, the first and sixth principal components are extracted from the Doppler signal, and the ratio is used as an index to weight the power signal of the blood flow. In this way, the residual clutter components are appropriately suppressed.

[0005] Furthermore, the technology disclosed in Patent Document 2 is a method of suppressing clutter components by adjusting the signal strength of a display image based on information obtained by principal component analysis of an ultrasonic signal.

[0006] Prior Art Literature

[0007] Patent Literature

[0008] Patent Document 1: Japanese Patent Application Publication No. 2019-54938

[0009] Patent Document 2: Japanese Patent Application Publication No. 2020-185122 Summary of the invention

[0010] -Problems to be solved by the invention-

[0011] In recent years, there has been a desire for an ultrasonic imaging device that can extract the blood flow of tiny blood vessels with high precision. Therefore, the inventors have focused on the need to suppress the noise of the background image and the pixels of the blood vessel wall that are prone to high brightness to a level higher than the existing technology. Principal component analysis based on special value decomposition and intrinsic value decomposition is an effective method for extracting tiny blood vessel structures, but depending on the speed of movement of tissues and blood flow or the size of the brightness, the suppression of noise components may be insufficient. In this case, the noise components remain in the displayed image, and the correct blood vessel structure cannot be visually identified, resulting in a decrease in diagnostic accuracy and an increase in the pressure of reading the image.

[0012] The technology described in Patent Document 1 extracts a first specific principal component (e.g., the first principal component) and a second specific principal component (e.g., the sixth principal component) from a Doppler signal by principal component analysis, and adjusts (weights) the brightness using the ratio as an index. For example, when the ratio of the sixth principal component to the first principal component is used as an index, if the pixel is located in a blood vessel, the index is close to 1, and if the pixel is located in an organ tissue, the index is close to 0.

[0013] However, for example, when organ tissue contains high-brightness components, organ tissue moves violently, or tissue and blood flow information is dispersed over a wide range of main components, even for pixels located in organ tissue, the above index is not 0 but a relatively high value. As a result, it may not be possible to completely suppress residual clutter.

[0014] On the other hand, the technology described in Patent Document 2 performs principal component analysis on the acquired time series ultrasonic signal, calculates the brightness adjustment amount based on the result, and adjusts the brightness of the displayed image. That is, it is a method of calculating the residual clutter component based on the data of the acquired ultrasonic signal. However, since the ultrasonic signal also contains the blood flow component, there is not only residual clutter, but also the possibility of excessively suppressing the blood flow component, which may result in the impairment of blood vessel visual recognition.

[0015] Generally, when clutter suppression is performed by principal component analysis, a plurality of transmissions and receptions are performed to obtain a plurality of ultrasonic signals (Doppler signals) and perform principal component analysis, thereby improving the clutter suppression effect. However, transmission and reception takes time, and real-time performance is reduced.

[0016] An object of the present invention is to provide an ultrasonic imaging device capable of extracting blood flow in a microscopic blood vessel with high accuracy in a short time.

[0017] -Methods for solving problems-

[0018] According to the present invention, there is provided the following ultrasonic imaging device.

[0019] That is, the ultrasonic imaging device of the present invention has a data collection unit, a matrix conversion unit, a matrix analysis unit, a filter element generation unit and an image processing unit. The data collection unit receives a plurality of reception signals obtained by receiving the ultrasound waves reflected in sequence in the depth z direction of the subject to which the ultrasound waves are transmitted by using an array of a plurality of vibrators arranged in the x direction. The data collection unit repeatedly arranges the received reception signals on the zx plane to generate frame data, and generates N frames of frame data. The matrix conversion unit generates a correlation matrix by arranging the data of the corresponding position zx of the frame data into vectors of N frames. The matrix analysis unit performs singular value decomposition on the correlation matrix, and calculates the singular value and singular vector of each level of the N levels of the received signal. The filter element generation unit generates at least one of the first filter element and the second filter element as a filter element. The image processing unit weights the pixel values ​​of the frame data or the image generated based on the frame data according to the filter elements generated by the filter element generation unit to generate a clutter suppression image. The first filter element is calculated based on the variance between the data of the position zx corresponding to the plurality of blood flow component frame data obtained by multiplying the plurality of specific vectors of a given level range above a predetermined threshold level k by the plurality of frame data. The second filter element is calculated based on the tissue component frame data obtained by multiplying the specific vector of a given level less than the threshold level k by the frame data of 1 or more.

[0020] -Effects of the Invention-

[0021] According to the present invention, clutter components from tissues can be effectively suppressed in a short time by principal component analysis, and blood flow in microscopic blood vessels can be extracted with high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a block diagram of the ultrasonic imaging device according to the first embodiment.

[0023] Figure 2 1 is a flowchart showing the processing procedure of the blood flow image generating unit 30 of the ultrasonic imaging device according to the first embodiment.

[0024] Figure 3 Use graphs and charts to Figure 2 A flowchart for explaining each process.

[0025] Figure 4 This is an explanatory diagram showing an example of a display screen of the ultrasonic imaging device according to the first embodiment.

[0026] Figure 5 This is a graph showing the relationship between the correlation value (similarity) of a plurality of frame data calculated by the ultrasonic imaging device according to the second embodiment and the coefficient α of the second filter element.

[0027] Figure 6 1 is a flowchart showing a processing procedure of the blood flow image generating unit 30 of the ultrasonic imaging device according to the second embodiment.

[0028] Figure 7 (a) and (b) are explanatory diagrams showing examples of display screens of the ultrasonic imaging device according to the second embodiment.

[0029] Figure 8 This is a flowchart showing another example of the processing procedure of the blood flow image generating unit 30 of the ultrasonic imaging device according to the second embodiment.

[0030] Fig. 9 1 is a flowchart showing a processing procedure of the blood flow image generating unit 30 of the ultrasonic imaging device according to the third embodiment.

[0031] Fig.10 This is a graph showing the relationship between the correlation value (similarity) of a plurality of frame data calculated by the ultrasonic imaging device according to the third embodiment and the number of acquired frame data.

[0032] -Explanation of symbols-

[0033] 1 ... subject, 10 ... probe, 20 ... transmission and reception control unit, 21 ... transmission beamformer, 22 ... reception beamformer, 30 ... blood flow image generation unit, 31 ... data collection unit, 32 ... matrix conversion unit, 33 ... matrix analysis unit, 34 ... singular value analysis unit, 35 ... filter element generation unit, 36 ... mask image generation unit, 37 ... image reconstruction unit, 38 ... image processing unit, 40 ... memory, 50 ... B-mode image generation unit, 60 ... display processing unit, 70 ... display device, 80 ... external input device, 100 ... ultrasonic imaging device, 300 ... three-dimensional data, 301 ... frame data, 310 ... graph, 403 ... area, bar ... 404, 405 ... mark, 501 ... straight line graph, 502, 503 ... curve graph, 901 ... straight line graph, 902, 903 ... curve graph, 1001 ... graph. DETAILED DESCRIPTION

[0034] An ultrasonic imaging device according to an embodiment of the present invention will be described with reference to the drawings.

[0035] 《<Implementation Method 1>》

[0036] use Figures 1 to 3 The ultrasonic imaging device 100 according to the first embodiment will be described. Figure 1 is a diagram showing a schematic configuration of an ultrasonic imaging device according to Embodiment 1. Figure 2 This is a flowchart showing the processing procedure of the blood flow image generating unit of the ultrasonic imaging device. Figure 3 Use graphs and charts to Figure 2 A flowchart for explaining each processing step.

[0037] use Figure 1 , the structure of the ultrasonic imaging device 100 of Embodiment 1 is described. The ultrasonic imaging device 100 is connected to the probe 10, the external input device 80 for receiving instructions from the outside, and the display device 70. The probe 10 has an array of vibrators arranged in the x direction. The vibrators are composed of piezoelectric elements and the like.

[0038] The ultrasonic imaging device 100 includes a transmission / reception control unit 20 , a blood flow image generating unit 30 , a memory 40 , a B-mode image generating unit 50 , and a display processing unit 60 .

[0039] The transmission and reception control unit 20 includes a transmission beamformer 21, a reception beamformer 22, and a quadrature detector (not shown). The transmission beamformer 21 generates a transmission signal and outputs it to each vibrator of the probe 10, and transmits it in the depth direction (z direction) of the subject 1. The reception beamformer 22 receives a reception signal output by each vibrator of the probe 10 receiving an ultrasonic wave reflected by the subject 1, etc., and performs a predetermined phase modulation process to generate reception beam data. The quadrature detector performs quadrature detection on the reception beam data to generate beam data as a complex signal.

[0040] The blood flow image generator 30 includes a data collector 31, a matrix converter 32, a matrix analyzer 33, a singular value analyzer 34, a filter element generator 35, a mask image generator 36, and an image reconstructor 37. The mask image generator 36 and the image reconstructor 37 constitute an image processor 38.

[0041] The blood flow image generating unit 30 includes a computer having a processor such as a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit) and a memory. The processor reads and executes a program pre-stored in the memory, thereby realizing the functions of the above-mentioned parts 31 to 38 by software. In addition, the blood flow image generating unit 30 can also be partially or entirely composed of hardware. For example, a custom IC such as an ASIC (Application Specific Integrated Circuit) or a programmable IC such as an FPGA (Field-Programmable Gate Array) can be used to design a circuit so as to realize the functions of each part 31 to 38.

[0042] Hereinafter, the operation of each part of the ultrasonic imaging device 100 according to the present embodiment will be described.

[0043] The transmission beam former 21 outputs the transmission signals with a given time delay to the vibrators of the probe 10. As a result, the vibrators of the probe 10 irradiate ultrasonic waves toward the organ tissues in the living body (subject 1) that is the imaging object through electroacoustic conversion based on the piezoelectric effect. The ultrasonic waves propagate in the living body in a manner of connecting the focus at a given depth. During the propagation process, the ultrasonic waves are repeatedly reflected at interfaces with different acoustic impedances.

[0044] The reflected wave (ultrasound) reaches the vibrator of the probe 10 again and is received and converted into a received signal (electrical signal). The receiving beam former 22 performs phase modulation processing on the received signal and generates a plurality of receiving beam data in the depth z direction in the x direction so as to return the time delay set when sending. The received signal after the phase modulation processing is sent to the blood flow image generation unit 30. Among them, as long as the received signal processed in the present invention is always obtained by reflection from the imaging object, there is no need to specifically limit the transmission and reception method. For example, even if the vibrator of the probe 10 is not given a time delay, the signal obtained by sending a plane wave to the imaging object is also valid.

[0045] When generating a blood flow image, the blood flow image generating unit 30 repeatedly transmits and receives ultrasonic waves N times with respect to the same imaging plane of the subject 1 .

[0046] Next, use Figure 2 as well as Figure 3 The flowchart of FIG. 1 will explain the processing of each unit 31 to 38 of the blood flow image generating unit 30 .

[0047] <Step 201>

[0048] The data collection unit 31 receives a plurality of receive beam data after phase adjustment by the receive beam former 22, arranges the received receive beam data in the zx plane, and generates frame data 301. The data collection unit 31 receives signals when the probe 10 transmits and receives ultrasound waves N times in the same imaging range, and obtains N frames of frame data 301. The data collection unit 31 stores the time series (N frames) of frame data 301 in the form of IQ (In-Phase / Quadrature-Phase) signals in the memory 40. Thus, as shown in FIG. Figure 3 As shown in step 201 , three-dimensional data 300 arranged in the x, z, and N directions are stored in the memory 40 .

[0049] The ultrasonic waves reflected by the blood flow in the subject 1 are very few, and most of the ultrasonic waves are reflected at the interface of the structure in the subject 1. Therefore, most of each frame data 301 is the background component (clutter component) reflected by the structure, and the proportion of the blood flow component is very small. Therefore, through the following processing, the clutter component of each frame data 301 is removed and suppressed, and the blood flow component is extracted.

[0050] <Step 202>

[0051] The matrix conversion unit 32 extracts data u of one pixel (position zx) of the frame data 301 from the corresponding pixels (position zx) of the N frames 301. zxi , and arrange N frames, thus generating vector u as in equation (1) zx In addition, i represents the number of the frame data in the N direction. The vector u zx It is the integrated data of singular value decomposition.

[0052] Furthermore, according to formula (2), according to the vector u zx Generate the correlation matrix R zx .

[0053] [Formula 1]

[0054] u zx =[u zx1 ,…,u zxN ] T …(1)

[0055] [Formula 2]

[0056]

[0057] <Step 203>

[0058] The matrix conversion unit 32 obtains the correlation matrix R for each of a plurality of pixels (for example, all pixels). zx, the obtained correlation matrix is ​​averaged by formula (3) to calculate the average correlation matrix R - (In addition, R - of" - ” represents the “-” on the character R in formula (3)).

[0059] [Formula 3]

[0060]

[0061] <Step 204>

[0062] The matrix analysis unit 33 calculates the average correlation matrix R obtained by the matrix conversion unit 32. - , according to formula (4), the specific value λ of the Nth principal component is calculated from the first principal component according to each level i i and the unique vector w i (i=1, 2...N). (Here, level i represents the number of each principal component obtained by principal component analysis based on special value decomposition, the first principal component equivalent to the upper level is called the principal component of level 1, and the Nth principal component is called the principal component of level N.) In addition, formula (4) is the calculation process of principal component analysis, but when the frame data 301 is a square matrix, eigenvalue decomposition can also be used as an alternative method.

[0063] [Formula 4]

[0064]

[0065] <Steps 205, 206>

[0066] Next, the singular value analyzing unit 34 determines a threshold level for separating the main component of the blood flow component from the main component of the clutter component by utilizing the fact that the blood flow component included in the received signal is less dominant than the clutter component.

[0067] The singular value analysis unit 34 calculates the singular value λ of each principal component calculated by singular value decomposition. i (diagonal components of matrix Λ). Specifically, the singular value λ is obtained i Divide by the singular value λ i The total sum of (∑λ i ) and the obtained λ i / ∑λ i (Step 204). i / ∑λ i It indicates how much each principal component contributes to the reproduction of the original data, that is, the information dominance.

[0068] like Figure 3 As shown in step 205, the i / ∑λ iThe graph 310 is a graph with level i (i=1 to N) as the vertical axis and level i (i=1 to N) as the horizontal axis. i / ∑λ i The level of the principal component corresponding to the threshold value (level threshold k). i / ∑λ i An example of the threshold value of is 0.01, and thus the energy of the blood flow component relative to the energy of the tissue component is approximately 1%.

[0069] In addition, a preset value may be used as the level threshold k.

[0070] <Steps 207, 208>

[0071] The unique vector (w i , i=k to N) can function as a filter for suppressing clutter components included in the received signal and extracting blood flow components. The image reconstruction unit 37 reconstructs a blood flow image from which the blood flow components are extracted using the above-mentioned unique vectors.

[0072] Specifically, the image reconstruction unit 37 generates the sum of products of a plurality of unique vectors and unique values ​​in a given level range from the threshold level k to N as the blood flow extraction filter P using equation (5): k (Step 207).

[0073] Next, the image reconstruction unit 37 multiplies the plurality of frame data 301 (for example, all three-dimensional data U of the frame data 301) by the blood flow extraction filter P using equations (6) and (7). k , and obtain the blood flow component frame data U^. (In addition, the "^" in U^ represents the "^" on the character U in formula (6)).

[0074] Furthermore, the image reconstruction unit 37 performs addition averaging on the obtained three-dimensional data (N frames of data) U^ for each pixel (position zx) in the time axis direction (N direction) by using equation (8), thereby generating a blood flow image U avg (Step 208).

[0075] [Formula 5]

[0076]

[0077] [Formula 6]

[0078]

[0079] [Formula 7]

[0080]

[0081] [Formula 8]

[0082]

[0083] <Steps 209-211>

[0084] The filter element generation unit 35 generates at least one of the first filter element and the second filter element as a filter element. In the present embodiment, both the first filter element and the second filter element are calculated.

[0085] (First filter element)

[0086] The first filter element is based on a plurality of unique vectors (w) in a given level range that are greater than a threshold level k (that is, including levels below level k). i , i=k to N) by multiplying the plurality of frame data 301 by the variances between the data at the positions zx corresponding to the plurality of blood flow component frame data (step 209). Here, the first filter element is the inverse of the variance value.

[0087] Specifically, the calculation method of the first filter element is described. The filter element generation unit 35 uses the unique vectors (w i , i = k ~ N), the blood flow extraction filter P is obtained by formula (5) k , through formula (6), multiple frame data 301 (N pieces of all frame data 301, that is, three-dimensional data U) are multiplied by the blood flow extraction filter P k , and obtain N blood flow component frame data U^. The filter element generation unit 35 calculates the variance (var[U^]) in the time axis direction (N direction) for the corresponding pixels (position zx) of the N blood flow component frame data U^ through equation (9), and uses the inverse of the variance as the first filter element U^. var . (Wherein, in formula (9), var[] represents a function for calculating the brightness variance in the time axis direction).

[0088] [Formula 9]

[0089]

[0090] (Second filter element)

[0091] The second filter element is based on the specific vector w of a given level that is less than a threshold level k. i The filter element generation unit 35 uses the unique vector of level 1 as the unique vector of a given level smaller than the threshold level k. The second filter element is the value of the pixel of the tissue component frame data.

[0092] Specifically, the calculation method of the second filter element is described. The filter element generation unit 35 generates a clutter extraction filter P1 (step 210) according to the product of the unique vector w1 and the unique value λ1 of the first principal component by using formula (10) based on the result of the principal component analysis. The filter element generation unit 35 multiplies the clutter extraction filter P1 by the N-frame data 301 (three-dimensional data U) and calculates the N-frame tissue component frame data (three-dimensional data) U1^ by using formula (11).

[0093] The filter element generation unit 35 adds and averages the corresponding pixels of the N tissue component frame data (three-dimensional data) U1^ of the obtained time series in the time axis direction (N direction), thereby obtaining the frame data U1 representing the distribution of brightness centered on the clutter component, and uses it as the second filter element. In the formula (12), avg[] represents a function for calculating the average in the time axis direction (N direction).

[0094] [Formula 10]

[0095]

[0096] [Formula 11]

[0097]

[0098] [Formula 12]

[0099]

[0100] The image processing unit 38 processes the image by the first filter element U var The image processing unit 38 generates a clutter-suppressed image by weighting the pixel values ​​of at least one of the frame data 301 or the image generated based on the frame data in the second filter element U1. var The second filter element U1 has an effect on the blood flow image U obtained in step 208. avg The pixel values ​​are weighted to generate a clutter suppressed image.

[0101] Specifically, the mask image generation unit 36 ​​generates the first filter element U by using equation (13): var The second filter element U1 is multiplied by the coefficients β and α (the sum of the weight coefficients α and β is 1) and weighted addition is performed to generate a mask image MASK (a two-dimensional coefficient matrix (mask matrix)) whose reciprocal value is used as the pixel value (step 212). The relationship between α and β is added and set to 1 as described above, which has the advantage of being able to make the variable parameter one and simplifying the user's adjustment. In particular, when image quality is prioritized, it is sometimes effective to treat α and β as independent variables, and a design that switches according to the operator's purpose is effective.

[0102] [Formula 13]

[0103]

[0104] As mentioned above, the first filter element U var It is the inverse of the brightness variance in the time direction (N direction) of the blood flow component frame data. Generally speaking, the blood flow component with three-dimensional flow has a larger time variation of brightness than the movement of organ tissue. Therefore, by setting the inverse of the brightness variance as the pixel value of the mask, a mask image MASK with a high numerical value (pixel value) can be generated in the pixels of the tissue area of ​​the subject 1.

[0105] On the other hand, the second filter element U1 is the brightness average of the noise component (tissue component) frame data (three-dimensional data) U1^ extracted by the specific vector of the first principal component. Compared with the brightness data of the frame data of the received signal, the noise component is a dominant factor in that it is a signal that is stable in space and time and is not easily affected by blood flow and noise. Therefore, by setting the reciprocal of the second filter element U1 as the pixel value of the mask, a mask image MASK that can selectively extract the boundary and high brightness area of ​​the organ tissue can be generated.

[0106] In other words, the mask image generator 36 uses the first filter element and the second filter element showing high values ​​in the tissue region that is a clutter source and sets their reciprocals as mask pixel values, thereby generating a mask image MASK having a clutter suppression effect (step 213).

[0107] The image reconstruction unit 37 uses equation (14) to reconstruct the blood flow image U generated in step 208. avg The mask image MASK is multiplied to perform image reconstruction, thereby generating a blood flow image U with less residual clutter and enhanced blood flow and high visual recognition. MASK (Step 213).

[0108] [Formula 14]

[0109] U MASK =U avg ×MASK…(14)

[0110] The image reconstruction unit 37 can generate a blood flow image U with less residual clutter and an enhanced blood flow area. MASK The image reconstruction unit 37 converts the blood flow image U MASK Output to the display processing unit 60.

[0111] On the other hand, the B-mode image generation unit 50 receives the reception beam data from the reception beamformer 22 , generates a B-mode image, and outputs the B-mode image to the display processing unit 60 .

[0112] The display processing unit 60 is, for example, Figure 4 As shown, the B-mode image and the blood flow image U with the blood flow area emphasized are MASK The image reconstruction unit 37 may also display the blood flow image U generated in step 208 in areas 401 and 402 of the screen of the display device 70, respectively. avg Displayed together with area 403.

[0113] The mask image generating unit 36 ​​may be configured to receive from the user on the display screen of the display device 70 the first filter element U used when generating the mask image MASK according to the equation (13). var and the coefficients β and α (α+β=1) of the second filter element U1. Figure 4 The bar 404 is displayed on the display screen. The user moves the mark 405 on the bar 404 to the left or right using an external input device 80 such as a mouse to set the ratio of the coefficient α to the coefficient β. The mask image generation unit 36 ​​generates the mask image MASK in step 212 using the set ratio of the coefficient α to the coefficient β.

[0114] The first feature of this embodiment is to use the second filter element U1 using the first principal component as a tissue component (clutter component). Thus, the mask image can be used to suppress not only low-brightness areas of organ tissues but also clutter components remaining in high-brightness areas.

[0115] That is, the second filter element U1 is an image of level 1 (first principal component), and is therefore an image of only tissue (background), and the blood flow component does not contain electrical noise. Therefore, in order to remove high-brightness components such as blood vessel walls contained in the blood flow image, it is more appropriate to use the original image (frame data).

[0116] In addition, the second feature of this embodiment is to use the blood flow image U avg The time-direction variance of the first filter element U var According to the blood flow image U avg The brightness change over time is used to adjust the blood flow image U using the mask image avg The brightness of the image can be expected not only to suppress the clutter component but also to enhance the blood flow component.

[0117] That is, the first filter element U is calculated by using N blood flow component frame data U^ richly containing blood flow components. var , thus being able to effectively remove blood flow.

[0118] In addition, in this embodiment, the blood flow image U avgThe generation of the first and second filter elements is also performed in all principal component analyses. Therefore, compared with the case where a different process from the principal component analysis is performed to suppress clutter components, the calculation efficiency is high, and the blood flow of fine blood vessels can be extracted with high accuracy in a short time. In addition, no additional processing circuit is required, and the device structure can be simplified.

[0119] 《<Implementation Method 2>》

[0120] The ultrasonic imaging device of the second embodiment is described. The ultrasonic imaging device of the second embodiment has the same configuration as that of the device of the first embodiment, but further includes a method of appropriately setting the first filter element U var The function of multiplying the second filter element U1 by the coefficients β and α (α+β=1) is that the image processing unit 38 calculates the similarity of the plurality of frame data 301 and sets the weight coefficients α and β based on the calculation result.

[0121] When the correlation (similarity) between the N frames of data 301 for which principal component analysis is performed is high, it is appropriate for the coefficients β and α to be around 0.5. This is because when the correlation (similarity) between the N frames of data 301 is high, the clutter component and the blood flow component can be effectively separated in the principal component analysis, and the brightness of the clutter component also becomes high in the calculation result of the second filter element U1. Therefore, the first filter element U1 is generated. var The inverse mask image MASK is obtained by weighted addition of the second filter element U1 and multiplied by the blood flow image U avg , thus achieving clutter suppression effect.

[0122] On the other hand, when the correlation of the N frames of data 301 is low, it is assumed that the clutter component crosses the plurality of principal components, and the brightness of the clutter component of the second filter element U1 becomes a low value. In this state, when the mask image MASK is constructed in step 212, the mask image MASK is combined with the blood flow image U1. avg Multiplication can produce a phenomenon in which the clutter component of the blood flow image avg is emphasized instead.

[0123] In order to avoid this phenomenon, when the correlation between the N frames of frame data 301 is low, it is effective to set the value of the coefficient α to be low and suppress the contribution ratio of the second filter element U1 to the mask image MASK.

[0124] For example, Figure 5 As shown, the straight line graph 501 and the curved line graphs 502 and 503 are set and stored in the memory 40 so that the coefficient α of the second filter element U1 decreases as the correlation value of the N frames of frame data 301 decreases.

[0125] like Figure 6 As shown in the process, the mask image generating unit 36 ​​executes steps 214 and 215 to calculate the correlation value of N frames of data 301. Figure 5 The coefficient α corresponding to the calculated correlation value is obtained from any one of the graphs 501, 502 and 503. Then, in step 212, the coefficients α and β (=1-α) obtained are used to calculate the second filter element U1 and the first filter element U2 by equation (13). var Perform weighting to obtain the mask image MASK.

[0126] Thereby, an appropriate coefficient α can be automatically set in the ultrasonic imaging device.

[0127] like Figure 7 As shown in (a) and (b) of FIG. 21 , the graph 1001 showing the time variation of the correlation value calculated in step 214 (data stability) and the blood flow image U after residual clutter suppression generated in step 213 may be compared. avg are displayed together on the display device 70 . Figure 7 (a) is a case where the correlation value is maintained high and the data stability is high, and the coefficient α is set to a large value close to 0.5. On the other hand, Figure 7 (b) is a case where the correlation value is maintained low and the stability of the data is low, and the ratio of the coefficient α is set to be small.

[0128] In addition, the calculation of the correlation value of the N frames of frame data 301 in step 214 uses a generally known similarity calculation method such as a cross-correlation calculation.

[0129] exist Figure 6 In the process of Figure 2 The process is the same, so the description is omitted.

[0130] In addition, if Figure 8 As shown in the process of , the mask image generating unit 36 ​​can also use the tissue component frame data (U1^) obtained using the first principal component to determine the similarity of the N frames of frame data 301.

[0131] The tissue component frame data is a component (three-dimensional data) based on the tissue structure extracted by the specific vector of the first principal component. As mentioned above, the reflection signal from the biological body is mainly the component from the organ tissue, and the component from the blood flow is less than 10%-1%. That is, it is appropriate to judge the component of the organ tissue as the main component based on the similarity of the data.

[0132] Therefore, the mask image generation unit 36 ​​executes steps 216 and 217 before step 212 to calculate the correlation value of the tissue component frame data. Figure 5The coefficient α corresponding to the calculated correlation value is obtained from the graph.

[0133] By calculating the correlation value of N frames of data 301 using tissue component frame data, the influence of tissue movement and noise that affect the principal component analysis can be eliminated, and the correlation can be determined more effectively. As a result, an appropriate coefficient α can be automatically set in the ultrasonic imaging device. In addition, α and β are treated as dependent variables (α+β=1) here, but this limitation is only a setting that focuses on the simplicity of operation. While maintaining the described device mode, there is no limitation on treating the two as independent variables (in other words, 0<=α<=1, 0<=β<=1).

[0134] exist Figure 8 In the process of Figure 2 The process is the same, so the description is omitted.

[0135] Note that the configuration, operation, and effects of the ultrasonic imaging device of the second embodiment other than those described above are the same as those of the first embodiment, and thus description thereof will be omitted.

[0136] 《〈Implementation Method 3>》

[0137] The ultrasonic imaging device of Embodiment 3 will be described. The ultrasonic imaging device of Embodiment 3 has the same configuration as that of Embodiment 1, but further has a function of determining the correlation (similarity) between N frames of frame data 301 and adjusting the number of acquired data.

[0138] In the matrix conversion unit 32, the matrix analysis unit 33, the singular value analysis unit 34, and the filter element generation unit 35, when the number N of frame data 301 used in the principal component analysis is large, the number of principal components to be distinguished increases, and the separation accuracy of clutter and blood flow becomes higher. However, the calculation load in the blood flow image generation unit 30 becomes heavy, which may lead to a decrease in the frame rate.

[0139] Therefore, in the third embodiment, when the correlation between the frame data 301 is sufficiently high, the number of acquired frames N is reduced, thereby achieving both the effects of clutter suppression and frame rate improvement.

[0140] Specifically, if Fig. 9 As shown in the flow, before step 201 of collecting N frames of data 301, step 200 of setting the number N of frames of data 301 to be collected is executed.

[0141] In step 200, the correlation value of the tissue component frame data calculated in the previous step 216 is received, for example, according to Fig.10Thus, the graphs 901 , 902 , 903 and the like showing the relationship between the predetermined correlation value and the number of frames are set to an appropriate number of frames N. As described in the second embodiment, the correlation value of the tissue component frame data corresponds to the correlation value of the N frames of frame data 301 .

[0142] like Fig.10 As shown, the straight line graph 901 and the curved line graphs 902 and 903 are set so that the higher the correlation value of the N-frame data 301 is, the smaller the number of acquired frames N is. These graphs 901 to 903 are stored in the memory 40 .

[0143] By using Fig. 9 The process can use the correlation value of the tissue component frame data calculated in the previous step 216 to set the number N of frame data 301 used in the current principal component analysis to an appropriate number.

[0144] In addition, as a method for calculating the correlation value of the N-frame data 301, in addition to the method of calculating the correlation value of the tissue component frame data, it is also possible to calculate the correlation value of the N-frame data 301 by Figure 6 In step 214 of the process, the correlation value is calculated directly from the frame data 301.

[0145] In addition, if the value of the appropriate number of frames N changes significantly each time the principal component analysis is performed, brightness flickers at the timing of switching the displayed blood flow image, which may impair the visual recognition of the dynamic image. In order to suppress this, it is effective to have a restriction measure such as limiting the change in the number of data between consecutive frames to less than 2 frames.

[0146] Note that the configuration, operation, and effects of the ultrasonic imaging device of the third embodiment other than those described above are the same as those of the first and second embodiments, and thus description thereof will be omitted.

Claims

1. An ultrasonic imaging device, characterized in that: have: A data collection unit receives a plurality of reception signals obtained by receiving sequentially reflected ultrasound waves using an array of a plurality of vibrators arranged in the x direction in the depth z direction of the subject to which the ultrasound waves have been transmitted, and repeatedly arranges the reception signals on the zx plane to generate frame data, thereby generating N frames of frame data; A matrix conversion unit generates a correlation matrix based on a vector obtained by arranging data at a corresponding position zx of the frame data by N frames; A matrix analysis unit, performing singular value decomposition on the correlation matrix, and calculating a singular value and a singular vector of each level of N levels of the received signal; a filter element generating unit that generates at least one of a first filter element and a second filter element as a filter element; as well as an image processing unit that generates a clutter-suppressed image by weighting the pixel values ​​of the frame data or an image generated based on the frame data using the filter elements generated by the filter element generating unit, The first filter element is calculated based on the variance between the data of the position zx corresponding to the plurality of blood flow component frame data obtained by multiplying the plurality of frame data by a plurality of specific vectors in a given level range above a predetermined threshold level k, The second filter element is calculated based on tissue component frame data obtained by multiplying a specific vector of a given level smaller than the threshold level k by the frame data equal to or greater than 1.

2. The ultrasonic imaging device according to claim 1, characterized in that: The first filter element is the inverse of the value of the variance.

3. The ultrasonic imaging device according to claim 1, characterized in that: The ultrasonic imaging device further includes an image reconstruction unit that reconstructs a blood flow image from which blood flow components are extracted. The image reconstruction unit multiplies a plurality of the frame data by a plurality of specific vectors in a given level range from the threshold level k to N as blood flow extraction filters to obtain a plurality of blood flow component frame data, and generates the blood flow image by averaging the obtained plurality of blood flow component frame data at each position zx as pixel values, The image processing unit generates the clutter-suppressed image by weighting pixel values ​​of the blood flow image using at least one of the first filter element and the second filter element.

4. The ultrasonic imaging device according to claim 3, characterized in that: The image processing unit generates, for each position zx, a mask image having a pixel value obtained by multiplying the first filter element and the second filter element by a weight coefficient and averaging the values, and multiplies the pixel value of the mask image at a position corresponding to the pixel value of the blood flow image, thereby generating the clutter suppressed image.

5. The ultrasonic imaging device according to claim 1, characterized in that: The filter element generation unit calculates a first filter element using a plurality of unique vectors in a level range from the threshold level k to level N as a blood flow component filter.

6. The ultrasonic imaging device according to claim 1, characterized in that: The filter element generation unit generates a second filter element using a unique vector of level 1 as a unique vector of a given level smaller than the threshold level k.

7. The ultrasonic imaging device according to claim 1, characterized in that: The matrix conversion unit obtains the correlation matrix for each of the plurality of positions zx, and obtains a correlation matrix by averaging the correlation matrices. The matrix analysis unit calculates the unique value and the unique vector based on the average correlation matrix obtained by the matrix conversion unit.

8. The ultrasonic imaging device according to claim 4, characterized in that: The sum of the weight coefficient of the first filter element and the weight coefficient of the second filter element is 1.

9. The ultrasonic imaging device according to claim 1, characterized in that: The ultrasonic imaging device further includes a singular value analyzing unit configured to analyze the singular value and calculate the threshold level k.

10. The ultrasonic imaging device according to claim 4, characterized in that: The image processing unit calculates similarities between the plurality of frame data and sets the weight coefficient based on the calculation result.

11. The ultrasonic imaging device according to claim 10, characterized in that: The image processing unit calculates similarity of the plurality of frame data by calculating similarity of the tissue component frame data.

12. The ultrasonic imaging device according to claim 4, characterized in that: The data collection unit calculates similarities between the plurality of frame data and sets the number N of frames of data to be collected based on the calculation result.

13. A signal processing method, characterized in that: have: A data collection step of receiving a plurality of reception signals obtained by receiving sequentially reflected ultrasound waves using an array of a plurality of vibrators arranged in the x direction in the depth z direction of the subject to which the ultrasound waves have been transmitted, and repeatedly arranging the reception signals on the zx plane to generate frame data, thereby generating N frames of frame data; A matrix transformation step, generating a correlation matrix according to a vector obtained by arranging the data at the position zx corresponding to the frame data by N frames; A matrix analysis step, performing singular value decomposition on the correlation matrix, and calculating the singular value and singular vector of each level of N levels of the received signal; A filter element generating step of generating at least one of a first filter element and a second filter element as a filter element; as well as an image processing step of weighting the pixel values ​​of the frame data or an image generated based on the frame data using the filter elements generated by the filter element generating unit to generate a clutter suppressed image, In the filter element generation step, The first filter element is calculated based on the variance between the data of the position zx corresponding to the plurality of blood flow component frame data obtained by multiplying the plurality of frame data by a plurality of specific vectors in a given level range above a predetermined threshold level k, The second filter element is calculated based on tissue component frame data obtained by multiplying a specific vector of a given level smaller than the threshold level k by the frame data equal to or greater than 1.

14. A computer program product comprising a signal processing program, characterized in that The signal processing program causes the computer to execute the following steps: A data collection step of receiving a plurality of reception signals obtained by receiving sequentially reflected ultrasound waves using an array of a plurality of vibrators arranged in the x direction in the depth z direction of the subject to which the ultrasound waves have been transmitted, and repeatedly arranging the reception signals on the zx plane to generate frame data, thereby generating N frames of frame data; A matrix transformation step, generating a correlation matrix according to a vector obtained by arranging the data at the position zx corresponding to the frame data by N frames; A matrix analysis step, performing singular value decomposition on the correlation matrix, and calculating the singular value and singular vector of each level of N levels of the received signal; A filter element generating step of generating at least one of a first filter element and a second filter element as a filter element; as well as an image processing step of weighting the pixel values ​​of the frame data or an image generated based on the frame data using the filter elements generated by the filter element generating unit to generate a clutter suppressed image, In the filter element generation step, The first filter element is calculated based on the variance between the data of the position zx corresponding to the plurality of blood flow component frame data obtained by multiplying the plurality of frame data by a plurality of specific vectors in a given level range above a predetermined threshold level k, The second filter element is calculated based on tissue component frame data obtained by multiplying a specific vector of a given level smaller than the threshold level k by the frame data equal to or greater than 1.

Images