A 4D intracardiac ultrasound imaging apparatus and method based on row-column addressing PMUT, medium

Abstract

The application relates to the technical field of ultrasonic imaging, and discloses a 4D intracardiac ultrasonic imaging device and method based on row-column addressing PMUT and a medium. It can be seen that the application can simplify the number of lead wires through row-column addressing wiring, so that the PMUT array can meet the requirements of a narrow space of a heart cavity and realize independent control of each array element. Moreover, a first selected area with a signal-to-noise ratio less than a first threshold value is obtained through signal-to-noise ratio distribution. A correlation matrix is obtained according to the signal-to-noise ratio intensity and the emission amplitude matrix of the intracardiac ultrasonic image at a historical moment in the first selected area. The initial emission amplitude matrix at the current moment is optimized according to the correlation, so that the PMUT array can enhance the signal release intensity of the low signal-to-noise ratio area according to the correlation degree, and then 4D intracardiac ultrasonic imaging is performed through the optimized PMUT array, thereby improving the imaging precision of 4D ICE.

Description

Technical Field

[0001] This invention relates to the field of ultrasound imaging technology, and in particular to a 4D intracardiac ultrasound imaging device, method, and medium based on row and column addressing PMUT. Background Technology

[0002] 4D intracardiac ultrasound imaging (4D ICE) is a real-time three-dimensional intracardiac ultrasound that incorporates the time dimension into traditional three-dimensional static images, enabling continuous and dynamic real-time imaging of the heart structure.

[0003] In existing technologies, mainstream 4D ICE devices based on traditional piezoelectric ceramics (PZT) suffer from high latency and low frame rate, affecting the accuracy of real-time imaging. Furthermore, their large overall thickness makes them susceptible to damage during imaging. In contrast, micromachined piezoelectric ultrasonic transducers (PMUTs) utilize a silicon-based thin-film stacked structure, eliminating the need for matching and backing layers required by traditional PZTs. Their smaller overall thickness reduces the risk of damage. Moreover, PMUTs are fabricated using MEMS wafer-level processes, exhibiting superior energy conversion efficiency, low power consumption, and high integration density, and can operate stably at mid-to-high frequencies. Therefore, PMUT-based 4D ICE technology outperforms traditional PZTs in miniaturization, integration, and imaging quality.

[0004] However, when ultrasound propagates through the heart, the amplitude of the far-field echo is lower than that of the near-field echo, and the energy decays exponentially with depth, resulting in low grayscale in the far field and forming a dark area. Simultaneously, ultrasound waves are completely reflected or absorbed when encountering strongly reflective or highly attenuated structures (such as calcified valves, metal occluders, ribs, etc.), preventing imaging of tissues behind these structures. Furthermore, PMUT may experience decreased imaging accuracy due to insufficient signal-to-noise ratio in areas with greater detection depth, severe tissue acoustic attenuation, or complex signal paths. These factors affect the imaging quality of 4D ICE, and existing 4D ICE technologies suffer from low imaging accuracy. Summary of the Invention

[0005] This invention provides a 4D intracardiac ultrasound imaging device, method, and medium based on row and column addressing PMUT, which can improve the imaging quality of 4D ICE.

[0006] The first aspect of this invention discloses a 4D intracardiac ultrasound imaging method based on row-column addressing PMUT, the method comprising: An initial ultrasound image set is acquired, which includes all 4D intracardiac ultrasound images at the current and historical times. The PMUT array includes multiple PMUT elements, and the PMUT array uses row and column addressing wiring to achieve independent control of each element. Based on the initial set of ultrasound images, the signal-to-noise ratio distribution and emission amplitude matrix corresponding to each 4D intracardiac ultrasound image are obtained. For each 4D intracardiac ultrasound image: the signal-to-noise ratio distribution is used to represent the signal-to-noise ratio of each voxel on the 4D intracardiac ultrasound image, and each element in the emission amplitude matrix is ​​used to represent the signal release intensity of the PMUT array element corresponding to that element when the 4D intracardiac ultrasound image is acquired. Based on the signal-to-noise ratio (SNR) distribution at the current moment, a first selected region is obtained. The first selected region is the region corresponding to the first voxel, which is the voxel in the 4D intracardiac ultrasound image at the current moment whose SNR is less than a preset first threshold. In the first selected region: based on all SNR distributions and all emission amplitude matrices corresponding to all 4D intracardiac ultrasound images at all historical moments, a correlation matrix is ​​obtained. The elements in the correlation matrix are used to represent the degree of correlation between the change in signal release intensity and the change in SNR corresponding to each PMUT element. Based on the correlation matrix, the initial emission amplitude matrix at the current moment is optimized to optimize the PMUT array, resulting in an optimized PMUT array used for 4D intracardiac ultrasound imaging.

[0007] As an optional implementation, in the first aspect of the present invention, the specific operation of optimizing the initial transmission amplitude matrix at the current moment according to the correlation matrix to obtain the optimized PMUT array includes: constructing a first perturbation matrix according to the correlation matrix and a preset perturbation parameter, wherein the sum of all elements in the first perturbation matrix is ​​equal to the preset perturbation parameter, the perturbation parameter is used to adjust the signal release intensity of the PMUT array elements, and the dimension of the first perturbation matrix is ​​the same as that of the initial transmission amplitude matrix; The initial emission amplitude matrix is ​​perturbed and updated according to the first perturbation matrix to obtain the transition emission amplitude matrix; the signal-to-noise ratio (SNR) distribution of the transition ultrasound image and the transition ultrasound image is obtained according to the transition emission amplitude matrix; the SNR level increment corresponding to the transition ultrasound image is obtained, where the SNR level is the total SNR of all voxels in the transition ultrasound image corresponding to the first selected region, and the SNR level increment is the increase in the SNR level between the current perturbation update round and the previous round; When it is determined that the signal-to-noise ratio increment corresponding to the transition ultrasound image is greater than a preset second threshold, multiple rounds of perturbation update operations are performed on the transition emission amplitude matrix based on the first perturbation matrix until it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix is ​​less than the preset second threshold. Then, the transition emission amplitude matrix obtained from the previous round of perturbation update operation is output to optimize the PMUT array and obtain the optimized PMUT array.

[0008] As an optional implementation, in the first aspect of the present invention, the method further includes: When it is determined that the signal-to-noise ratio increment corresponding to the transition ultrasound image is less than a preset second threshold, in the first selected area: according to the signal-to-noise ratio distribution corresponding to the transition ultrasound image and the transition emission amplitude matrix, the correlation matrix is ​​updated to construct a second perturbation matrix based on the result of the transition update, and the transition emission amplitude matrix is ​​perturbed and updated based on the second perturbation matrix to obtain the updated transition emission amplitude matrix; When it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix is ​​greater than the preset second threshold, multiple rounds of perturbation update operations are performed on the updated transition emission amplitude matrix based on the second perturbation matrix until it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix after the perturbation update is less than the preset second threshold. Then, the transition emission amplitude matrix obtained in the previous round of perturbation update operation is output to optimize the PMUT array and obtain the optimized PMUT array.

[0009] As an optional implementation, in the first aspect of the present invention, after perturbating and updating the transition emission amplitude matrix according to the second perturbation matrix to obtain the updated transition emission amplitude matrix, the method further includes: When it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix is ​​less than the preset second threshold, the 4D intracardiac ultrasound image to be processed is obtained according to the PMUT array in the initial state. Based on the 4D intracardiac ultrasound image to be processed, its corresponding signal-to-noise ratio distribution is obtained; based on the signal-to-noise ratio distribution, the attention distribution corresponding to each voxel in the 4D intracardiac ultrasound image to be processed is obtained; based on a preset image enhancement model, the 4D intracardiac ultrasound image to be processed is enhanced according to the attention distribution to obtain an enhanced 4D intracardiac ultrasound image, which is used as the final 4D intracardiac ultrasound imaging result.

[0010] As an optional implementation, in the first aspect of the present invention, the method further includes: Construct a first loss function and a second loss function. The first loss function is used to measure the noise intensity of the image, and the second loss function is used to measure the edge gradient of the image. The training process of the image enhancement model is constrained by the first loss function and the second loss function to obtain a trained image enhancement model. The trained image enhancement model is used to enhance the 4D intracardiac ultrasound image to be processed according to the attention distribution.

[0011] As an optional implementation, in the first aspect of the present invention, the method further includes: Spatial clustering is performed on the first voxel to obtain the first voxel cluster; the location range corresponding to the first voxel cluster whose total number of all voxels is greater than a preset third threshold is determined as the undetermined selection area; the proportion of low quality voxels is obtained, wherein the proportion of low quality voxels is the proportion of all voxels in the undetermined selection area relative to all voxels in the current 4D intracardiac ultrasound image. When it is determined that the proportion of low-quality voxels is less than a preset fourth threshold, 4D intracardiac ultrasound imaging results are obtained based on the PMUT array. When it is determined that the proportion of low-quality voxels is greater than the preset fourth threshold, the undetermined selection area is determined as the first selection area.

[0012] As an optional implementation, in the first aspect of the present invention, the step of spatially clustering the first voxel to obtain a first voxel cluster includes: Determine the neighboring voxels corresponding to the target first voxel, wherein the target first voxel is any first voxel, and the spatial distance between the neighboring voxel and the target first voxel is less than a preset adjacency radius; obtain the signal-to-noise ratio difference, wherein the signal-to-noise ratio difference is the difference between the signal-to-noise ratio corresponding to the target first voxel and the signal-to-noise ratio corresponding to the neighboring voxel; Neighboring voxels with a signal-to-noise ratio difference less than a preset second threshold are designated as second voxels; based on all first voxels and all second voxels, a first voxel cluster is obtained.

[0013] A second aspect of the present invention discloses a 4D intracardiac ultrasound imaging device based on row and column addressing PMUT, the device comprising: A PMUT array is used to acquire an initial set of ultrasound images, which includes all 4D intracardiac ultrasound images at the current and historical times. The PMUT array uses row and column addressing wiring to achieve independent control of each array element. The preprocessing module is used to obtain the signal-to-noise ratio distribution and emission amplitude matrix corresponding to each 4D intracardiac ultrasound image based on the initial ultrasound image set. The core processing module is used to obtain a first selected region based on the signal-to-noise ratio distribution at the current moment. The first selected region is the region corresponding to a first voxel, which is a voxel in the 4D intracardiac ultrasound image at the current moment whose signal-to-noise ratio is less than a preset first threshold. In the first selected region, a correlation matrix is ​​obtained based on all signal-to-noise ratio distributions and all emission amplitude matrices corresponding to all 4D intracardiac ultrasound images at historical moments. The elements in the correlation matrix are used to represent the degree of correlation between the change in signal release intensity and the change in signal-to-noise ratio corresponding to each PMUT element. The array element adjustment module is used to optimize the initial emission amplitude matrix at the current time according to the correlation matrix, so as to optimize the PMUT array and obtain the optimized PMUT array, which is used for 4D intracardiac ultrasound imaging.

[0014] As an optional implementation, in a second aspect of the present invention, the array element adjustment module optimizes the initial emission amplitude matrix at the current moment based on the correlation matrix to obtain the optimized PMUT array's specific operation mode, including: A first perturbation matrix is ​​constructed based on the correlation matrix and preset perturbation parameters. The sum of all elements in the first perturbation matrix is ​​equal to the preset perturbation parameters. The perturbation parameters are used to adjust the signal release intensity of the PMUT array elements. The dimension of the first perturbation matrix is ​​the same as that of the initial transmission amplitude matrix. The initial emission amplitude matrix is ​​perturbed and updated according to the first perturbation matrix to obtain the transition emission amplitude matrix; the signal-to-noise ratio (SNR) distribution of the transition ultrasound image and the transition ultrasound image is obtained according to the transition emission amplitude matrix; the SNR level increment corresponding to the transition ultrasound image is obtained, where the SNR level is the total SNR of all voxels in the transition ultrasound image corresponding to the first selected region, and the SNR level increment is the increase in the SNR level between the current perturbation update round and the previous round; When it is determined that the signal-to-noise ratio increment corresponding to the transition ultrasound image is greater than a preset second threshold, multiple rounds of perturbation update operations are performed on the transition emission amplitude matrix based on the first perturbation matrix until it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix is ​​less than the preset second threshold. Then, the transition emission amplitude matrix obtained from the previous round of perturbation update operation is output to optimize the PMUT array and obtain the optimized PMUT array.

[0015] As an optional implementation, in a second aspect of the invention, the apparatus further includes: The first judgment module is used to, when it is determined that the signal-to-noise ratio level increment corresponding to the transition ultrasound image is less than a preset second threshold, in the first selected area: according to the signal-to-noise ratio distribution corresponding to the transition ultrasound image and the transition emission amplitude matrix, perform a transition update on the correlation matrix, construct a second perturbation matrix according to the result of the transition update, and perform a perturbation update on the transition emission amplitude matrix according to the second perturbation matrix to obtain the updated transition emission amplitude matrix; The second judgment module is used to perform multiple rounds of perturbation update operations on the updated transition emission amplitude matrix based on the second perturbation matrix when it is determined that the signal-to-noise ratio level increment corresponding to the updated transition emission amplitude matrix is ​​greater than a preset second threshold, until it is determined that the signal-to-noise ratio level increment corresponding to the updated transition emission amplitude matrix after the perturbation update is less than the preset second threshold. Then, it outputs the transition emission amplitude matrix obtained in the previous round of perturbation update operation to optimize the PMUT array and obtain an optimized PMUT array.

[0016] As an optional implementation, in a second aspect of the invention, the first determining module is further configured to, after perturbing and updating the transition emission amplitude matrix according to the second perturbation matrix to obtain the updated transition emission amplitude matrix: When it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix is ​​less than the preset second threshold, the 4D intracardiac ultrasound image to be processed is obtained according to the PMUT array in the initial state. Based on the 4D intracardiac ultrasound image to be processed, its corresponding signal-to-noise ratio distribution is obtained; based on the signal-to-noise ratio distribution, the attention distribution corresponding to each voxel in the 4D intracardiac ultrasound image to be processed is obtained; based on a preset image enhancement model, the 4D intracardiac ultrasound image to be processed is enhanced according to the attention distribution to obtain an enhanced 4D intracardiac ultrasound image, which is used as the final 4D intracardiac ultrasound imaging result.

[0017] As an optional implementation, in a second aspect of the invention, the apparatus further includes: The model training module is used to construct a first loss function and a second loss function. The first loss function is used to measure the noise intensity of the image, and the second loss function is used to measure the edge gradient of the image. The training process of the image enhancement model is constrained by the first loss function and the second loss function to obtain a trained image enhancement model. The trained image enhancement model is used to enhance the 4D intracardiac ultrasound image to be processed according to the attention distribution.

[0018] As an optional implementation, in a second aspect of the invention, the apparatus further includes: A spatial clustering module is used to perform spatial clustering on the first voxel to obtain a first voxel cluster; determine the location range corresponding to the first voxel cluster whose total number of all voxels is greater than a preset third threshold as a candidate region; obtain the proportion of low-quality voxels, which is the proportion of all voxels in the candidate region relative to all voxels in the current 4D intracardiac ultrasound image; when it is determined that the proportion of low-quality voxels is less than a preset fourth threshold, obtain the 4D intracardiac ultrasound imaging result based on the PMUT array; when it is determined that the proportion of low-quality voxels is greater than the preset fourth threshold, determine the candidate region as the first selection region.

[0019] As an optional implementation, in a second aspect of the present invention, the specific operation of the spatial clustering module device to spatially cluster the first voxel to obtain the first voxel cluster includes: Determine the neighboring voxels corresponding to the target first voxel, wherein the target first voxel is any first voxel, and the spatial distance between the neighboring voxel and the target first voxel is less than a preset adjacency radius; obtain the signal-to-noise ratio difference, wherein the signal-to-noise ratio difference is the difference between the signal-to-noise ratio corresponding to the target first voxel and the signal-to-noise ratio corresponding to the neighboring voxel; Neighboring voxels with a signal-to-noise ratio difference less than a preset second threshold are designated as second voxels; based on all first voxels and all second voxels, a first voxel cluster is obtained.

[0020] A third aspect of the present invention discloses an apparatus comprising a memory and a processor, the apparatus comprising: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute some or all of the steps in the 4D intracardiac ultrasound imaging method based on row and column addressing PMUT as described in any of the first aspects of the present invention.

[0021] The fourth aspect of the present invention discloses a computer storage medium storing computer instructions, which, when invoked, are used to execute some or all of the steps in the 4D intracardiac ultrasound imaging method based on row and column addressing PMUT as described in any of the first aspects of the present invention.

[0022] Compared with the prior art, the present invention has the following beneficial effects: This invention simplifies the number of leads through row and column addressing wiring, enabling the PMUT array to meet the requirements of the confined space of the cardiac chamber while achieving independent control of each array element. Based on this, a first selected region with a signal-to-noise ratio (SNR) less than a first threshold is obtained through SNR distribution. A correlation matrix is ​​obtained based on the SNR intensity and emission amplitude matrix of intracardiac ultrasound images from historical moments within the first selected region. The initial emission amplitude matrix at the current moment is optimized based on this correlation, allowing the PMUT array to enhance the signal release intensity in low SNR regions according to the correlation level. 4D intracardiac ultrasound imaging is then performed using the optimized PMUT array, thereby improving the imaging accuracy of 4D ICE. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a schematic diagram of the structure of a row-column addressing PMUT array disclosed in an embodiment of the present invention; Figure 2 This is a schematic flowchart of a 4D intracardiac ultrasound imaging method based on row and column addressing PMUT disclosed in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of a 4D intracardiac ultrasound imaging device based on row and column addressing PMUT disclosed in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of a device including a memory and a processor disclosed in an embodiment of the present invention. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or end that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or ends.

[0027] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0028] Figure 1 This is a schematic diagram of a row-column addressable PMUT array disclosed in an embodiment of the present invention. To better understand the 4D ICE principle in this solution, Figure 1 For example, traditional 4D imaging uses a single array rotating scan, which is slow and prone to distortion. Figure 1 The row and column addressing PMUT array shown uses two rectangular arrays that are spatially orthogonally stacked (one layer horizontally and one layer vertically alternately stacked). The lower array is dedicated to ultrasonic transmission, and the upper orthogonal array is dedicated to echo reception. The three-dimensional volume is directly obtained through an algorithm.

[0029] Furthermore, in traditional arrays where each element is individually wired (e.g., 1000 wires for 1000 elements), the wire bundle becomes too thick to meet the imaging requirements of confined spaces. Therefore, this embodiment of the invention uses row-column addressing, which can reduce the number of wires by 70%. In addition, this PMUT array uses an AIN thin film, eliminating the need for a matching layer and a backing layer. With an overall thickness of less than 20 micrometers, the catheter outer diameter can be controlled within 11 Fr (approximately 3.67 mm), making it suitable for interventional cardiology procedures for the entire population.

[0030] Example 1 Please see Figure 2 , Figure 2 This is a schematic flowchart of a 4D intracardiac ultrasound imaging method based on row-column addressing PMUT disclosed in an embodiment of the present invention. Figure 2 The described 4D intracardiac ultrasound imaging method based on row-column addressing PMUT can be applied to 4D intracardiac ultrasound imaging devices based on row-column addressing PMUT. For example... Figure 2As shown, this 4D intracardiac ultrasound imaging method based on row and column addressing PMUT can include the following operations: Step 201: Obtain the initial ultrasound image set.

[0031] In this embodiment of the invention, 4D intracardiac ultrasound imaging (4D ICE), also known as real-time three-dimensional imaging, can display the three-dimensional structure of the heart and capture the dynamic process of the heart changing over time. It allows for complete observation of the entire cycle of the valve annulus, leaflets, and defect shunt without the need for repeated catheter angle adjustments.

[0032] In existing technologies, 4D ICE based on block PZT piezoelectric arrays can be implemented in two ways: 1) One-dimensional linear phased array + micro-mechanical rotating structure: Relying on motor to drive the array to rotate and scan to acquire multiple two-dimensional cross-sections, and then stitching them together into a three-dimensional volume; the drawbacks are that there is a delay in mechanical motion, volume distortion, and low frame rate (<10vol / s), and it is impossible to capture high-speed leaflet opening and closing in real time; the swing mechanism increases the outer diameter of the duct, which increases the risk of damage.

[0033] 2) Matrix-type transceiver multiplexed PZT array: Rectangular PZT arrays can realize spatial structure volume imaging, but the traditional block PZT array has a complex manufacturing process, with multiple matching layers and backing layers resulting in large device thickness; all array elements have independent leads, the multi-channel array wiring required for 4D imaging is huge, the shielding performance of single-layer flexible circuit is poor, crosstalk is serious when multiple channels are acquired in parallel, three-dimensional volume artifacts are significant, and the volume frame rate is insufficient, which does not meet the clinical 4D real-time dynamic imaging standard.

[0034] Against the backdrop of the aforementioned 4D ICE, a PMUT (Piezoelectric Micromachined Ultrasonic Transducer) is fabricated using MEMS technology. This transducer utilizes piezoelectric materials (such as AlN aluminum nitride, AlScN scandium-doped aluminum nitride, etc.) to generate mechanical vibrations under electrical signal stimulation, thereby emitting ultrasonic waves or receiving echoes and converting them into electrical signals. The PMUT is extremely small, allowing for easy integration into the catheter tip. Furthermore, ICE requires high resolution to clearly display cardiac structures, and the ability to fabricate a two-dimensional matrix array of PMUTs is crucial for realizing 4D ICE, thus providing high-quality real-time images from inside the heart. Simultaneously, the PMUT's manufacturing process is compatible with CMOS (Complementary Metal-Oxide-Semiconductor) integrated circuit technology, allowing integration with signal processing circuitry on the same chip, reducing complex wiring within the catheter, improving signal quality, and lowering costs.

[0035] While PMUT arrays offer these advantages, their cumbersome high-density 4D sampling channel lead count makes them unsuitable for integration into delicate intraluminal catheters in practical applications. This invention addresses this by introducing row-column addressing, optimizing the M*N row-column arrangement of independent leads (N array elements = N lines) into an additive M+N lead arrangement. This provides physical space for the composite FPC on the catheter (suppressing crosstalk) and reduces sheath rigidity, improving catheter accessibility in tortuous blood vessels. Furthermore, the additive increase in the number of row-column addressing leads allows for subsequent higher-density sampling.

[0036] In this embodiment of the invention, the initial ultrasound image set includes all 4D intracardiac ultrasound images from the current moment and historical moments. Since 4D intracardiac ultrasound imaging is real-time three-dimensional ultrasound imaging, the acquired data is a time-continuous sequence of multiple three-dimensional images. Therefore, the current moment and historical moments can refer to continuous physical time in microseconds / seconds or to the frame acquisition time. Furthermore, it should be noted that to improve the accuracy of subsequent correlation analysis and imaging precision, the initial ultrasound image set can be images from the same phase of the cardiac cycle, or images calibrated according to the cardiac cycle to compensate for any shifts generated in the images. ECG gating can be used for cross-cycle accumulation.

[0037] In this embodiment of the invention, the PMUT array employs row-column addressing wiring to achieve independent control of each array element. For ease of understanding, specifically, a differential drive strategy is used, utilizing the positive and negative voltage difference to drive each array element individually. For example, in a 4x4 array, driving the element in the 2nd row and 3rd column individually can involve inputting a +50V high-voltage pulse to the corresponding row line in the 2nd row and simultaneously inputting a -50V high-voltage pulse to the 3rd column, resulting in a voltage difference of 100V, causing the thin film to vibrate violently and emit ultrasonic waves.

[0038] Step 202: Based on the initial ultrasound image set, obtain the signal-to-noise ratio distribution and emission amplitude matrix corresponding to each 4D intracardiac ultrasound image.

[0039] In this embodiment of the invention, for each 4D intracardiac ultrasound image: the signal-to-noise ratio (SNR) distribution is used to represent the SNR of each voxel in the 4D intracardiac ultrasound image, and each element in the emission amplitude matrix is ​​used to represent the signal release intensity of the PMUT element corresponding to that element when acquiring the 4D intracardiac ultrasound image. The SNR distribution can be a sequence recording the SNR and corresponding voxel coordinate information, and the SNR can be obtained by comparing the signal amplitude reflected by tissue with the background noise of the system circuit. The SNR distribution can objectively quantify the real-time image quality of each voxel; a higher SNR reflects better imaging quality.

[0040] In this matrix, each element corresponds one-to-one with each element in the PMUT array, and the value of each element represents the signal release intensity of that element. Specifically, the matrix controls the signal release intensity by adjusting the amplitude (peak voltage) of the driving voltage across the electrodes of each PMUT element, or by controlling the number of driving pulses. Signal release intensity refers to sound pressure level and total acoustic energy; controlling the voltage amplitude or the number of pulses (the number of cycles of the driving pulses) directly changes the signal release intensity.

[0041] Step 203: Obtain the first selected area based on the signal-to-noise ratio distribution at the current moment.

[0042] In this embodiment of the invention, the first selected region is the area corresponding to the first voxel, which is a voxel in the current 4D intracardiac ultrasound image with a signal-to-noise ratio (SNR) less than a preset first threshold. Specifically, the first selected region can be represented by coordinates of a range where the SNR is lower than the first threshold, reflecting the region with low imaging quality at the current moment. Subsequently, in subsequent operations, signal intensity enhancement is performed on this region with low imaging quality.

[0043] Step 204: In the first selected area: Based on all signal-to-noise ratio distributions and all emission amplitude matrices corresponding to all 4D intracardiac ultrasound images at all historical moments, obtain the correlation matrix.

[0044] In this embodiment of the invention, the elements in the correlation matrix represent the degree of correlation between the change in signal release intensity and the change in signal-to-noise ratio (SNR) for each PMUT element. Specifically, taking any element in the correlation matrix as an example, this correlation degree mainly reflects the contribution of the signal enhancement of the corresponding element to the increase in SNR (imaging quality enhancement). This can be achieved through a row-column addressing mechanism, allowing independent control of the signal release intensity of elements at historical moments, thus facilitating the calculation of the correlation for each element. The degree of correlation can be measured using the Pearson correlation coefficient or the Spearman correlation coefficient to measure the overall trend. Preferably, the gradient of signal release intensity with respect to SNR can directly reflect the contribution of signal enhancement to the change in SNR.

[0045] Step 205: Optimize the initial emission amplitude matrix at the current moment based on the correlation matrix to optimize the PMUT array and obtain the optimized PMUT array.

[0046] In this embodiment of the invention, the optimized PMUT array is used for 4D intracardiac ultrasound imaging. In ultrasound imaging, most diagnostic ultrasounds satisfy linear acoustics, where the sound pressure at a point in space is a linear superposition of the contributions from each array element. Therefore, optimizing the PMUT array to increase the signal release intensity of the array elements can effectively increase the sound pressure in the first selected region (low signal-to-noise ratio region), thereby enhancing the signal-to-noise ratio of the imaging result. Specifically, the process of optimizing the initial emission amplitude matrix at the current moment can involve adjusting the signal release intensity of the array elements according to their correlation. This adjustment must be based on ensuring that the total increase in power consumption does not exceed a safe range and that no area of ​​the image is overexposed, so that the low signal-to-noise ratio region receives a slightly stronger signal echo, thereby improving the imaging quality of the corresponding region when subsequently acquiring 4D ultrasound images.

[0047] As can be seen, this embodiment of the invention measures the correlation between the signal-to-noise ratio distribution and the emission amplitude matrix changes at historical moments during 4D intracardiac ultrasound imaging, thereby obtaining the influence of each array element signal on the low signal-to-noise ratio region (first selected region) at the current moment. Based on this correlation, the current PMUT array and emission amplitude matrix are optimized, thereby enhancing the signal strength of the optimized PMUT array in the low signal-to-noise ratio region, and thus improving the imaging accuracy.

[0048] Since this process requires independent control of each array element, and considering the limited space of the cardiac chamber makes it difficult to accommodate the dimensions of a high-density PMUT array for independent wiring of each element, this embodiment of the invention uses row-column addressing to simplify the wiring of the PMUT array. Simultaneously, row-column addressing is a necessary condition for the application of PMUT arrays in confined spaces. Therefore, the introduction of this mechanism is both the foundation for further implementation in this embodiment and a rigid requirement for the application of PMUT arrays in 4DICE. This relationship allows the row-column addressing mechanism to not only simplify wiring but also further highlight the special attribute of being able to independently control each array element in this environment.

[0049] In an optional embodiment, the aforementioned increase in signal release intensity is engineering and physically constrained. The PMUT itself has a relatively low output sound pressure level, and studies have shown that after increasing the driving voltage, the response of the transmitted signal first increases and then decreases, indicating an optimal value. Therefore, to avoid the negative optimization effect caused by increasing the signal, the aforementioned optimization of the initial transmission amplitude matrix at the current moment based on the correlation matrix to obtain an optimized PMUT array may include: A first perturbation matrix is ​​constructed based on the correlation matrix and preset perturbation parameters. The sum of all elements in the first perturbation matrix is ​​equal to the preset perturbation parameters. The perturbation parameters are used to adjust the signal release intensity of the PMUT array elements. The dimension of the first perturbation matrix is ​​the same as that of the initial transmission amplitude matrix. The initial emission amplitude matrix is ​​perturbed and updated according to the first perturbation matrix to obtain the transition emission amplitude matrix; the transition ultrasound image and the corresponding signal-to-noise ratio distribution are obtained according to the transition emission amplitude matrix; the signal-to-noise ratio level increment corresponding to the transition ultrasound image is obtained, the signal-to-noise ratio level is the total signal-to-noise ratio of all voxels in the transition ultrasound image corresponding to the first selected area, and the signal-to-noise ratio level increment is the increase in the signal-to-noise ratio level between the current perturbation update round and the previous round; When it is determined that the signal-to-noise ratio increment corresponding to the transition ultrasound image is greater than the preset second threshold, multiple rounds of perturbation update operations are performed on the transition emission amplitude matrix based on the first perturbation matrix until it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix is ​​less than the preset second threshold. Then, the transition emission amplitude matrix obtained from the previous round of perturbation update operation is output to optimize the PMUT array and obtain the optimized PMUT array.

[0050] In this optional embodiment, the purpose of the perturbation parameters, perturbation matrix, and perturbation update is to iteratively fine-tune each array element based on the transmission amplitude matrix, and to select the fine-tuning direction and rhythm of the array elements using a judgment feedback mechanism. The first perturbation matrix has the same dimension as the transmission amplitude matrix. The first perturbation matrix is ​​constructed based on the correlation matrix and the perturbation parameters. Specifically, the correlation matrix is ​​regarded as a weight parameter, and the perturbation parameters are assigned a perturbation sub-parameter to each array element in the transmission amplitude matrix based on the weight parameter. The perturbation sub-parameter is used to adjust the signal release intensity of the corresponding PMUT array element. Optionally, a small amount (negative correlation) is added or decreased in the driving voltage of the corresponding array element to adjust the signal release intensity. The purpose of this step is to facilitate subsequent judgment on whether the direction of optimization update meets the requirements for improving imaging quality, and then further adjustments can be made based on this.

[0051] In this optional embodiment, in the perturbation update operation, the signal-to-noise ratio (SNR) level increment can be the total SNR change of the corresponding region image (voxel), or the increase in SNR level between the current perturbation update round and the previous round. The first round is the increase in SNR level of the transition emission amplitude matrix relative to the initial emission amplitude matrix. The multi-round perturbation update operation here is triggered by subsequent threshold judgment.

[0052] In this optional embodiment, if the signal-to-noise ratio (SNR) increment is greater than a threshold, it means the perturbation update direction is correct. A low SNR increment (below the second threshold) is considered insufficient optimization, while a negative SNR increment indicates negative optimization. When the SNR increment meets the above requirements, the emission amplitude matrix continues to be updated based on this relativity matrix until the SNR increment is determined to be less than the threshold, meaning further optimization updates are difficult to meet. At this point, the perturbation update stops, and the previous round's results are output.

[0053] It should be noted that when the signal-to-noise ratio increment is less than the second threshold, the current initial transmission amplitude matrix does not need to be optimized, which will be further explained in subsequent embodiments.

[0054] As can be seen, this optional embodiment refines the optimization operation through perturbation updates, further improving the optimization process of the PMUT array into a directional automatic optimization method. Optionally, small perturbations will not have a significant impact on the original 4D ICE process and can provide a suitable optimization direction for easy optimization. Moreover, the optimization process stops when the benefits do not meet the requirements, and the optimized array (emission amplitude matrix) is output for imaging. Since the first perturbation matrix is ​​obtained based on the correlation of all image data before the current time, when the perturbation update based on the first perturbation matrix fails to meet the requirement (the signal-to-noise ratio increment of the transition ultrasound image is less than the second threshold), it indicates that the result of this perturbation update is a good comparative reference for completing the correlation information, and the correlation matrix can be further updated accordingly. In another optional embodiment, the 4D intracardiac ultrasound imaging method based on row and column addressing PMUT may further include: When it is determined that the signal-to-noise ratio increment corresponding to the transition ultrasound image is less than the preset second threshold, in the first selected area: according to the signal-to-noise ratio distribution corresponding to the transition ultrasound image and the transition emission amplitude matrix, the correlation matrix is ​​updated in a transitional manner, and a second perturbation matrix is ​​constructed based on the result of the transitional update. The transition emission amplitude matrix is ​​then perturbed and updated based on the second perturbation matrix to obtain the updated transition emission amplitude matrix. When it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix is ​​greater than the preset second threshold, multiple rounds of perturbation update operations are performed on the updated transition emission amplitude matrix based on the second perturbation matrix until it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix after perturbation update is less than the preset second threshold. Then, the transition emission amplitude matrix obtained from the previous round of perturbation update operation is output to optimize the PMUT array and obtain the optimized PMUT array.

[0055] In this optional embodiment, the operation of further updating the correlation matrix can be set as a regular operation where the signal-to-noise ratio increment is lower than the second threshold, depending on the actual task requirements. This optional embodiment describes one round of operation only as an example, and more rounds can be executed in practice as needed.

[0056] In this optional embodiment, perturbation update involves adding corresponding elements from the perturbation matrix to the transmission amplitude matrix to update the transmission amplitude matrix; the difference with transition update is that the updated object is the correlation matrix. The transition update method can be to add the transition ultrasound image to the initial ultrasound image set and then recalculate the correlation between the signal-to-noise ratio change and the corresponding array element signal release intensity change in the first selected region within these images; alternatively, it can be to subtract the initial transmission amplitude matrix from the transition transmission amplitude matrix to obtain the signal release intensity change corresponding to each array element, obtain the signal-to-noise ratio change corresponding to the first selected region from the current 4D ICE image to the transition ultrasound image, calculate the correlation between these two changes, and then perform a weighted correction on the original correlation matrix. Specifically, the correlation corresponding to each array element can be directly weighted and summed, and a larger weight can be assigned to the new correlation as the dominant factor; this optional embodiment does not limit this approach.

[0057] As can be seen, this optional embodiment updates the correlation matrix, further designs for situations where the perturbation update cannot meet the requirements, and makes full use of the characteristic that the transition emission amplitude matrix cannot achieve the expected effect to update the correlation matrix in a targeted manner, thereby further improving the accuracy and completeness of the PMUT array optimization method, and thus improving the imaging accuracy of the optimized PMUT array when acquiring 4D intracardiac ultrasound images.

[0058] In yet another optional embodiment, after perturbing and updating the transition emission amplitude matrix according to the second perturbation matrix to obtain the updated transition emission amplitude matrix, the 4D intracardiac ultrasound imaging method based on row and column addressing PMUT may further include: When it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix is ​​less than the preset second threshold, the 4D intracardiac ultrasound image to be processed is obtained according to the PMUT array in the initial state. Based on the 4D intracardiac ultrasound image to be processed, its corresponding signal-to-noise ratio distribution is obtained; based on the signal-to-noise ratio distribution, the attention distribution corresponding to each voxel in the 4D intracardiac ultrasound image to be processed is obtained; based on the preset image enhancement model, the 4D intracardiac ultrasound image to be processed is enhanced according to the attention distribution to obtain the enhanced 4D intracardiac ultrasound image, which is used as the final 4D intracardiac ultrasound imaging result.

[0059] In this optional embodiment, the image enhancement model can be the EdgeSRIE model, which suppresses speckle noise through an unsupervised denoising branch and restores blurred edges and details through a self-supervised deblurring branch. This can improve the contrast-to-noise ratio (CNR) of the image by nearly 60%, thereby improving the problem of low imaging accuracy in low signal-to-noise ratio areas. Furthermore, this model can achieve an inference speed of 64.1 frames per second on the target hardware, and has only 17,700 model parameters, resulting in low computational cost and allowing deployment on low-power on-chip systems. The image enhancement model can also be MWCFNet (Multi-Scale Wavelet Cross-Attention Fusion Network), which can suppress noise while improving resolution; this optional embodiment is not limited to this.

[0060] In this optional embodiment, the attention distribution, or signal-to-noise ratio-guided attention mechanism introduced into the image enhancement model, corresponds to the signal-to-noise ratio (SNR) of each voxel. The lower the SNR, the higher the attention the image enhancement model pays to that voxel, guiding the model to prioritize processing regions with low SNR. For example, different processing strategies are dynamically adopted based on the SNR level. In low SNR regions, non-local or long-range global information is used to aid in recovery. In high SNR regions, local or short-range neighborhood information is relied upon for processing.

[0061] In summary, this optional embodiment first adapts the PMUT array to the confined space of the cardiac cavity through row-column addressing wiring, and fully utilizes the independent control mechanism of the array elements through row-column addressing. Subsequently, based on this independent control, the correlation between the array element signal release intensity and the signal-to-noise ratio (SNR) change is introduced as a perturbation weight. A small amount can be added to the array element signal release intensity, and based on the feedback from the imaging SNR, it is determined whether to continue perturbation updates, thereby achieving a blind optimization effect. Furthermore, when the SNR condition is difficult to meet, the method of prioritizing the update of the correlation matrix is ​​selected. If the expected effect is still not achieved, image enhancement is performed on the first selected region with low SNR to improve the imaging quality of the first selected region. From an image perspective, the SNR is used as a safety measure when the above methods fail to achieve the expected SNR. Ultimately, this improves the imaging accuracy of 4D ICE.

[0062] Since the core challenge of ultrasound image enhancement is balancing the suppression of speckle noise with the preservation of detail information such as edges and textures, in another optional embodiment, this contradiction is addressed through a multi-task loss function. This 4D intracardiac ultrasound imaging method based on row-column addressing PMUT may further include: Construct a first loss function and a second loss function. The first loss function is used to measure the noise intensity of the image, and the second loss function is used to measure the edge gradient of the image. The training process of the image enhancement model is constrained by the first loss function and the second loss function to obtain a trained image enhancement model. The trained image enhancement model is used to enhance the 4D intracardiac ultrasound image to be processed according to the attention distribution.

[0063] In this optional embodiment, the first loss function focuses on reducing noise intensity, making the image cleaner in smooth areas, while the second loss function focuses on preserving edge gradients, penalizing edge blurring, and ensuring structural clarity. The edge gradient can be the location and intensity of drastic changes in pixel brightness or grayscale within the image. Specifically, the calculation method of the first loss function can include:

[0064] In the above formula, I represents the enhanced image after training. This formula encourages smoothing by penalizing the differences between adjacent pixels, thereby suppressing noise and preserving edges well. The loss is calculated as the sum of the absolute values ​​of pixel differences in the horizontal and vertical directions. The first loss function can also use a high-quality, clean image G as a reference, employing mean squared error (MSE) or mean absolute error (L1Loss), which can directly compare the differences between the enhanced image I and the reference image G.

[0065] Furthermore, the calculation method for the second loss function can include:

[0066] In the above formula, For gradient operators (such as the Sobel operator). This represents the gradient magnitude of the enhanced image; this loss minimizes the difference between the gradient magnitudes of the two images.

[0067] In yet another optional embodiment, the 4D intracardiac ultrasound imaging method based on row and column addressing PMUT may further include: Spatial clustering is performed on the first voxel to obtain the first voxel cluster; the location range corresponding to the first voxel cluster whose total number of all voxels is greater than the preset third threshold is determined as the undetermined selection area; the proportion of low quality voxels is obtained, which is the proportion of all voxels in the undetermined selection area relative to all voxels in the current 4D intracardiac ultrasound image. When the proportion of low-quality voxels is determined to be less than the preset fourth threshold, 4D intracardiac ultrasound imaging results are obtained based on the PMUT array. When it is determined that the proportion of low-quality voxels is greater than the preset fourth threshold, the undetermined selection area is determined as the first selection area.

[0068] In yet another optional embodiment, the specific implementation of spatial clustering of the first voxel to obtain the first voxel cluster may include: Determine the neighboring voxels corresponding to the target first voxel. The target first voxel can be any first voxel, and the spatial distance between the neighboring voxels and the target first voxel is less than a preset adjacency radius. Obtain the signal-to-noise ratio difference, which is the difference between the signal-to-noise ratio corresponding to the target first voxel and the signal-to-noise ratio corresponding to the neighboring voxel. Neighboring voxels with a signal-to-noise ratio difference less than a preset second threshold are designated as second voxels; based on all first voxels and all second voxels, a first voxel cluster is obtained.

[0069] Example 2 Please see Figure 3 , Figure 3 This is a schematic diagram of a 4D intracardiac ultrasound imaging device based on row and column addressing PMUT disclosed in an embodiment of the present invention. Figure 3 As shown, the 4D intracardiac ultrasound imaging device based on row and column addressing PMUT may include: PMUT array 301 is used to acquire an initial ultrasound image set, which includes all 4D intracardiac ultrasound images at the current and historical times. The PMUT array uses row and column addressing wiring to achieve independent control of each array element. The preprocessing module 302 is used to obtain the signal-to-noise ratio distribution and emission amplitude matrix corresponding to each 4D intracardiac ultrasound image based on the initial ultrasound image set. The core processing module 303 is used to obtain a first selected region based on the signal-to-noise ratio distribution at the current moment. The first selected region is the region corresponding to a first voxel. The first voxel is the voxel in the 4D intracardiac ultrasound image at the current moment whose signal-to-noise ratio is less than a preset first threshold. In the first selected region, a correlation matrix is ​​obtained based on all signal-to-noise ratio distributions and all emission amplitude matrices corresponding to all 4D intracardiac ultrasound images at all historical moments. The elements in the correlation matrix are used to represent the degree of correlation between the change in signal release intensity and the change in signal-to-noise ratio corresponding to each PMUT array element. The array element adjustment module 304 is used to optimize the initial emission amplitude matrix at the current moment based on the correlation matrix, so as to optimize the PMUT array and obtain the optimized PMUT array, which is used for 4D intracardiac ultrasound imaging.

[0070] In an optional embodiment, the array element adjustment module optimizes the initial emission amplitude matrix at the current moment based on the correlation matrix to obtain the optimized PMUT array's specific operation mode, which may include: A first perturbation matrix is ​​constructed based on the correlation matrix and preset perturbation parameters. The sum of all elements in the first perturbation matrix is ​​equal to the preset perturbation parameters. The perturbation parameters are used to adjust the signal release intensity of the PMUT array elements. The dimension of the first perturbation matrix is ​​the same as that of the initial transmission amplitude matrix. The initial emission amplitude matrix is ​​perturbed and updated according to the first perturbation matrix to obtain the transition emission amplitude matrix; the transition ultrasound image and the corresponding signal-to-noise ratio distribution are obtained according to the transition emission amplitude matrix; the signal-to-noise ratio level increment corresponding to the transition ultrasound image is obtained, the signal-to-noise ratio level is the total signal-to-noise ratio of all voxels in the transition ultrasound image corresponding to the first selected area, and the signal-to-noise ratio level increment is the increase in the signal-to-noise ratio level between the current perturbation update round and the previous round; When it is determined that the signal-to-noise ratio increment corresponding to the transition ultrasound image is greater than the preset second threshold, multiple rounds of perturbation update operations are performed on the transition emission amplitude matrix based on the first perturbation matrix until it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix is ​​less than the preset second threshold. Then, the transition emission amplitude matrix obtained from the previous round of perturbation update operation is output to optimize the PMUT array and obtain the optimized PMUT array.

[0071] In another alternative embodiment, the device may further include: The first judgment module is used to, when it is determined that the signal-to-noise ratio level increment corresponding to the transition ultrasound image is less than a preset second threshold, in the first selected area: according to the signal-to-noise ratio distribution corresponding to the transition ultrasound image and the transition emission amplitude matrix, perform transition update on the correlation matrix, construct a second perturbation matrix according to the result of the transition update, and perform perturbation update on the transition emission amplitude matrix according to the second perturbation matrix to obtain the updated transition emission amplitude matrix; The second judgment module is used to perform multiple rounds of perturbation update operations on the updated transition emission amplitude matrix based on the second perturbation matrix when it is determined that the signal-to-noise ratio level increment corresponding to the updated transition emission amplitude matrix is ​​greater than the preset second threshold. This continues until it is determined that the signal-to-noise ratio level increment corresponding to the updated transition emission amplitude matrix after the perturbation update is less than the preset second threshold. Then, the transition emission amplitude matrix obtained from the previous round of perturbation update operation is output to optimize the PMUT array and obtain the optimized PMUT array.

[0072] In yet another optional embodiment, the first determination module is further configured to, after perturbing and updating the transition transmission amplitude matrix according to the second perturbation matrix to obtain the updated transition transmission amplitude matrix: When it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix is ​​less than the preset second threshold, the 4D intracardiac ultrasound image to be processed is obtained according to the PMUT array in the initial state. Based on the 4D intracardiac ultrasound image to be processed, its corresponding signal-to-noise ratio distribution is obtained; based on the signal-to-noise ratio distribution, the attention distribution corresponding to each voxel in the 4D intracardiac ultrasound image to be processed is obtained; based on the preset image enhancement model, the 4D intracardiac ultrasound image to be processed is enhanced according to the attention distribution to obtain the enhanced 4D intracardiac ultrasound image, which is used as the final 4D intracardiac ultrasound imaging result.

[0073] In yet another alternative embodiment, the device may further include: The model training module is used to construct a first loss function and a second loss function. The first loss function is used to measure the noise intensity of the image, and the second loss function is used to measure the edge gradient of the image. The training process of the image enhancement model is constrained by the first loss function and the second loss function to obtain the trained image enhancement model. The trained image enhancement model is used to enhance the 4D intracardiac ultrasound image to be processed according to the attention distribution.

[0074] In yet another alternative embodiment, the device may further include: The spatial clustering module is used to perform spatial clustering on the first voxel to obtain the first voxel cluster; the location range corresponding to the first voxel cluster whose total number of all voxels is greater than a preset third threshold is determined as the undetermined selection area; the proportion of low-quality voxels is obtained, which is the proportion of all voxels in the undetermined selection area relative to all voxels in the 4D intracardiac ultrasound image at the current time; when it is determined that the proportion of low-quality voxels is less than a preset fourth threshold, the 4D intracardiac ultrasound imaging result is obtained based on the PMUT array; when it is determined that the proportion of low-quality voxels is greater than the preset fourth threshold, the undetermined selection area is determined as the first selection area.

[0075] In another optional embodiment, the specific operation of the spatial clustering module device to spatially cluster the first voxels to obtain the first voxel clusters may include: Determine the neighboring voxels corresponding to the target first voxel. The target first voxel can be any first voxel, and the spatial distance between the neighboring voxels and the target first voxel is less than a preset adjacency radius. Obtain the signal-to-noise ratio difference, which is the difference between the signal-to-noise ratio corresponding to the target first voxel and the signal-to-noise ratio corresponding to the neighboring voxel. Neighboring voxels with a signal-to-noise ratio difference less than a preset second threshold are designated as second voxels; based on all first voxels and all second voxels, a first voxel cluster is obtained.

[0076] Example 3 Please see Figure 4 , Figure 4 This is a schematic diagram of the structure of a device including a memory and a processor, as disclosed in an embodiment of the present invention. Figure 4 As shown, the device including memory and processor may include: Memory 401 storing executable program code; Processor 402 coupled to memory 401; The processor 402 calls the executable program code stored in the memory 401 to execute some or all of the steps in any of the row and column addressing PMUT-based 4D intracardiac ultrasound imaging methods in Embodiment 1 of the present invention.

[0077] Example 4 This invention discloses a computer storage medium storing computer instructions. When these computer instructions are invoked, they are used to execute some or all of the steps in any of the row-column addressing PMUT-based 4D intracardiac ultrasound imaging methods disclosed in Embodiment 1 of this invention.

[0078] The device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0079] Through the detailed description of the above embodiments, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-Erasable Programmable Read-Only Memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0080] Finally, it should be noted that the above embodiments are merely preferred embodiments of the present invention and are only used to illustrate the technical solutions of the present invention, not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A 4D intracardiac ultrasound imaging device based on row and column addressing PMUT, characterized in that, The device includes: A PMUT array is used to acquire an initial set of ultrasound images, which includes all 4D intracardiac ultrasound images at the current and historical times. The PMUT array uses row and column addressing wiring to achieve independent control of each array element. The preprocessing module is used to obtain the signal-to-noise ratio distribution and emission amplitude matrix corresponding to each 4D intracardiac ultrasound image based on the initial ultrasound image set. The core processing module is used to obtain a first selected region based on the signal-to-noise ratio distribution at the current moment. The first selected region is the region corresponding to a first voxel, which is a voxel in the 4D intracardiac ultrasound image at the current moment whose signal-to-noise ratio is less than a preset first threshold. In the first selected region, a correlation matrix is ​​obtained based on all signal-to-noise ratio distributions and all emission amplitude matrices corresponding to all 4D intracardiac ultrasound images at historical moments. The elements in the correlation matrix are used to represent the degree of correlation between the change in signal release intensity and the change in signal-to-noise ratio corresponding to each PMUT element. The array element adjustment module is used to optimize the initial emission amplitude matrix at the current moment according to the correlation matrix, so as to optimize the PMUT array and obtain the optimized PMUT array, which is used for 4D intracardiac ultrasound imaging.

2. A 4D intracardiac ultrasound imaging method based on row-column addressing PMUT, characterized in that, The method includes: An initial ultrasound image set is acquired, which includes all 4D intracardiac ultrasound images at the current and historical times. The PMUT array includes multiple PMUT elements, and the PMUT array uses row and column addressing wiring to achieve independent control of each element. Based on the initial set of ultrasound images, the signal-to-noise ratio distribution and emission amplitude matrix corresponding to each 4D intracardiac ultrasound image are obtained. For each 4D intracardiac ultrasound image: the signal-to-noise ratio distribution is used to represent the signal-to-noise ratio of each voxel on the 4D intracardiac ultrasound image, and each element in the emission amplitude matrix is ​​used to represent the signal release intensity of the PMUT array element corresponding to that element when the 4D intracardiac ultrasound image is acquired. Based on the signal-to-noise ratio (SNR) distribution at the current moment, a first selected region is obtained. The first selected region is the region corresponding to the first voxel, which is the voxel in the 4D intracardiac ultrasound image at the current moment whose SNR is less than a preset first threshold. In the first selected region: based on all SNR distributions and all emission amplitude matrices corresponding to all 4D intracardiac ultrasound images at all historical moments, a correlation matrix is ​​obtained. The elements in the correlation matrix are used to represent the degree of correlation between the change in signal release intensity and the change in SNR corresponding to each PMUT element. Based on the correlation matrix, the initial emission amplitude matrix at the current moment is optimized to optimize the PMUT array, resulting in an optimized PMUT array used for 4D intracardiac ultrasound imaging.

3. The 4D intracardiac ultrasound imaging method based on row and column addressing PMUT according to claim 2, characterized in that, The step of optimizing the initial emission amplitude matrix at the current moment based on the correlation matrix to obtain the optimized PMUT array includes: A first perturbation matrix is ​​constructed based on the correlation matrix and preset perturbation parameters. The sum of all elements in the first perturbation matrix is ​​equal to the preset perturbation parameters. The perturbation parameters are used to adjust the signal release intensity of the PMUT array elements. The dimension of the first perturbation matrix is ​​the same as that of the initial transmission amplitude matrix. The initial emission amplitude matrix is ​​perturbed and updated according to the first perturbation matrix to obtain the transition emission amplitude matrix; the signal-to-noise ratio (SNR) distribution of the transition ultrasound image and the transition ultrasound image is obtained according to the transition emission amplitude matrix; the SNR level increment corresponding to the transition ultrasound image is obtained, where the SNR level is the total SNR of all voxels in the transition ultrasound image corresponding to the first selected region, and the SNR level increment is the increase in the SNR level between the current perturbation update round and the previous round; When it is determined that the signal-to-noise ratio increment corresponding to the transition ultrasound image is greater than a preset second threshold, multiple rounds of perturbation update operations are performed on the transition emission amplitude matrix based on the first perturbation matrix until it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix is ​​less than the preset second threshold. Then, the transition emission amplitude matrix obtained from the previous round of perturbation update operation is output to optimize the PMUT array and obtain the optimized PMUT array.

4. The 4D intracardiac ultrasound imaging method based on row and column addressing PMUT according to claim 3, characterized in that, The method further includes: When it is determined that the signal-to-noise ratio increment corresponding to the transition ultrasound image is less than a preset second threshold, in the first selected area: according to the signal-to-noise ratio distribution corresponding to the transition ultrasound image and the transition emission amplitude matrix, the correlation matrix is ​​updated to construct a second perturbation matrix based on the result of the transition update, and the transition emission amplitude matrix is ​​perturbed and updated based on the second perturbation matrix to obtain the updated transition emission amplitude matrix; When it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix is ​​greater than the preset second threshold, multiple rounds of perturbation update operations are performed on the updated transition emission amplitude matrix based on the second perturbation matrix until it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix after the perturbation update is less than the preset second threshold. Then, the transition emission amplitude matrix obtained in the previous round of perturbation update operation is output to optimize the PMUT array and obtain the optimized PMUT array.

5. The 4D intracardiac ultrasound imaging method based on row and column addressing PMUT according to claim 4, characterized in that, After perturbating and updating the transition emission amplitude matrix according to the second perturbation matrix to obtain the updated transition emission amplitude matrix, the method further includes: When it is determined that the signal-to-noise ratio increment corresponding to the updated transition emission amplitude matrix is ​​less than the preset second threshold, the 4D intracardiac ultrasound image to be processed is obtained according to the PMUT array in the initial state. Based on the 4D intracardiac ultrasound image to be processed, its corresponding signal-to-noise ratio distribution is obtained; based on the signal-to-noise ratio distribution, the attention distribution corresponding to each voxel in the 4D intracardiac ultrasound image to be processed is obtained; based on a preset image enhancement model, the 4D intracardiac ultrasound image to be processed is enhanced according to the attention distribution to obtain an enhanced 4D intracardiac ultrasound image, which is used as the final 4D intracardiac ultrasound imaging result.

6. The 4D intracardiac ultrasound imaging method based on row and column addressing PMUT according to claim 5, characterized in that, The method further includes: Construct a first loss function and a second loss function. The first loss function is used to measure the noise intensity of the image, and the second loss function is used to measure the edge gradient of the image. The training process of the image enhancement model is constrained by the first loss function and the second loss function to obtain a trained image enhancement model. The trained image enhancement model is used to enhance the 4D intracardiac ultrasound image to be processed according to the attention distribution.

7. The 4D intracardiac ultrasound imaging method based on row and column addressing PMUT according to claim 5, characterized in that, The method further includes: Spatial clustering is performed on the first voxel to obtain the first voxel cluster; the location range corresponding to the first voxel cluster whose total number of all voxels is greater than a preset third threshold is determined as the undetermined selection area; the proportion of low quality voxels is obtained, wherein the proportion of low quality voxels is the proportion of all voxels in the undetermined selection area relative to all voxels in the current 4D intracardiac ultrasound image. When it is determined that the proportion of low-quality voxels is less than a preset fourth threshold, 4D intracardiac ultrasound imaging results are obtained based on the PMUT array. When it is determined that the proportion of low-quality voxels is greater than the preset fourth threshold, the undetermined selection area is determined as the first selection area.

8. The 4D intracardiac ultrasound imaging method based on row and column addressing PMUT according to claim 6, characterized in that, The step of spatially clustering the first voxel to obtain the first voxel cluster includes: Determine the neighboring voxels corresponding to the target first voxel, wherein the target first voxel is any first voxel, and the spatial distance between the neighboring voxel and the target first voxel is less than a preset adjacency radius; obtain the signal-to-noise ratio difference, wherein the signal-to-noise ratio difference is the difference between the signal-to-noise ratio corresponding to the target first voxel and the signal-to-noise ratio corresponding to the neighboring voxel; Neighboring voxels with a signal-to-noise ratio difference less than a preset second threshold are designated as second voxels; based on all first voxels and all second voxels, a first voxel cluster is obtained.

9. An apparatus comprising a memory and a processor, characterized in that, The device includes: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to perform some or all of the steps in the 4D intracardiac ultrasound imaging method based on row and column addressing PMUT as described in any one of claims 2-8.

10. A computer storage medium, characterized in that, The computer storage medium stores computer instructions, which, when invoked by a processor, execute some or all of the steps in the 4D intracardiac ultrasound imaging method based on row and column addressing PMUT as described in any one of claims 2-8.

Images