An ultrasound contrast image enhancement method for nasojejunal tube placement

Abstract

The present application relates to the technical field of image processing, in view of the problem that the existing MSR algorithm influences the image enhancement effect when the intestinal cavity environment changes, the present application provides an ultrasound contrast image enhancement method for nasoenteral tube placement, the method comprises the following steps: collecting and preprocessing the ultrasound contrast image sequence of the nasoenteral tube placement area, highlighting the microbubble signal features in the ultrasound contrast image sequence, establishing the quantitative correlation between the intestinal cavity physicochemical environment and the microbubble signal features, obtaining the expected attenuation degree of the microbubble signal features under the current environment, dynamically adjusting the core parameters of the MSR algorithm according to the expected attenuation degree, constructing the MSR image processing model, inputting the preprocessed ultrasound contrast image sequence into the MSR image processing model for enhancement processing, highlighting the nasoenteral tube structure, and realizing the accurate positioning of the distal end of the nasoenteral tube. The present application can significantly improve the contrast of the nasoenteral tube structure and the background, and can maintain the stability of the image enhancement effect under different intestinal environments.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to an ultrasound contrast imaging method for nasoenteric tube insertion. Background Technology

[0002] Nasoenteric tubes are an enteral nutrition route used to support patients with bowel dysfunction or those unable to eat orally for other reasons. They are commonly used for patients with nasopharyngeal tumors, undergoing radiotherapy, oropharyngeal tumors, or esophageal strictures. It can be any nasogastric feeding tube that passes through the esophagus and pylorus; if the distal end is in the duodenum, it is called a nasoduodenal tube, and if it is in the jejunum, it is called a nasojejunal tube. Clinically, nasoenteric tube placement is a core procedure for enteral nutrition support in critically ill patients. The accurate determination of the distal position of the tube directly determines the effectiveness of nutrient delivery and patient safety, and is a crucial step in ensuring the quality of clinical treatment.

[0003] Currently, contrast-enhanced ultrasound, as an advanced application of conventional ultrasound, significantly improves the imaging capabilities of the nasoenteric tube and its surrounding structures through intravenous injection of microbubble contrast agents, providing an important visual basis for tube end localization. However, even after contrast enhancement, ultrasound images are often affected by noise, low contrast, and local occlusion, impacting the distinction between the tube and surrounding tissues. Therefore, existing techniques typically rely on multi-scale retinex (MSR) algorithms to preprocess chemically enhanced ultrasound images, thereby improving structural clarity. Existing MSR algorithms separate the illumination and reflection components in the image through multi-scale Gaussian filtering, utilizing the detailed information contained in the reflection component to enhance the contrast between the nasoenteric tube structure and surrounding tissues, thus improving the clarity of the tube edges and the overall visual quality of the image, providing a more reliable image basis for subsequent tube end localization.

[0004] The core of existing MSR algorithms lies in removing the influence of illumination through multi-scale filtering estimation, thereby enhancing the details and contrast in images. However, when the physicochemical environment of the intestinal lumen changes significantly, it directly affects the stability and dispersion uniformity of gas microbubbles in the contrast agent. For example, unevenly viscous intestinal fluid (such as in patients with inflammatory bowel disease) can affect the damping of microbubble movement, weakening its ability to generate nonlinear harmonics; disordered intestinal peristalsis rhythm (such as in neuroendocrine patients) may affect the stability of the microbubble phospholipid shell, causing microbubbles to rupture or merge prematurely before reaching the target area. These environmentally induced microbubble anomalies can lead to uneven distribution of microbubble signal intensity in the image, resulting in distorted illumination estimation. This makes it difficult to effectively distinguish nasoenteric tube details from background noise in the reflection component, thus interfering with the discriminative ability of the MSR algorithm in illumination correction and detail enhancement, and reducing the image enhancement effect on nasoenteric tube structures.

[0005] In view of this, it is particularly important to study an image enhancement method that can quantitatively perceive the intestinal environment and adaptively adjust the core parameters of the MSR algorithm during the insertion process of the nasoenteric tube, so as to stably improve the recognition and positioning accuracy of the nasoenteric tube boundary in the complex and ever-changing intestinal environment. Summary of the Invention

[0006] To address the shortcomings of the existing technologies, this invention aims to provide an ultrasound contrast imaging enhancement method for nasoenteric tube insertion. By establishing a quantitative correlation between the physicochemical environment of the intestinal lumen and the characteristics of microbubble signals, the expected attenuation level of the microbubble signals is inferred using quantitative intestinal peristalsis characteristics and intestinal fluid viscosity characteristics. Based on this, the core parameters of the MSR algorithm are dynamically adjusted to achieve optimal adaptation between the MSR algorithm and the actual imaging environment, thereby effectively solving the problems in the background technology section.

[0007] To achieve the above-mentioned technical objectives, the present invention adopts the following technical solution:

[0008] The first aspect of the technical solution disclosed in this invention is: a method for enhancing ultrasound contrast images for nasoenteric tube insertion, comprising the following steps:

[0009] Acquired and preprocessed ultrasound contrast imaging sequences of the nasoenteric tube insertion area to highlight the microbubble signal characteristics in the ultrasound contrast imaging sequences;

[0010] Establish a quantitative correlation between the physicochemical environment of the intestinal lumen and the characteristics of microbubble signals, obtain the expected attenuation degree of microbubble signal characteristics under the current environment, dynamically adjust the core parameters of the MSR algorithm based on the expected attenuation degree, and construct an MSR image processing model.

[0011] The preprocessed ultrasound contrast imaging sequence is input into the MSR image processing model for enhancement, highlighting the nasoenteric tube structure. Based on the enhanced ultrasound contrast imaging sequence, the distal end of the nasoenteric tube is accurately located.

[0012] As a further preferred option, the steps of acquiring and preprocessing ultrasound contrast-enhanced image sequences of the nasoenteric tube insertion area include:

[0013] The real-time contrast imaging process, including key parts of the expected course of the nasoenteric tube, was captured using harmonic imaging mode, resulting in a sequence of ultrasound contrast images containing microbubble signals.

[0014] Noise suppression processing was performed on the ultrasound contrast imaging sequence;

[0015] A registration algorithm is used to perform motion compensation on the noise-suppressed ultrasound contrast imaging sequence, so that each frame in the ultrasound contrast imaging sequence is spatially aligned.

[0016] As a further preferred option, the steps for constructing the MSR image processing model include:

[0017] Based on the extraction of intestinal peristalsis features and intestinal fluid viscosity features from ultrasound contrast imaging sequences, an environmental perception assessment index was established.

[0018] Microbubble state prediction factors were constructed based on environmental perception and assessment indicators.

[0019] The core parameters of the MSR algorithm are dynamically adjusted using microbubble state predictors, and an MSR image processing model is established with the adjusted optimal core parameters.

[0020] As a further preferred approach, intestinal peristalsis features and intestinal fluid viscosity features are extracted based on ultrasound contrast imaging sequences to establish environmental perception assessment indicators, specifically including:

[0021] Extract the region of interest (ROI) from each frame of ultrasound contrast imaging image, identify the outline of the bright pixel clusters based on the ROI, and construct the bounding box using the outline of the bright pixel clusters.

[0022] The uniformity of intestinal fluid viscosity is obtained by quantifying the uniformity of the distribution of bright pixels within the bounding box and the uniformity of brightness within the cluster.

[0023] The stability of intestinal peristalsis can be obtained by quantifying the continuity of the direction of microbubble signal movement.

[0024] An environmental perception assessment index was established based on the uniformity of intestinal fluid viscosity and the stability of intestinal peristalsis, serving as a comprehensive indicator for assessing the physicochemical state of the intestine.

[0025] As a further preferred option, the step of constructing the minimum bounding rectangle of the highlight pixel cluster outline as the bounding box includes:

[0026] The K-means clustering algorithm is used to perform pixel clustering on grayscale images. The pixel clustering results include bright pixel clusters representing microbubble signals and low-bright pixel clusters representing tissue background. Regions of interest are extracted based on the pixel classification results.

[0027] The contour detection algorithm identifies discrete contour points of bright pixel clusters within the region of interest, and uses the smallest bounding rectangle of the discrete contour points of the bright pixel clusters as the bounding box.

[0028] As a further preferred option, a grayscale histogram of image grayscale level and pixel count is established for all pixels within the bounding box. Based on the grayscale histogram, all significant peak points are found. From multiple significant peak points, the two peak points with the greatest grayscale distance are selected as the characteristic peaks of high-brightness microbubble signal clusters and low-brightness tissue background, respectively.

[0029] As a further preferred option, microbubble state prediction factors are constructed based on environmental perception assessment indicators, including:

[0030] Based on environmental perception assessment indicators, we obtain the adverse frames and favorable frames of the intestinal physicochemical environment.

[0031] By combining bad frames and good frames to adjust the weights, a microbubble state prediction factor is constructed.

[0032] As a further optimization, an MSR image processing model is established using the adjusted optimal core parameters. The expression of the MSR image processing model is as follows:

[0033] ;

[0034] in, This represents the core parameters after dynamic adjustment. These represent the core parameters of the Gaussian filter for the corresponding time-series image. This represents the predictor of microbubble state.

[0035] As a further preferred embodiment, the preprocessed ultrasound contrast-enhanced image sequence is input into the MSR image processing model for enhancement processing to highlight the nasoenteric canal structure, including:

[0036] A logarithmic transformation is performed on the preprocessed ultrasound contrast image sequence to map the pixel gray values ​​to a more compact range;

[0037] The dynamically adjusted MSR image processing model is used to further perform multi-scale Gaussian filtering and reflection component calculation, followed by exponential transformation and grayscale stretching to output an enhanced ultrasound contrast image sequence.

[0038] The contour of the nasoenteric tube was extracted from the enhanced ultrasound imaging sequence, and the spatial localization of the distal end of the nasoenteric tube was achieved by combining the tube's movement trajectory.

[0039] The second aspect of the technical solution disclosed in this invention is: an ultrasound contrast imaging enhancement system for nasoenteric tube insertion, the system being used to implement the steps of the above-described method, the system comprising:

[0040] The acquisition module is used to acquire ultrasound contrast imaging sequences of the nasoenteric tube insertion area;

[0041] The image processing module is used to preprocess the ultrasound contrast imaging image sequence to highlight the microbubble signal characteristics in the ultrasound contrast imaging image sequence;

[0042] The signal feature change acquisition module establishes a quantitative correlation between the intestinal lumen physicochemical environment and microbubble signal features, and obtains the expected attenuation degree of microbubble signal features under the current environment;

[0043] The model building module dynamically adjusts the core parameters of the MSR algorithm based on the expected attenuation level to build the MSR image processing model;

[0044] The enhancement module is used to input the pre-processed ultrasound contrast image sequence into the MSR image processing model for enhancement processing to highlight the nasoenteric canal structure.

[0045] The positioning module achieves precise positioning of the distal end of the nasoenteric tube based on enhanced ultrasound contrast imaging sequences.

[0046] In addition, the present invention also discloses an electronic device, the electronic device including at least one processor and a memory communicatively connected to the processor; wherein the memory stores instructions executable by the processor, the instructions being executed by the processor to enable the processor to perform the steps of the above-described method.

[0047] Compared with the prior art, the beneficial effects of the present invention are reflected in:

[0048] 1. This invention extracts features such as the distribution density, grayscale separation degree, and centroid displacement vector of high-brightness microbubble clusters from ultrasound contrast imaging sequences to quantitatively evaluate the uniformity of intestinal fluid viscosity and the stability of intestinal peristalsis. It also constructs a comprehensive index for evaluating the state of the intestinal physicochemical environment and a microbubble state predictor, forming a quantitative correlation between the intestinal physicochemical environment and the attenuation characteristics of microbubble signals. This allows for dynamic perception of the current imaging environment and adaptive adjustment of the core parameters of the MSR algorithm. Compared to the traditional MSR algorithm which uses fixed parameters, the method of this invention can maintain the stability of image enhancement effects under different intestinal environments and significantly improve the contrast between the nasoenteric tube structure and the background.

[0049] 2. The method of the present invention can reduce the filtering scale and reduce the weight of large-scale components under harsh environmental conditions, enhance the ability to capture weak edge and texture information, and effectively highlight the boundary structure of the nasoenteric tube. Under good environmental conditions, it retains the global filtering characteristics, maintains the naturalness and anatomical authenticity of the image. This parameter adaptive adjustment mechanism helps to improve the accuracy and reliability of tube end positioning.

[0050] 3. The method of the present invention fully considers the influence of changes in the physicochemical environment in the intestine on microbubble signals, and is suitable for patient groups such as inflammatory bowel disease and neuroendocrine tumors that are prone to abnormal intestinal environment. Through environmental perception and parameter adaptive control, the robustness of ultrasound contrast imaging under complex physiological conditions is enhanced, and the applicability of ultrasound guidance technology in nasoenteric tube placement is expanded. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0052] Figure 1This is a flowchart of an ultrasound contrast imaging enhancement method for nasoenteric tube insertion according to the present invention.

[0053] Figure 2 When the ultrasound contrast agent of the present invention is injected into the nasoenteric lumen, the nasoenteric tube appears as a continuous, highly enhanced thin strip echo pattern after imaging, wherein the arrow indicates the nasoenteric tube;

[0054] Figure 3 This is a schematic diagram illustrating the generation of the region of interest and bounding box for extracting bright pixel clusters according to the present invention.

[0055] Figure 4 This is a schematic diagram illustrating the extraction of the centroid and displacement vector of the bright pixel clusters within the bounding box according to the present invention. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0057] To address the issue of tube tip localization during nasoenteric tube insertion, clinical practice currently relies on multi-scale image enhancement (MSR) algorithms to improve image structural recognition. However, when significant changes occur in the intestinal physicochemical environment of some patients, such as uneven intestinal fluid viscosity or disordered intestinal peristalsis rhythm, the stability and dispersion uniformity of gas microbubbles in the contrast agent can be directly affected, thus interfering with the discrimination ability of traditional MSR algorithms in illumination correction and detail enhancement. Based on this, the inventors have researched and proposed a nasoenteric tube ultrasound contrast imaging image enhancement method based on quantitative perception of the intestinal physicochemical environment and adaptive parameter adjustment. This method establishes attenuation characteristics that correlate the intestinal physicochemical environment with microbubble quantification, adjusts the core parameters of the MSR algorithm, and achieves optimal adaptation between the optimized MSR algorithm and the actual imaging environment, thereby improving the recognition and localization accuracy of the nasoenteric tube boundary in complex intestinal environments.

[0058] Example 1:

[0059] Reference Figures 1-4 This invention provides a method for enhancing ultrasound contrast images for nasoenteric tube insertion, comprising the following steps:

[0060] S1: Acquire and preprocess ultrasound contrast imaging sequences of the nasoenteric tube insertion area to highlight the microbubble signal characteristics in the ultrasound contrast imaging sequences;

[0061] S2: Establish a quantitative correlation between the physicochemical environment of the intestinal lumen and the characteristics of microbubble signals, obtain the expected attenuation degree of microbubble signal characteristics under the current environment, dynamically adjust the core parameters of the MSR algorithm based on the expected attenuation degree, and construct the MSR image processing model.

[0062] S3: Input the preprocessed ultrasound contrast image sequence into the MSR image processing model for enhancement processing to highlight the nasoenteric tube structure, and achieve accurate localization of the distal end of the nasoenteric tube based on the enhanced ultrasound contrast image sequence.

[0063] Specifically, in step S1, before acquiring the ultrasound contrast image sequence of the nasoenteric tube insertion area, an ultrasound contrast agent is injected into a peripheral vein. In this embodiment, the contrast agent is a microbubble formulation containing inert sulfur hexafluoride gas, which has better stability and harmonic characteristics, helping to obtain clearer images of the lumen. Then, a clinical ultrasound diagnostic device is used to scan the target area. The scanned target area focuses on the key anatomical sites where the nasoenteric tube is expected to run, such as the pylorus and duodenum, in order to capture the flow and distribution of the contrast agent in the area in real time, providing a high-quality original image basis for subsequent image enhancement and distal tube positioning.

[0064] As a preferred implementation, the clinical ultrasound diagnostic equipment here employs harmonic imaging mode. By receiving the nonlinear echo signal of microbubbles, it effectively suppresses tissue background and highlights contrast agent signals. The image acquisition frequency is set to 20 frames per second (fps), producing a continuous dynamic image sequence (denoted as P). k The method captures the real-time imaging process containing the expected course of the nasoenteric tube (in the form of k=1,2,…,N), aiming to clearly record the flow trajectory of microbubbles in the digestive tract.

[0065] For the acquired ultrasound contrast imaging image sequence, a registration algorithm based on feature points or mutual information is used to perform motion compensation on the image sequence. By spatially aligning each frame of the ultrasound contrast imaging image sequence, non-rigid deformation and translation are eliminated. Specifically, stable feature points at the inflection points of tissue edges in the ultrasound contrast imaging image sequence are detected, local descriptors of each feature point are extracted and continuous inter-frame matching is completed to establish reliable corresponding point pairs. Then, based on the matching point pairs, the spatial transformation model of non-rigid deformation between the matching frames is estimated. Geometric transformation and resampling are performed on each frame in this way to achieve spatial alignment with the reference frame and complete non-rigid motion compensation.

[0066] Traditional MSR algorithms typically estimate the illumination component and separate the reflection component based on preset fixed scales and weight parameters, which cannot adapt to changes in microbubble signals under different intestinal environments, leading to unstable enhancement effects. Therefore, in step S2 of this embodiment, a quantitative correlation is established between the intestinal physicochemical environment and microbubble signal characteristics. The expected attenuation level of the microbubble signal is inferred by quantifying intestinal peristalsis characteristics and intestinal fluid viscosity characteristics. The core parameters of the MSR algorithm are then dynamically adjusted based on the expected attenuation level of the microbubble signal. The specific steps are as follows:

[0067] (1) Based on the ultrasound contrast imaging sequence, extract the intestinal peristalsis features and intestinal fluid viscosity features to establish environmental perception assessment indicators;

[0068] The attenuation process of microbubble contrast imaging signals is not dominated by a single environmental factor, but is the result of the combined effects of intestinal fluid viscosity and intestinal peristalsis in the intestinal environment. When the intestinal fluid viscosity is heterogeneous, there are significant differences in the shear force and diffusion resistance of the liquid in different viscosity regions, which leads to selective retention and distribution imbalance of microbubbles at the viscosity difference interface. This manifests as regional distribution heterogeneity and morphological diversity in a single frame image. When the intestinal peristalsis rhythm is disordered, the irregular contraction and relaxation of the intestinal tract will directly affect the stability of the movement trajectory of the microbubble contrast agent. At this time, the microbubbles in the ultrasound contrast imaging sequence will show the characteristic of local deviation of the movement trajectory direction. Therefore, combining intestinal fluid viscosity and intestinal peristalsis for subsequent prediction of microbubble attenuation factors can improve the accuracy and reliability of attenuation law analysis.

[0069] In this step, by analyzing the static features of a single frame and the dynamic features between frames of the ultrasound contrast imaging sequence, the morphological features and spatiotemporal distribution patterns of the microbubble contrast agent are accurately extracted. This allows for the quantification of the uniformity of intestinal fluid viscosity and the degree of intestinal peristalsis disorder, thereby constructing a multi-dimensional intestinal physiochemical environment perception and evaluation index. This provides a standardized quantitative basis for the subsequent accurate evaluation of the microbubble contrast imaging signal attenuation characteristics. Specifically:

[0070] For single-frame image analysis, the K-means clustering algorithm is first used to cluster the pixels of each frame of ultrasound contrast-enhanced image. The number of clusters is set to 2, corresponding to bright pixel clusters representing microbubble signals and low-bright pixel clusters representing tissue background. Pixel classification is completed by calculating the grayscale distance between pixels and cluster centers. Then, the connected regions corresponding to the bright pixel clusters are extracted as regions of interest (ROIs), such as... Figure 3 As shown;

[0071] Then, the Suzuki contour tracking algorithm is used to scan the clustered binary image (including foreground and background) pixel by pixel. When the pixel value changes from background to foreground, the starting point of the contour is determined. Then, the contour is tracked along the boundary direction, and the discrete point sequence that constitutes the contour is recorded in sequence. The discrete contour points of the bright pixel cluster are accurately identified. Then, the minimum bounding rectangle of the contour is constructed as the bounding box, and the core analysis area is accurately selected by the bounding box.

[0072] Next, grayscale values ​​are extracted for all pixels within the bounding box. A grayscale histogram is constructed with the image grayscale level as the x-axis and the number of pixels at the corresponding grayscale level as the y-axis to quantify its grayscale distribution characteristics. The global minimum grayscale value is [value missing]. The global maximum grayscale value is The grayscale range of the X-axis of the histogram is [ , The number of pixels corresponding to each grayscale value within the specified range is counted, with the Y-axis representing the number of pixels falling within each grayscale value. Binning is performed based on the image grayscale resolution and analysis accuracy requirements. The Sturges formula k=log2m (where m is the total number of pixels within the bounding box) is used to calculate the number of bins, uniformly dividing the grayscale range to ensure that each pixel falls within a unique range.

[0073] The grayscale histogram is smoothed to suppress minor fluctuations caused by noise. Then, all local maxima (i.e. significant peaks) are found based on the smoothed grayscale histogram. From all the extracted significant peaks, the two peaks with the greatest grayscale distance are selected. The peak with the higher grayscale level is defined as the characteristic peak of the bright microbubble signal cluster, and the peak with the lower grayscale level is defined as the characteristic peak of the low-brightness tissue background.

[0074] For clusters of bright pixels within the bounding box, a weighted centroid extraction method based on pixel coordinates and grayscale values ​​is employed. The grayscale value of each pixel is used as a weighting factor to calculate the centroid coordinates of the cluster, ensuring the centroid position more closely aligns with the core region where microbubble signal concentration is high. This effectively avoids centroid shift caused by uneven pixel distribution or local grayscale fluctuations. Figure 4 As shown, by combining the temporal continuity characteristics of ultrasound contrast imaging image sequences, the centroid coordinates of the same bright pixel cluster in consecutive frame images are tracked and matched frame by frame. The difference between the centroid coordinates between two adjacent frames is extracted to construct a centroid displacement vector. The magnitude of the vector corresponds to the displacement distance of the centroid, and the direction of the vector corresponds to the direction of the centroid's movement.

[0075] Furthermore, the following intestinal environment characteristics were analyzed for each ultrasound contrast imaging sequence:

[0076] ① By quantifying the uniformity of the distribution of bright pixels within the bounding box and the uniformity of brightness within the cluster, the uniformity of intestinal fluid viscosity is obtained, and its expression is:

[0077] ;

[0078] In the formula: Indicates the uniformity of intestinal fluid viscosity. This represents the area of ​​the bounding box in the k-th frame. This represents the number of pixels in the highlighted cluster within the bounding box of the k-th frame. The hyperbolic tangent function tanh(⋅) aims to normalize this calculation result. This represents the total number of pixels in the grayscale histogram of the k-th frame image that represent the two main peaks (the characteristic peak of the microbubble signal cluster and the characteristic peak of the low-brightness background). This represents the total number of pixels covering all gray levels between the two peaks.

[0079] in, / The spatial filling density of microbubble signals within the analysis area was quantified: the larger the ratio, the more widely and densely the bright pixels are distributed within the area, suggesting that the microbubbles have low diffusion resistance and are evenly distributed in the intestinal fluid, and the intestinal fluid viscosity is likely to be uniformly distributed; the smaller the ratio, the sparsely distributed or locally concentrated microbubbles are, suggesting that there may be a viscosity barrier that hinders the uniform diffusion of microbubbles, and the intestinal fluid viscosity is likely to be unevenly distributed.

[0080] in, The ratio quantifies the degree of separation between the target signal and the background in grayscale in ultrasound contrast imaging: the closer the ratio is to 1, the more blurred the grayscale boundary between the bright pixels and the background pixels, which means that the brightness inside the microbubble cluster is uneven, and the shape may be broken or diverse. This indirectly reflects the heterogeneity of imaging caused by the different hydrodynamic effects of the surrounding fluid on the microbubbles due to the uneven viscosity of the intestinal fluid. At this time, the intestinal fluid viscosity is more likely to be unevenly distributed. If the ratio deviates from 1, it indicates that the signal and the background are clearly separated, and the brightness of the microbubble cluster is relatively concentrated and uniform. At this time, the intestinal fluid viscosity is more likely to be uniformly distributed.

[0081] ② By quantifying the continuity of the direction of microbubble signal movement, the stability of intestinal peristalsis is obtained, and its expression is:

[0082] ;

[0083] In the formula: Represents the displacement vector and The included angle between them, the range of values ​​is , , and Let represent the magnitudes (i.e., Euclidean distances) of the displacement vectors connecting the centroid of the k-th frame to its previous centroid and its subsequent centroid, respectively. The natural exponential function exp(-⋅) will... and difference normalization, wherein the greater the difference, the smaller the value of this item.

[0084] wherein,[ ] / 180° quantifies the continuity of the movement direction of microbubbles: the closer the ratio is to 1, it indicates that the movement direction has a large-amplitude deflection in a short time, reflecting that the movement direction of the intestinal wall that pushes the microbubbles is disorderly; the closer the ratio is to 0, it indicates that the movement direction is basically kept consistent, reflecting that the peristalsis direction of the intestinal tract is stable.

[0085] An environment perception evaluation index is established based on the uniformity of intestinal fluid viscosity and the stability of intestinal peristalsis, which is used as a comprehensive index for evaluating the physical and chemical environment state of the intestinal tract. The expression of the comprehensive index is as follows:

[0086] ;

[0087] in the formula: represents the comprehensive index for evaluating the physical and chemical environment state of the intestinal tract obtained by combining the uniformity index of intestinal fluid viscosity and the disorder degree index of intestinal peristalsis.

[0088] it is further necessary to set a physical and chemical environment state evaluation threshold T1: when < T1, it is determined that the physical and chemical environment state of the intestinal tract in the current frame is relatively poor, and the microbubble signal may have significant attenuation or distribution distortion; when ≥ T1, it is determined that the intestinal environment state in the current frame is relatively good, and the microbubble signal is relatively stable. In this embodiment, a reference threshold T1=0.4 is given, which can be used to balance the sensitivity and specificity of environment state evaluation. In practical applications, this threshold can be finely adjusted in combination with specific clinical scenarios and contrast agent characteristics.

[0089] (2) Constructing a microbubble state prediction factor according to the environment perception evaluation index;

[0090] Based on the environment perception evaluation index, poor frames and good frames of the intestinal physical and chemical environment are obtained, and the microbubble state prediction factor is constructed by establishing the mapping relationship between the poor frames and good frames and the ideal filtering scale and adjustment weight. The expression thereof is:

[0091] ;

[0092] in the formula: is the microbubble state prediction factor, N is the total number of collected time-series images, is the number of all environment-poor frames in the time-series images, is a very small positive number preventing the denominator from being 0, / N quantifies the coverage of poor environment states in the whole observation period. The larger the ratio, the greater the possibility and severity of the attenuation trend of the microbubble signal state. and These represent the image indices of the first and last frames in all the harsh environment frames, respectively. for and The number of good-environment frames between these two frames. The continuity of adverse environmental conditions over time was quantified. The smaller the ratio, the greater the likelihood and severity of the decay trend of the microbubble signal.

[0093] (3) The core parameters of the MSR algorithm are dynamically adjusted using the microbubble state predictor, and the MSR image processing model is established with the adjusted optimal core parameters. The expression is as follows:

[0094] ;

[0095] In the formula: This represents the core parameters after dynamic adjustment. This represents the core parameters of the Gaussian filter for the corresponding time-series image (including the filter scale). With large-scale component weights ).

[0096] when When the value is large, the expected attenuation of the microbubble harmonic signal is strong, and the effective tube wall detail signal in the ultrasound contrast image is easily submerged by noise or background. In this case, the scale of the Gaussian filter should be reduced and the weight of the large-scale component should be reduced during synthesis to make the filtering operation more localized, thereby more sensitively capturing weak edge and texture changes and enhancing detail features; when When the value is small, the expected attenuation of the microbubble signal is weak, and the ultrasound contrast image itself already has a relatively clear contrast. At this time, the original filtering scale and weight allocation should be maintained to ensure the global nature of the filtering operation, thereby maximizing the preservation of the naturalness of the image and the realism of the anatomical structure, and providing a reliable image basis for subsequent nasoenteric tube localization.

[0097] Furthermore, in step S3, based on the completion of environmental perception and parameter adaptive adjustment, the preprocessed ultrasound contrast image sequence is input into the MSR image processing model. Through image enhancement processing that matches the environmental state, the contrast between the nasoenteric tube structure and the background is significantly improved, highlighting the nasoenteric tube structure. Then, based on the enhanced stable image features, the accurate spatial localization of the distal end of the nasoenteric tube is achieved. The specific steps are as follows:

[0098] First, a logarithmic transformation is performed on the preprocessed ultrasound contrast image sequence to map the pixel gray values ​​to a more compact range, laying the foundation for subsequent separation of illumination and reflection components.

[0099] Subsequently, based on the dynamically adjusted Gaussian filter scale With large-scale component weights The process involves performing multi-scale Gaussian filtering, constructing a corresponding Gaussian kernel function for each filtering scale, and obtaining the Gaussian filtered image at the corresponding scale. The filtered images at each scale are then weighted and averaged to obtain the estimated illumination component. The reflection component is then obtained by calculating the difference between the logarithmic image and the illumination component. Finally, after exponential transformation and grayscale stretching, the enhanced ultrasound contrast image is output.

[0100] Finally, on the enhanced ultrasound contrast image sequence, edge detection was performed on each frame of the enhanced ultrasound contrast image using the Canny operator to extract potential tube edge maps in the image. At the same time, morphological operations were combined to process the edge maps. The edge segments that were broken due to local weak tube signals were connected by the closing operation, and isolated small-scale noise points were filtered out by the opening operation, thereby forming a continuous nasoenteric tube outline. Combined with the tube movement trajectory in the ultrasound contrast image sequence, the spatial positioning of the distal end of the nasoenteric tube was realized, providing visual guidance and positioning support for clinical operation.

[0101] Example 2:

[0102] The present invention also provides an ultrasound contrast imaging enhancement system for nasoenteric tube insertion, the system being used to implement the steps of the above-described method, the system comprising:

[0103] The acquisition module is used to acquire ultrasound contrast imaging sequences of the nasoenteric tube insertion area;

[0104] The image processing module is used to preprocess the ultrasound contrast imaging image sequence to highlight the microbubble signal characteristics in the ultrasound contrast imaging image sequence;

[0105] The signal feature change acquisition module establishes a quantitative correlation between the intestinal lumen physicochemical environment and microbubble signal features, and obtains the expected attenuation degree of microbubble signal features under the current environment;

[0106] The model building module dynamically adjusts the core parameters of the MSR algorithm based on the expected attenuation level to build the MSR image processing model;

[0107] The enhancement module is used to input the pre-processed ultrasound contrast image sequence into the MSR image processing model for enhancement processing to highlight the nasoenteric canal structure.

[0108] The positioning module achieves precise positioning of the distal end of the nasoenteric tube based on enhanced ultrasound contrast imaging sequences.

[0109] In addition, the present invention also discloses an electronic device, the electronic device including at least one processor and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the protection method described in the above embodiments.

[0110] The memory and processor are connected via a bus, which can include any number of interconnecting buses and bridges, connecting various circuits of one or more processors and memories. The bus can also connect various other circuits such as peripherals, voltage regulators, and power management circuits. The bus interface provides an interface between the bus and the transceiver, which can be a single component or multiple components, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over the wireless medium via an antenna, which further receives data and transmits it back to the processor.

[0111] The processor is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management and other control functions, while the memory can be used to store the data used by the processor when performing operations.

[0112] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for enhancing ultrasound contrast images for nasoenteric tube insertion, characterized in that, Includes the following steps: Acquired and preprocessed ultrasound contrast imaging sequences of the nasoenteric tube insertion area to highlight the microbubble signal characteristics in the ultrasound contrast imaging sequences; A quantitative correlation was established between the intestinal physicochemical environment and microbubble signal characteristics. The expected attenuation level of microbubble signal characteristics under the current environment was obtained. The core parameters of the MSR algorithm were dynamically adjusted based on the expected attenuation level to construct an MSR image processing model. The steps for constructing the MSR image processing model include: Based on the extraction of intestinal peristalsis features and intestinal fluid viscosity features from ultrasound contrast imaging sequences, an environmental perception assessment index was established. Microbubble state prediction factors were constructed based on environmental perception and assessment indicators. The core parameters of the MSR algorithm are dynamically adjusted using microbubble state predictors, and an MSR image processing model is established with the adjusted optimal core parameters. The preprocessed ultrasound contrast imaging sequence is input into the MSR image processing model for enhancement, highlighting the nasoenteric tube structure. Based on the enhanced ultrasound contrast imaging sequence, the distal end of the nasoenteric tube is accurately located.

2. The ultrasound contrast imaging enhancement method for nasoenteric tube insertion according to claim 1, characterized in that, The steps for acquiring and preprocessing ultrasound contrast-enhanced images of the nasoenteric tube insertion area include: The real-time contrast imaging process, including key parts of the expected course of the nasoenteric tube, was captured using harmonic imaging mode, resulting in a sequence of ultrasound contrast images containing microbubble signals. Noise suppression processing was performed on the ultrasound contrast imaging sequence; A registration algorithm is used to perform motion compensation on the noise-suppressed ultrasound contrast imaging sequence, so that each frame in the ultrasound contrast imaging sequence is spatially aligned.

3. The ultrasound contrast imaging enhancement method for nasoenteric tube insertion according to claim 2, characterized in that, Based on the extraction of intestinal peristalsis and intestinal fluid viscosity characteristics from ultrasound contrast imaging sequences, an environmental perception assessment index is established, specifically including: Extract the region of interest (ROI) from each frame of ultrasound contrast imaging image, identify the outline of the bright pixel clusters based on the ROI, and construct the bounding box using the outline of the bright pixel clusters. The uniformity of intestinal fluid viscosity is obtained by quantifying the uniformity of the distribution of bright pixels within the bounding box and the uniformity of brightness within the cluster. The stability of intestinal peristalsis can be obtained by quantifying the continuity of the direction of microbubble signal movement. An environmental perception assessment index was established based on the uniformity of intestinal fluid viscosity and the stability of intestinal peristalsis, serving as a comprehensive indicator for assessing the physicochemical state of the intestine.

4. The ultrasound contrast imaging enhancement method for nasoenteric tube insertion according to claim 3, characterized in that, The steps to construct the minimum bounding rectangle of the highlighted pixel cluster as the bounding box include: The K-means clustering algorithm is used to perform pixel clustering on grayscale images. The pixel clustering results include bright pixel clusters representing microbubble signals and low-bright pixel clusters representing tissue background. Regions of interest are extracted based on the pixel classification results. The contour detection algorithm identifies discrete contour points of bright pixel clusters within the region of interest, and uses the smallest bounding rectangle of the discrete contour points of the bright pixel clusters as the bounding box.

5. The ultrasound contrast imaging enhancement method for nasoenteric tube insertion according to claim 4, characterized in that, For all pixels within the bounding box, a grayscale histogram of image grayscale level and pixel count is constructed. Based on the grayscale histogram, all significant peak points are found. From multiple significant peak points, the two peak points with the greatest grayscale distance are selected as the characteristic peaks of high-brightness microbubble signal clusters and low-brightness tissue background, respectively.

6. The ultrasound contrast imaging enhancement method for nasoenteric tube insertion according to claim 5, characterized in that, Microbubble state prediction factors were constructed based on environmental perception assessment indicators, including: Based on environmental perception assessment indicators, we obtain the adverse frames and favorable frames of the intestinal physicochemical environment. By combining bad frames and good frames to adjust the weights, a microbubble state prediction factor is constructed.

7. The ultrasound contrast imaging enhancement method for nasoenteric tube insertion according to claim 6, characterized in that, The MSR image processing model is established using the adjusted optimal core parameters. The expression of the MSR image processing model is as follows: ； in, This represents the core parameters after dynamic adjustment. These represent the core parameters of the Gaussian filter for the corresponding time-series image. This represents the predictor of microbubble state.

8. The ultrasound contrast imaging enhancement method for nasoenteric tube insertion according to claim 7, characterized in that, The preprocessed ultrasound contrast-enhanced image sequence was input into the MSR image processing model for enhancement, highlighting the nasoenteric canal structure, including: A logarithmic transformation is performed on the preprocessed ultrasound contrast image sequence to map the pixel gray values ​​to a more compact range; The dynamically adjusted MSR image processing model is used to further perform multi-scale Gaussian filtering and reflection component calculation, followed by exponential transformation and grayscale stretching to output an enhanced ultrasound contrast image sequence. The contour of the nasoenteric tube was extracted from the enhanced ultrasound imaging sequence, and the spatial localization of the distal end of the nasoenteric tube was achieved by combining the tube's movement trajectory.

Images