A method and system for assisting in the care of defecation

Abstract

A nursing method and system for assisting defecation includes: a multimodal data acquisition unit for acquiring multimodal data of the anal sphincter region, including image data and biosignal data; a feature parameter extraction unit for segmenting the image data using a convolutional neural network segmentation model to generate a binary mask of the anal sphincter region and extracting feature parameters of the sphincter core region, including sphincter thickness and sphincter continuity; a stimulation strategy control unit for registering the sphincter core region after binary masking with a temperature matrix image using anatomical markers on the sphincter surface to obtain a temperature matrix of the sphincter core region, calculating gradient values ​​on the temperature matrix to obtain the temperature gradient of the temperature data damage area; calculating the anal resting pressure morphological coupling coefficient based on sphincter thickness, sphincter continuity, and anal resting pressure in the biosignal data; and a stimulation unit for adjusting the electroacupuncture stimulation parameters of the electrical stimulation terminal according to the coupling coefficient.

Description

Technical Field

[0001] This invention relates to the field of postoperative care for anorectal surgery, and in particular to a method and system for assisting defecation. Background Technology

[0002] Patients undergoing pelvic surgeries such as rectal cancer often face the serious challenge of postoperative bowel dysfunction, with low anterior resection syndrome (LARS) occurring in 40%-90% of cases. This syndrome manifests as frequent bowel movements, urgency, and fecal incontinence, and 50% of patients still experience moderate to severe dysfunction one year post-surgery, significantly increasing the risk of anxiety and depression, as well as the medical burden. Current postoperative bowel function rehabilitation methods have significant limitations.

[0003] Existing treatment options include drug therapy, such as loperamide and probiotics, which can only relieve symptoms in the short term, are ineffective against organic dysfunction caused by neuromuscular damage, and long-term use can easily lead to intestinal flora imbalance.

[0004] Biofeedback training relies on the patient's active cooperation, which is difficult for patients who are weak or have cognitive impairment after surgery to complete, resulting in insufficient clinical effectiveness.

[0005] Surgical interventions carry both trauma and risks. For example, invasive treatments such as sacral nerve modulation (SNS) require the implantation of permanent electrodes, which carries risks of infection and electrode displacement. Furthermore, the efficacy of these treatments varies greatly among patients with LARS, with an effectiveness rate of only 40%-70%. While fecal diversion stomas can temporarily resolve fecal incontinence, they require a second surgery to restore them, increasing the patient's physical and psychological trauma.

[0006] Although traditional Chinese medicine electroacupuncture has shown a regulatory effect on the electrical activity of anorectal muscles in clinical observations, most existing nerve electrical stimulation devices use fixed parameters, such as constant frequency and constant current, to stimulate nerves, which cannot adapt to the dynamic changes of the anal sphincter in postoperative patients.

[0007] Therefore, existing technologies have shortcomings and need further improvement and development. Summary of the Invention

[0008] (I) Purpose of the invention: In order to solve the problems existing in the prior art, the purpose of the present invention is to provide a more efficient nursing method and system for assisting defecation with electrical stimulation.

[0009] (II) Technical Solution: To solve the above-mentioned technical problems, this technical solution provides a nursing method to assist defecation, specifically including the following steps:

[0010] Multimodal data of the anal sphincter region were collected, including image data and biosignal data.

[0011] A convolutional neural network segmentation model is used to segment the acquired image data to generate a binary mask of the anal sphincter region. The sphincter core region in the multimodal data is then cropped based on the binary mask, and feature parameters of the sphincter core region are extracted. The feature parameters include sphincter thickness and sphincter continuity.

[0012] By registering the sphincter core region after binary masking and temperature matrix image with the anatomical markers on the sphincter surface, the temperature matrix of the sphincter core region is obtained. The gradient value of the temperature matrix is ​​then calculated to obtain the temperature gradient of the temperature data damage area.

[0013] Based on the sphincter thickness, sphincter continuity, and anal resting pressure in the biosignal data, the anal resting pressure morphological coupling coefficient is calculated, and the electroacupuncture stimulation parameters of the electrical stimulation terminal are adjusted according to the coupling coefficient.

[0014] The aforementioned method for assisting defecation includes image data comprising RGB color image data of the anal sphincter region, a temperature matrix formed by infrared thermal imaging temperature distribution data, and depth distance image data; the biosignal data also includes anal contraction pressure.

[0015] The aforementioned method for assisting defecation involves establishing virtual directional markers based on the anal center point, bony anatomical reference points, and soft tissue functional points, following a clockwork orientation. The bony anatomical reference points include the tip of the coccyx, the ischial tuberosity, and the lower edge of the pubic symphysis. The soft tissue functional points include the anal center point, the sphincter contraction center point, the Changqiang acupoint (located at the midpoint of the line connecting the coccyx tip and the anal center point), and the apex of the anal margin fold.

[0016] The aforementioned method for assisting defecation further includes spatiotemporal correlation of multimodal data, calculating the dynamic contraction speed of the sphincter by tracking the displacement of key structural points in three consecutive frames of RGB color image data, generating a contraction speed-time curve. The key structural points include the anal center point and the sphincter contraction center point among the anatomical markers on the sphincter surface. When the absolute value of the slope of the contraction speed-time curve is >0.5mm / s, it is determined to be an abnormal marker of sphincter contraction speed. The depth distance image data is registered with infrared thermal imaging temperature distribution data through the anatomical markers on the sphincter surface to locate the low-temperature depression complex abnormal area, and the low-temperature depression complex abnormal area is used as the target point for electroacupuncture stimulation.

[0017] The aforementioned method for assisting defecation, wherein the formula for calculating the anal resting pressure morphological coupling coefficient K is as follows: , where P 静息The collected anal resting pressure is P0, which is the average anal resting pressure of a normal adult. T is the collected sphincter thickness, which is T0, which is the average sphincter thickness of a normal adult. D is the proportion of the damaged area; the proportion of the damaged area D is the ratio of the area of ​​the damaged area to the area of ​​the core sphincter region.

[0018] The aforementioned method for assisting defecation includes determining the stimulation phase of the stimulation unit based on the anal resting pressure morphological coupling coefficient, and adjusting the electroacupuncture stimulation parameters in real time based on the characteristic parameters of the stimulation phase; the stimulation phase includes a repair stagnation phase, a repair initiation phase, and a repair acceleration phase; the characteristic parameters also include the proportion of the damaged area D, the temperature gradient of the damaged area G, and the sphincter contraction speed slope k. b .

[0019] The aforementioned method for assisting defecation involves stimulating the anal sphincter when the anal resting pressure morphological coupling coefficient K < 0.3 and the damaged area percentage D > 50%, indicating that the stimulation phase is in the repair stagnation phase. The electroacupuncture stimulation parameters are adjusted as follows: frequency 8-10Hz, intensity 2.5-3mA, target point is Changqiang acupoint and acupoint about 1-2cm directly below the anal margin, and the duration of a single stimulation session is 30 minutes.

[0020] The aforementioned method for assisting defecation, wherein when 0.3 ≤ anal resting pressure morphological coupling coefficient K < 0.6 and the temperature gradient of the damaged area G < 0.3℃ / mm, the stimulation stage is determined to be in the repair initiation stage; the electroacupuncture stimulation parameters are adjusted to: frequency 20-25Hz, intensity 1.5-2mA, target point is low temperature depression complex abnormal area, and the duration of a single stimulation is extended to 40 minutes.

[0021] The aforementioned nursing method for assisting defecation includes a condition where the anal resting pressure morphological coupling coefficient K ≥ 0.6 and the sphincter contraction velocity slope k b If the anal sphincter pressure is ≥-0.5 mm / s² and the anal sphincter pressure is ≥120 mmHg, the stimulation phase is determined to be in the repair acceleration phase. The electroacupuncture stimulation parameters are adjusted to: frequency 30-35 Hz, intensity 1-1.5 mA, and the target points are the three acupoints of Zhongliao, Xialiao and Huiyang, which are soft tissue functional points. The duration of a single stimulation session is shortened to 20 minutes.

[0022] A defecation assistance care system, comprising:

[0023] A multimodal data acquisition unit is used to acquire multimodal data of the anal sphincter region, including image data and biosignal data.

[0024] The feature parameter extraction unit is used to segment the acquired image data using a convolutional neural network segmentation model to generate a binary mask of the anal sphincter region, crop the sphincter core region in the multimodal data according to the binary mask, and extract the feature parameters of the sphincter core region, including sphincter thickness and sphincter continuity.

[0025] The stimulation strategy control unit is used to register the sphincter core region after binary masking and the temperature matrix image through anatomical markers on the sphincter surface to obtain the temperature matrix of the sphincter core region, calculate the gradient value of the temperature matrix to obtain the temperature gradient of the temperature data damage area, and calculate the anal resting pressure morphological coupling coefficient based on the sphincter thickness, sphincter continuity and anal resting pressure in the biosignal data.

[0026] The stimulation unit is used to adjust the electroacupuncture stimulation parameters of the electrical stimulation terminal according to the coupling coefficient.

[0027] (III) Beneficial effects: The nursing method and system for assisting defecation provided by the present invention significantly improves computational efficiency, realizes phased monitoring and nursing, and is suitable for a variety of clinical application scenarios; the multimodal data fusion optimization adopted by the present invention supports individualized treatment, reduces clinical misjudgment, simplifies clinical operation procedures, and improves treatment accuracy. Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating a nursing method for assisting defecation according to the present invention;

[0029] Figure 2 This is a preferred embodiment of the feature parameter extraction unit for trimming the core area of ​​the sphincter muscle in a nursing method for assisting defecation according to the present invention;

[0030] Figure 3 This is a preferred embodiment of the extraction of characteristic parameters of the core region of the sphincter muscle in a nursing method for assisting defecation according to the present invention;

[0031] Figure 4 This is a schematic diagram of the structure of a nursing system for assisting defecation according to the present invention. Detailed Implementation

[0032] The present invention will be further described in detail below with reference to preferred embodiments. More details are set forth in the following description in order to provide a full understanding of the present invention. However, the present invention can obviously be implemented in many other ways different from those described herein. Those skilled in the art can make similar extensions and derivations based on actual application situations without departing from the spirit of the present invention. Therefore, the scope of protection of the present invention should not be limited by the content of this specific embodiment.

[0033] The accompanying drawings are schematic diagrams of embodiments of the present invention. It should be noted that these drawings are for illustrative purposes only and are not drawn to scale, and should not be construed as limiting the actual scope of protection of the present invention.

[0034] This invention provides a nursing method for assisting defecation, and a nursing system for assisting defecation that uses electrical stimulation terminals to stimulate different states of an individual to improve their defecation ability, such as... Figure 1 , Figure 4 As shown, the specific steps include:

[0035] Step 101: The multimodal data acquisition unit acquires multimodal data of the anal sphincter region, including image data and biosignal data;

[0036] Step 102: The feature parameter extraction unit segments the acquired image data using a convolutional neural network segmentation model to generate a binary mask of the anal sphincter region. Based on the binary mask, the sphincter core region in the multimodal data is cropped, and the feature parameters of the sphincter core region are extracted. The feature parameters include sphincter thickness and sphincter continuity.

[0037] Step 103: The stimulation strategy control unit registers the sphincter core region after binary masking and the temperature matrix image through the sphincter surface anatomical markers to obtain the temperature matrix of the sphincter core region, and calculates the gradient value of the temperature matrix to obtain the temperature gradient of the temperature data damage area.

[0038] Step 104: The stimulation unit calculates the anal resting pressure morphological coupling coefficient based on the sphincter thickness, sphincter continuity, and anal resting pressure in the biosignal data, and adjusts the electroacupuncture stimulation parameters of the electrical stimulation terminal according to the coupling coefficient.

[0039] The multimodal data acquisition unit collects multimodal data of the anal sphincter region, which can be achieved using a medical-grade miniature endoscope. The endoscope is equipped with an image data acquisition module and a biosignal acquisition module. The image data acquisition module includes an RGB color image sensor, an infrared thermal imaging sensor, and a depth / distance image sensor.

[0040] The RGB color image sensor acquires RGB images, specifically capturing texture details of the anal sphincter area, such as anal margin folds and mucosal color. The RGB color image sensor can be positioned at the center of the front end face of the endoscope probe.

[0041] The infrared thermal imaging sensor acquires infrared images and generates a two-dimensional temperature matrix for the sphincter region.

[0042] The depth distance image sensor acquires depth images. Specifically, the depth distance image sensor collects three-dimensional spatial geometric information of the anal sphincter region based on the depth images, and constructs three-dimensional point cloud data of the depth distance image of the target region through non-contact measurement, including stereo positioning of the sphincter structure, morphological analysis and multimodal data registration to provide spatial coordinate reference.

[0043] The biosignal acquisition module includes a pressure sensor array, and the biosignal data signals acquired by the pressure sensor array include anal resting pressure and anal contraction pressure acquired by direct contact with the rectal mucosa.

[0044] The image data includes RGB color image data of the anal sphincter region, a temperature matrix formed from infrared thermal imaging temperature distribution data, and depth distance image data. The biosignal data also includes anal contraction pressure.

[0045] In the feature parameter extraction unit, the convolutional neural network segmentation model is not restricted in generating a binary mask of the anal sphincter region by segmenting the acquired image data using a convolutional neural network segmentation model. The convolutional neural network segmentation model can be an improved U-Net++ model used to achieve automatic segmentation of the anal sphincter region and obtain a binary mask of the anal sphincter region from the acquired image data. The specific process is as follows:

[0046] I. Data Preprocessing Stage;

[0047] First, the RGB image and the depth image are fused to obtain a fused image. For example, a 512×512 RGB image is superimposed with a depth image normalized to [0,255] as a 4-channel input (R / G / B / Depth) to enhance spatial features.

[0048] Then, the fused image is labeled with sphincter region = 1 and background = 0 to obtain a binarized image, which constitutes the standard dataset. The binarized image is an image including a binarized mask, where each pixel has only two possible values: a value of 1 represents the anal sphincter region and a value of 0 represents the background. The preprocessed image is labeled with sphincter region = 1 and background = 0, and the labeled results are fused using the STAPLE algorithm to generate an accurate standard dataset.

[0049] Finally, the standard dataset is augmented to obtain an enhanced dataset. Augmentation of the standard dataset can be achieved by randomly rotating it from -15° to +15°, adjusting the contrast by ±20%, and setting the Gaussian blur standard deviation to 0.5. This expands the samples in the standard dataset to include 10,000 samples.

[0050] The encoder of the convolutional neural network segmentation model includes an input layer and a downsampling module. The input layer receives a 4-channel image, i.e., a fused image. The downsampling module includes four levels of downsampling, each level containing two residual convolutional blocks and one 2×2 max pooling operation, and introduces an attention gate mechanism to suppress background noise, such as intestinal gas and feces. For example, the feature map size after the first level of downsampling is 256×256×64, and that after the fourth level is 32×32×1024.

[0051] The decoder of the convolutional neural network segmentation model includes an upsampling module and an output layer. The upsampling module includes a 4-level transposed convolution with a stride of 2, and fuses the multi-scale feature maps passed from the encoder. The output layer includes a 1×1 convolutional kernel that outputs a single-channel probability map of 512×512×1, where the pixel values ​​in the single-channel probability map represent the probability of belonging to the sphincter region, and the pixel values ​​include 0 and 1.

[0052] II. Post-processing of binary masking;

[0053] A 3×3 matrix is ​​used to perform erosion on regions of interest (ROIs) in the augmentation dataset sample images. By convolving the image with a structuring element, small bumps or noise points at the edges of the ROIs are eliminated. Following the erosion, two dilation operations are performed on the eroded ROIs. The structuring element is used to expand the area of ​​the ROIs, connecting any fragmented parts caused by erosion, and further smoothing the edges. This invention, through the combination of erosion and dilation operations, effectively eliminates noise regions smaller than 50 pixels, achieving morphological optimization of the image and improving sample quality.

[0054] This invention performs two dilation operations on the region of interest in the eroded image. The morphological enhancement process, which aims to restore details of over-erosion and enhance regional connectivity, can make the feature parameters of the sphincter core region described in this invention more refined and accurate. Detailed embodiments are as follows.

[0055] For the corroded image, a morphological probe in morphological operations is adopted. The morphological probe is usually an n×n matrix, such as 3×3, 5×5. The morphological probe moves pixel by pixel on the corroded image. Dilation operations are performed through the shape of the matrix and the change of the matrix element values being 0 or 1. The meaning of 0 is the area in the morphological probe that does not participate in detection, and 1 is the area in the morphological probe that needs to participate in detection. Rules of the morphological probe: When the morphological probe moves to a certain position of the corroded image, only when all the positions marked as 1 in the morphological probe correspond to the target pixel value of 1 in the corroded image, this position will be retained as the target; otherwise, it will be corroded to the background 0. If there are 0s in the morphological probe, these positions do not participate in the judgment, and whether the corresponding position in the corroded image is 0 or 1 does not affect the result.

[0056] The purpose of the dilation operation of the present invention is to expand the target area. When the matrix elements of the morphological probe are all-1 structural elements, the detection range is the largest, which is suitable for overall shrinking / expanding of the target area, such as removing large noises, merging large areas, and is also suitable for operations of smoothing edges and enhancing overall connectivity.

[0057] The structural elements of the morphological probe of the present invention containing 0, such as cross-shaped and annular shapes, are used for limited detection ranges and are suitable for retaining specific direction features of the image. For example, the cross-shaped probe can preferentially dilate along the horizontal / vertical directions without affecting the details in the diagonal directions. The cross-shaped probe of the present invention is used. The matrix elements of the cross-shaped probe can contain 0. When dilating the Chinese character "十", it will only thicken along the horizontal and vertical directions and will not expand obliquely. The cross-shaped probe is suitable for retaining the edge details of the target area.

[0058] The purpose of the present invention using the morphological probe to perform the first dilation operation is to restore the overly shrunk target main body area in the corrosion operation, such as broken edges and narrow connected parts, and fill the first holes in the target area. The first holes refer to the holes with diameters smaller than the size of the structural element. The operation logic is to traverse each pixel in the overly shrunk target main body area of the corroded image. If at least one pixel in the area covered by the morphological probe is the target pixel 1, then this pixel is set to 1; otherwise, it is retained as 0. The first dilation operation expands the boundary of the target area outward, reconnects the broken areas, and the area is close to the original size before corrosion.

[0059] This invention uses a morphological probe for a second dilation, aiming to enhance expansion and morphological adjustment, further enlarge the target region, improve its connectivity with surrounding similar regions, and merge adjacent small targets. This invention selects to fill the second holes that still exist after the first dilation, with the diameter of the second hole ≤ twice the size of the structuring element. The operational logic is to iterate through each pixel of the over-shrunken target region in the eroded image; if at least one pixel within the morphological probe's coverage area is target pixel 1, then that pixel is set to 1; otherwise, it is left as 0. The morphological probe is used to perform a second dilation on the image after the first dilation, increasing the target region area by approximately 1-2 pixel layers compared to the first dilation, resulting in smoother edges and significantly improved overall connectivity.

[0060] The feature parameter extraction unit filters connected components, retains the largest connected component, and removes interfering areas such as intestinal edges and air bubbles, ultimately generating a binary mask of the sphincter region, for example: 512×512 pixels, with the white area representing the sphincter.

[0061] The preferred embodiment of the feature parameter extraction unit for cropping the core region of the sphincter is as follows:

[0062] Based on binary masking, core regions in RGB / infrared / depth data are cropped, such as... Figure 2 As shown, the steps are as follows:

[0063] 1. Determine the edge line of the area of ​​interest, which refers to the core area of ​​the sphincter muscle;

[0064] Calculate the minimum bounding rectangle of the binary mask: x_min, y_min, x_max, y_max, for example: x_min=120, y_min=90, x_max=380, y_max=420.

[0065] For anatomical marker constraints, if the minimum bounding rectangle does not include the anal center point, the minimum bounding rectangle is extended outward by 10%, for example, x_min is adjusted to 108 and x_max is adjusted to 392, to ensure that the minimum bounding rectangle includes the anatomical marker and to guarantee the integrity of the sphincter core area.

[0066] 2. Multimodal data is cropped synchronously. The RGB image is cropped to 256×256 pixels according to the coordinates of the region of interest (380-120=260, rounded to 256). Infrared temperature matrix: The 160×120 infrared image is resized to 256×256 using bilinear interpolation, and then cropped according to the region of interest. Depth image: Cropped in the same way as the RGB image, preserving the three-dimensional distance information of the sphincter region.

[0067] like Figure 3As shown, the preferred embodiments for extracting the characteristic parameters of the sphincter core region, including sphincter thickness and sphincter continuity, are as follows:

[0068] First, the sphincter thickness T is extracted based on the cropped depth image data, and the steps are as follows:

[0069] Establish a coordinate system: With the center point of the anus as the origin O, establish a polar coordinate system (r, θ), where θ is the circumferential angle from 0° to 360°, and r is the radial distance. Set sampling lines within the coordinate system: Take one sampling line every 10° along the θ direction, for a total of 36 lines. Each line extends from the inner edge r1 of the sphincter to the outer edge r2. The thickness T of each sampling line is... i =r2-r1, take the average value of all sampling lines (36 lines) as the final thickness T:

[0070] ,

[0071] If the average thickness of the 36 sampling lines is 6.2 mm, then the sphincter thickness T = 6.2 mm, and the average sphincter thickness of a normal adult is T0 = 5.5 mm.

[0072] The steps for extracting the continuous C-section of the sphincter are as follows:

[0073] Based on the morphological features of the binary mask, the number of connected components N is calculated: count the number of connected components with a pixel value of 1 in the mask (8-neighbor connections), and for a normal complete sphincter, N=1.

[0074] Calculate the area ratio of holes V: Calculate the ratio of the area of ​​holes in the binary mask, that is, the area of ​​pixels with a value of 0 but surrounded by 1, to the total area of ​​the mask.

[0075] The sphincter continuity index C is calculated by combining the number of connected components N and the proportion of the orifice area V. The formula is as follows:

[0076]

[0077] If the mask is a simply connected domain N=1 and the area of ​​the hole is V=5%, then the sphincter continuity C=0.95. The closer the value of the sphincter continuity is to 1, the better the continuity.

[0078] This invention achieves precise segmentation of the anal sphincter region through multimodal image fusion and convolutional neural network segmentation model; based on depth image and morphological analysis, it quantitatively extracts key feature parameters such as sphincter thickness and sphincter continuity.

[0079] The stimulation strategy control unit registers the sphincter core region after binary masking and the temperature matrix image using anatomical markers on the sphincter surface to obtain the temperature matrix of the sphincter core region. The gradient value of the temperature matrix is ​​then calculated to obtain the temperature gradient of the temperature data damage area. A preferred embodiment is as follows:

[0080] Determine the anatomical landmarks of the sphincter: Based on the pelvic floor anatomy, select three key surface landmarks, such as the midpoint A of the upper edge of the pubic symphysis, and the most prominent points B and C of the two ischial tuberosities. Use these three key surface landmarks to determine the approximate range of the core area of ​​the sphincter, i.e. the range of the first region, as the spatial reference for subsequent cutting and registration.

[0081] Generate a binary mask and crop the sphincter core region:

[0082] The original image is acquired by an infrared thermal imaging sensor that collects a binarized image of the pelvic floor region. The foreground is the sphincter area with a pixel value of 255 and the background is 0. The image resolution is 200×200 pixels.

[0083] Draw a ROI mask, using marked points A, B, and C as vertices to generate a triangular binary mask. Pixels within the mask are set to 1, and all others to 0.

[0084] Masking matrix M (200×200): M[i,j] = 1 (if pixel (i,j) is located inside △ABC) M[i,j] = 0 (otherwise)

[0085] The core region is cropped by multiplying the pixel value matrix of the original binarized image with the pixel value matrix of the binary mask M, preserving the sphincter region within the mask, thus obtaining the cropped binarized image I. cropped The binarized image I cropped The size is approximately 100×80 pixels.

[0086] A preferred embodiment of the present invention for registering the sphincter core region with a temperature matrix image is as follows:

[0087] A temperature matrix is ​​acquired, and a pelvic floor temperature matrix image T is collected synchronously. The resolution of the pelvic floor temperature matrix image T is 200×200 pixels, and each pixel value represents the temperature in °C, ranging from 32 to 38 °C.

[0088] Spatial registration of pelvic floor temperature matrix image T and sphincter core region.

[0089] Coordinate transformation: Due to the binary masking I of the sphincter core region cropped The temperature matrix image T is based on the same coordinate system (aligned with the surface markers), and I is directly plotted using the coordinates of markers A, B, and C. cropped The ROI is mapped to T, and the temperature data of the corresponding region is extracted.

[0090] Interpolation processing: If the temperature matrix and the binary image resolution are inconsistent, bilinear interpolation is used to resample the temperature data to I. cropped The size (100×80 pixels) is used to obtain the registered temperature matrix T of the sphincter core region. core .

[0091] Calculate the gradient value of the temperature matrix:

[0092] Gradient calculation, using the Sobel operator to calculate T core The gradients along the x-axis (horizontal direction) and y-axis (vertical direction) are calculated using the following formulas:

[0093] gradient G in the x-direction x = T core Sobel x (Sobel) x (3×3 horizontal gradient template)

[0094] gradient G in the y-direction y = T core Sobel y (Sobel) y (3×3 vertical gradient template)

[0095] Gradient magnitude G = sqrt(G x ² + G y ² (represents the magnitude of the rate of temperature change)

[0096] Example data: Assume T core A temperature matrix segment of a damaged area, the temperature matrix being 5×5 pixels, is shown below (unit: °C):

[0097]

[0098] After calculating the gradient, the gradient magnitude G at the edge of the damaged area of ​​the sphincter may be significantly higher than that in the normal area. For example, G≈0.1~0.3℃ / pixel in the normal area and G≈0.5~0.8℃ / pixel in the damaged area.

[0099] Identify the temperature data of the core sphincter region to identify the damaged area:

[0100] Threshold segmentation, setting a gradient magnitude threshold G thresh = 0.4℃ / pixel, G > G thresh The area was identified as an abnormal temperature gradient region, that is, a region with damaged temperature data in the core area of ​​the sphincter.

[0101] The output of the results marks the coordinate range of the temperature data damage area, such as pixel coordinates (20~30, 40~50)), and calculates the average gradient value of the area, such as 0.65℃ / pixel, as a quantitative indicator of sphincter dysfunction.

[0102] This invention uses a process of marker positioning, masking and cropping, image registration, and gradient calculation to accurately extract the temperature gradient features of the core region of the sphincter muscle from the temperature matrix, which helps to locate potential injury areas. In addition, it can also combine clinical data to optimize the selection of markers and gradient thresholds to improve accuracy.

[0103] Using the center point of the anus as the center, virtual orientation markers are established based on bony anatomical reference points and soft tissue functional points, following the orientation of a clock. The bony anatomical reference points include the tip of the coccyx, the ischial tuberosity, and the lower edge of the pubic symphysis. The soft tissue functional points include the center point of the anus, the center point of sphincter contraction, the Changqiang acupoint at the midpoint of the line connecting the tip of the coccyx and the center point of the anus, and the apex of the anal margin fold.

[0104] Using the center point of the anus as the origin (O) of the three-dimensional coordinate system, a standardized virtual orientation marking system is constructed by integrating bony anatomical landmarks and soft tissue functional characteristics. This system includes the following steps:

[0105] Selection of bony anatomical reference points

[0106] Coccyx tip (P1): Serves as a posterior positioning reference, located approximately 4-5 cm directly behind the center of the anus. Its coordinates (0, -5, 0) are determined by palpation or X-ray imaging.

[0107] Bilateral ischial tuberosities (P2, P3): Marked as left and right boundaries, with the left ischial tuberosity coordinates (-3, 0, 0) and the right ischial tuberosity coordinates (3, 0, 0). The midpoint of the line connecting the two points coincides with the center point of the anus.

[0108] Lower margin of the pubic symphysis (P4): Serves as an anterior reference point, located approximately 6 cm directly in front of the center of the anus, with coordinates (0, 6, 0).

[0109] Soft tissue function point identification,

[0110] Anal center point (O): Automatically located through morphological analysis of anal verge folds, serving as the center of a circle for orientation marking.

[0111] Sphincter contraction center point (S): Under pelvic floor muscle electrical stimulation, the geometric center of the sphincter at maximum contraction is monitored by real-time ultrasound, and the coordinates are dynamically corrected (usually deviating from point O by <0.5cm).

[0112] Changqiang acupoint (C): The midpoint of the line connecting the tip of the coccyx (P1) and the center point of the anus (O), with coordinates (0, -2.5, 0), is used as a key target point for stimulation of the meridians in traditional Chinese medicine.

[0113] Anal margin fold apex (F1~F4): Four symmetrically distributed anal margin skin fold apexes are selected to help verify the circumferential uniformity of the orientation markers.

[0114] A clock-like orientation system is constructed, with the following orientation division rules: A clock-like orientation system is established with the anal center point (O) as the center. This clock-like orientation system is a 360° orientation system. The 12 o'clock direction points to the lower edge of the pubic symphysis (P4), corresponding to the anterior anatomical region; the 6 o'clock direction points to the tip of the coccyx (P1), corresponding to the posterior region; the 3 o'clock / 9 o'clock directions correspond to the right / left ischial tuberosities (P3 / P2) respectively, forming a transverse axis. Within this clock-like orientation system, the scale is subdivided, with each 30° increment serving as a secondary marker (e.g., 1.30, 4 o'clock, etc.) for precise local lesion location or stimulation targets.

[0115] The multimodal data registration method of this invention includes the following steps:

[0116] Image fusion: Align RGB images, infrared temperature matrix, and depth distance images using reference point coordinates to ensure that the error between virtual orientation and actual anatomical position is <1mm.

[0117] Dynamic calibration: When the patient's position changes, such as from lying on their side to lying on their back, the spatial displacement of the ischial tuberosities (P2, P3) is tracked in real time, and the rotation angle of the orientation markers is automatically corrected.

[0118] Spatiotemporal correlation was performed on multimodal data. The dynamic contraction velocity of the sphincter was calculated by tracking the displacement of key structural points in three consecutive frames of RGB color image data, generating a contraction velocity-time curve. The key structural points included the anal center point and the sphincter contraction center point among the anatomical markers on the sphincter surface. When the absolute value of the slope of the contraction velocity-time curve was >0.5 mm / s, it was identified as an abnormal marker of sphincter contraction velocity. The depth distance image data was registered with the infrared thermal imaging temperature distribution data through the anatomical markers on the sphincter surface to locate the low-temperature depression complex abnormal area, which was then used as the target point for electroacupuncture stimulation.

[0119] The low-temperature depression complex abnormality area of ​​this invention refers to the intersection of the sphincter structure depression area detected by depth imaging and the area with a temperature below 35°C shown by infrared thermography, characterizing the site of neuromuscular damage. The low-temperature depression complex abnormality area mainly represents a complex lesion site in the anal sphincter and surrounding pelvic floor tissues where there are structural and morphological abnormalities and metabolic dysfunction. This area exhibits impaired blood circulation or nerve innervation, such as sphincter muscle fiber ischemia and weakened nerve conduction function, leading to reduced metabolic activity; simultaneously, there is sphincter muscle fiber atrophy, scarring, or anatomical defects, such as postpartum tears or poor tissue repair after surgical injury. The low-temperature depression complex abnormality area is a key pathological target for anal sphincter dysfunction and fecal incontinence.

[0120] A specific example of monitoring sphincter contraction velocity will now be described.

[0121] I. Focus on the sphincter area

[0122] Input: 3 consecutive RGB images, for example, Frame1, Frame2, Frame3, frame rate 25fps, time interval Δt=0.04s.

[0123] ROI cropping: Based on the coordinates of the anal center point, a predefined binary mask is used to directly crop out the 100×100 pixel sphincter core area, shielding background interference.

[0124] Grayscale conversion and filtering: Convert the ROI to an 8-bit grayscale image, use a 5×5 mean filter to smooth the noise, and set the kernel value to 1 / 25 to preserve the sphincter edge contour.

[0125] II. Feature point extraction, i.e., tracking and shrinking marker points

[0126] It can track four key markers on the sphincter contour, located at the clock positions of 12 o'clock, 3 o'clock, 6 o'clock, and 9 o'clock:

[0127] Edge detection: Perform Canny edge detection on the filtered ROI with a threshold of 50-150 to extract the sphincter closure contour.

[0128] Marker point positioning: 12 points: the highest point at the top of the contour (minimum y-coordinate); 6 points: the lowest point at the bottom of the contour (maximum y-coordinate); 3 points / 9 points: the widest points on the left and right sides of the contour (maximum / minimum x-coordinate).

[0129] Output the pixel coordinates (x1, y1) ~ (x4, y4) of the four marker points in Frame1, Frame2, and Frame3.

[0130] III. Motion Vector Calculation

[0131] Template block definition,

[0132] In Frame1, take a 5×5 pixel template block centered on each marker point (reducing the computational load to 1 / 4 of the original 10×10), and denote it as Block1.

[0133] Simplified block matching formula

[0134] In Frame2, a search area of ​​±10 pixels is defined, and the following calculations are performed for each candidate block Block2:

[0135]

[0136] i = 1 to 25 represents all pixels in a 5×5 block, and the average gray level difference is used as the matching basis. The smaller the value, the higher the matching degree.

[0137] Displacement acquisition

[0138] The candidate block with the smallest matching degree is selected, and its offset (Δx, Δy) relative to its original position is the pixel displacement (unit: pixels) between the two frames.

[0139] IV. Calculation and Generation of Contraction Rate-Time Curve

[0140] Pixel-to-physical scale conversion: Given camera calibration parameters 1 pixel = 0.1 mm, actual displacement:

[0141]

[0142] Speed ​​calculation: speed of adjacent frames Given Δt = 0.04s, we obtain:

[0143] Frame1→Frame2 speed:

[0144] Frame2→Frame3 speed:

[0145] Contraction speed-time curve generation: Plot points with time as the horizontal axis (0s, 0.04s, 0.08s) and v1 and v2 as the vertical axis, and connect them using linear interpolation. For example, the slope of the line segment from 0s to 0.04s is v1, and the slope of the line segment from 0.04s to 0.08s is v2.

[0146] V. Determining anomaly markers using slope thresholds

[0147] Slope calculation: Directly use the slope of the contraction speed-time curve of the speed difference between two frames:

[0148]

[0149] Anomaly detection: If |k| > the first preset threshold, mark the time period as an abnormal sphincter contraction speed and trigger an alert signal. The first preset threshold is a user-defined threshold, specifically 0.5 mm / s. 2 .

[0150] By reducing the number of feature points, simplifying the matching algorithm and curve generation, the computational load is significantly reduced while ensuring the accuracy of key motion parameters.

[0151] The formula for calculating the anal resting pressure morphological coupling coefficient K is as follows: , where P 静息 The collected anal resting pressure is P0, which is the average anal resting pressure of a normal adult. T is the collected sphincter thickness, which is T0, which is the average sphincter thickness of a normal adult. D is the proportion of the damaged area; the proportion of the damaged area D is the ratio of the area of ​​the damaged area to the area of ​​the core sphincter region.

[0152] The stimulation phase of the stimulation unit is determined based on the anal resting pressure morphological coupling coefficient, and the stimulation unit adjusts the electroacupuncture stimulation parameters in real time based on the characteristic parameters of the stimulation phase; the stimulation phase includes a repair stagnation phase, a repair initiation phase, and a repair acceleration phase; the characteristic parameters also include the proportion of the damaged area D, the temperature gradient of the damaged area G, and the sphincter contraction velocity slope k. b

[0153] When the anal resting pressure morphological coupling coefficient K < 0.3 and the proportion of the damaged area D > 50%, the stimulation stage is determined to be in the repair stagnation stage; the electroacupuncture stimulation parameters are adjusted to: frequency 8-10Hz, intensity 2.5-3mA, target point is Changqiang acupoint and acupoint about 1-2cm directly below the anal margin, and the duration of a single stimulation is 30 minutes.

[0154] When 0.3 ≤ anal resting pressure morphological coupling coefficient K < 0.6 and the temperature gradient of the damaged area G < 0.3℃ / mm, the stimulation stage is determined to be in the repair initiation stage; the electroacupuncture stimulation parameters are adjusted to: frequency 20-25Hz, intensity 1.5-2mA, target point is low temperature depression complex abnormal area, and the duration of a single stimulation is extended to 40 minutes.

[0155] When the anal resting pressure morphological coupling coefficient K ≥ 0.6, the sphincter contraction velocity slope k b If the anal sphincter pressure is ≥-0.5 mm / s² and the anal sphincter pressure is ≥120 mmHg, the stimulation phase is determined to be in the repair acceleration phase. The electroacupuncture stimulation parameters are adjusted to: frequency 30-35 Hz, intensity 1-1.5 mA, and the target points are the three acupoints of Zhongliao, Xialiao and Huiyang, which are soft tissue functional points. The duration of a single stimulation session is shortened to 20 minutes.

[0156] The present invention provides a nursing system for assisting defecation, a preferred embodiment of which includes the following:

[0157] A multimodal data acquisition unit is used to acquire multimodal data of the anal sphincter region, including image data and biosignal data.

[0158] The feature parameter extraction unit is used to segment the acquired image data using a convolutional neural network segmentation model to generate a binary mask of the anal sphincter region, crop the sphincter core region in the multimodal data according to the binary mask, and extract the feature parameters of the sphincter core region, including sphincter thickness and sphincter continuity.

[0159] The stimulation strategy control unit is used to register the sphincter core region after binary masking and the temperature matrix image through anatomical markers on the sphincter surface to obtain the temperature matrix of the sphincter core region, calculate the gradient value of the temperature matrix to obtain the temperature gradient of the temperature data damage area, and calculate the anal resting pressure morphological coupling coefficient based on the sphincter thickness, sphincter continuity and anal resting pressure in the biosignal data.

[0160] The stimulation unit is used to adjust the electroacupuncture stimulation parameters of the electrical stimulation terminal according to the coupling coefficient.

[0161] The present invention provides a nursing method and system for assisting defecation, which achieves the following technical effects:

[0162] 1. Significantly improves computational efficiency and enables real-time monitoring.

[0163] By simplifying key technical aspects, such as tracking only four feature points, the computational load is reduced, and the processing time per frame is shortened, thus meeting real-time requirements. This also lowers the hardware computing power requirements, eliminating the need for high-performance GPUs or dedicated image processing chips; ordinary embedded devices can operate on this basis, reducing equipment costs and facilitating its adoption in primary healthcare institutions, community hospitals, and home settings. It can be used for long-term dynamic monitoring of functional disorders.

[0164] 2. Balancing detection accuracy and robustness to reduce clinical misdiagnosis.

[0165] Accuracy Guarantee: By employing the "ROI Focusing + Edge Feature Point Tracking" strategy, the errors in sphincter contraction displacement extraction and velocity calculation are reduced, meeting the requirements for determining clinical diagnostic thresholds.

[0166] Anti-interference capability: Mean filtering preprocessing and block matching search range are limited to ±10 pixels, effectively suppressing interference such as skin wrinkles and lighting changes, and significantly improving the accuracy of anomaly marking.

[0167] 3. Simplify clinical procedures and improve treatment precision.

[0168] High degree of automation: No manual annotation of feature points is required, and binary masking automatically cropes ROI. Doctors only need to start the device to complete data collection and analysis, greatly reducing operation time.

[0169] Precise target localization: By registering depth images with anatomical markers and infrared thermography, the localization error of the low-temperature depression complex abnormal area is ensured, the hit rate of electrical stimulation target is improved, and the damage of ineffective stimulation to normal tissues is reduced.

[0170] 4. Multimodal data fusion optimization supports personalized treatment.

[0171] By integrating RGB dynamic functional data, contraction speed and structural-temperature static data, including complex regions such as deep indentations and low-temperature zones, a multi-dimensional correlation between functional abnormalities, anatomical abnormalities, and metabolic abnormalities is achieved. This provides a quantitative basis for the individualized adjustment of electrical stimulation parameters, including frequency, intensity, and duration of action, thereby improving the effectiveness of clinical treatment.

[0172] The above description illustrates preferred embodiments of the present invention and helps those skilled in the art to more fully understand the technical solution of the present invention. However, these embodiments are merely illustrative and should not be construed as limiting the specific implementation of the present invention to these embodiments. For those skilled in the art, several simple deductions and modifications can be made without departing from the inventive concept, and all such modifications should be considered within the protection scope of the present invention.

Claims

1. A nursing system for assisting defecation, characterized in that, include: A multimodal data acquisition unit is used to acquire multimodal data of the anal sphincter region, including image data and biosignal data. The feature parameter extraction unit is used to segment the acquired image data using a convolutional neural network segmentation model to generate a binary mask of the anal sphincter region, crop the sphincter core region in the multimodal data according to the binary mask, and extract the feature parameters of the sphincter core region, including sphincter thickness and sphincter continuity. The stimulation strategy control unit is used to register the sphincter core region after binary masking and the temperature matrix image through anatomical markers on the sphincter surface to obtain the temperature matrix of the sphincter core region, calculate the gradient value of the temperature matrix to obtain the temperature gradient of the temperature data damage area, and calculate the anal resting pressure morphological coupling coefficient based on the sphincter thickness, sphincter continuity and anal resting pressure in the biosignal data. The stimulation unit is used to adjust the electroacupuncture stimulation parameters of the electrical stimulation terminal according to the coupling coefficient. The formula for calculating the anal resting pressure morphological coupling coefficient K is as follows: , where P 静息 The collected anal resting pressure is P0, which is the average anal resting pressure in normal adults. T is the collected sphincter thickness, which is T0, which is the average sphincter thickness in normal adults. D is the proportion of the damaged area; the proportion of the damaged area D is the ratio of the area of ​​the damaged area to the area of ​​the sphincter core area. The stimulation phase of the stimulation unit is determined based on the anal resting pressure morphological coupling coefficient, and the stimulation unit adjusts the electroacupuncture stimulation parameters in real time based on the characteristic parameters of the stimulation phase; the stimulation phase includes a repair stagnation phase, a repair initiation phase, and a repair acceleration phase; the characteristic parameters also include the proportion of the damaged area D, the temperature gradient of the damaged area G, and the sphincter contraction velocity slope k. b .

2. The nursing system for assisting defecation according to claim 1, characterized in that, The image data includes RGB color image data of the anal sphincter region, a temperature matrix formed by infrared thermal imaging temperature distribution data, and depth distance image data; the biosignal data also includes anal contraction pressure.

3. The nursing system for assisting defecation according to claim 1, characterized in that, Using the center point of the anus as the center, virtual orientation markers are established based on bony anatomical reference points and soft tissue functional points, following the orientation of a clock. The bony anatomical reference points include the tip of the coccyx, the ischial tuberosity, and the lower edge of the pubic symphysis. The soft tissue functional points include the center point of the anus, the center point of sphincter contraction, the Changqiang acupoint at the midpoint of the line connecting the tip of the coccyx and the center point of the anus, and the apex of the anal margin fold.

4. The nursing system for assisting defecation according to claim 3, characterized in that, It also includes spatiotemporal correlation of multimodal data, calculating the dynamic contraction speed of the sphincter by tracking the displacement of key structural points in three consecutive frames of RGB color image data, and generating a contraction speed-time curve. The key structural points include the anal center point and the sphincter contraction center point among the anatomical markers on the sphincter surface. When the absolute value of the slope of the contraction speed-time curve is >0.5mm / s, it is determined to be an abnormal marker of sphincter contraction speed. The depth distance image data is registered with the infrared thermal imaging temperature distribution data through the anatomical markers on the sphincter surface to locate the low temperature depression complex abnormal area, and the low temperature depression complex abnormal area is used as the target point for electroacupuncture stimulation.

5. The nursing system for assisting defecation according to claim 1, characterized in that, When the anal resting pressure morphological coupling coefficient K < 0.3 and the proportion of the damaged area D > 50%, the stimulation phase is in the repair stagnation phase; adjust the electroacupuncture stimulation parameters as follows: frequency 8-10Hz, intensity 2.5-3mA, target point is Changqiang acupoint and acupoint 1-2cm directly below the anal margin, single stimulation duration 30 minutes.

6. The nursing system for assisting defecation according to claim 1, characterized in that, When 0.3 ≤ anal resting pressure morphological coupling coefficient K < 0.6 and the temperature gradient of the damaged area G < 0.3℃ / mm, the stimulation stage is determined to be in the repair initiation stage; the electroacupuncture stimulation parameters are adjusted to: frequency 20-25Hz, intensity 1.5-2mA, target point is low temperature depression complex abnormal area, and the duration of a single stimulation is extended to 40 minutes.

7. The nursing system for assisting defecation according to claim 1, characterized in that, When the anal resting pressure morphological coupling coefficient K ≥ 0.6, the sphincter contraction velocity slope k b If the anal sphincter pressure is ≥-0.5 mm / s² and the anal sphincter pressure is ≥120 mmHg, the stimulation phase is determined to be in the repair acceleration phase. The electroacupuncture stimulation parameters are adjusted to: frequency 30-35 Hz, intensity 1-1.5 mA, and the target points are the three acupoints of Zhongliao, Xialiao and Huiyang, which are soft tissue functional points. The duration of a single stimulation session is shortened to 20 minutes.

Images