A method and system for segmenting and extracting features of intraluminal structures of the urinary system

Abstract

The application discloses a kind of urinary system intracavity structure feature segmentation extraction method and system, it is related to image segmentation technical field, the method includes: obtaining urinary intracavity image, and synchronously obtaining acquisition setting parameter;To the urinary intracavity image is extracted, obtains mucosa texture feature parameter and image feature parameter;Based on mucosa texture feature parameter, carry out abnormal identification, and to urinary intracavity image is segmented, obtains initial segmentation result set;The effectiveness parameter of initial segmentation result is calculated and obtained;Based on effectiveness parameter, in combination with image feature parameter, acquisition setting parameter is optimized, and optimization acquisition parameter is obtained;Urinary intracavity image acquisition is carried out, and optimization intracavity image and optimization image acquisition parameter set are obtained and are re-identified and segmented, and optimization segmentation result is obtained.The application solves the technical problem of poor urinary system intracavity image segmentation extraction effect in the prior art.

Description

Technical Field

[0001] This invention relates to the field of image segmentation technology, specifically to a method and system for segmenting and extracting intracavitary structural features of the urinary system. Background Technology

[0002] In the field of intraluminal image processing of the urinary system, existing technologies typically employ fixed acquisition parameters for image acquisition and rely on a single image segmentation method for anomaly identification of mucosal structures. However, the intraluminal environment of the urinary tract is complex and variable. Differences in mucosal conditions among different patients and variations in endoscopic equipment can lead to significant fluctuations in the clarity, contrast, and texture of acquired images. Traditional methods often neglect the coupling relationship between acquisition parameters and subsequent segmentation results, performing static segmentation after a single acquisition, which makes reliable segmentation extraction difficult. Furthermore, the lack of a self-verification mechanism for the validity of segmentation results results in poor accuracy when image quality is poor or texture features are not obvious. Summary of the Invention

[0003] This application provides a method and system for segmenting and extracting intraluminal structural features of the urinary system, which addresses the technical problem of poor image segmentation and extraction performance of the urinary system in the prior art.

[0004] In view of the above problems, this application provides a method and system for segmenting and extracting intraluminal structural features of the urinary system.

[0005] In a first aspect, this application provides a method for segmenting and extracting intraluminal structural features of the urinary system, the method comprising:

[0006] Acquire images of the urinary tract and simultaneously acquire the acquisition settings parameters;

[0007] Feature extraction is performed on the image inside the urinary cavity to obtain mucosal texture feature parameters, and image feature extraction is performed to obtain image clarity and image contrast as image feature parameters;

[0008] Based on the mucosal texture feature parameters, anomaly identification is performed, and the anomaly identification results are used to segment the image of the urinary cavity to obtain an initial segmentation result set, wherein the initial segmentation result set includes multiple initial abnormal segmentation images and initial normal segmentation images.

[0009] Calculate the deviation between the mucosal texture features of multiple initial abnormal segmentation images and the mucosal texture features of multiple normal segmentation images to obtain the validity parameters of the initial segmentation results;

[0010] Based on the validity parameters and combined with the image feature parameters, the acquisition settings parameters are optimized to obtain optimized acquisition parameters;

[0011] Using the optimized acquisition parameters, images of the urinary cavity are acquired, and optimized intracavitary images and optimized image acquisition parameter sets are obtained. Then, re-identification and segmentation are performed to obtain optimized segmentation results.

[0012] Secondly, this application provides a system for segmenting and extracting intraluminal structural features of the urinary system, comprising:

[0013] The image acquisition module is used to acquire images inside the urinary cavity and simultaneously acquire the acquisition settings parameters;

[0014] The feature extraction module is used to extract features from the image inside the urinary cavity, obtain mucosal texture feature parameters, and perform image feature extraction to obtain image clarity and image contrast as image feature parameters.

[0015] The initial image segmentation module is used to perform anomaly identification based on the mucosal texture feature parameters, obtain the anomaly identification result, perform image segmentation on the urinary cavity image, and obtain an initial segmentation result set, wherein the initial segmentation result set includes multiple initial abnormal segmentation images and initial normal segmentation images;

[0016] The validity analysis module is used to calculate the deviation between the mucosal texture features of multiple initial abnormal segmentation images and the mucosal texture features of multiple normal segmentation images, and to obtain the validity parameters of the initial segmentation results.

[0017] The acquisition parameter optimization module is used to optimize the acquisition setting parameters based on the validity parameters and the image feature parameters to obtain optimized acquisition parameters;

[0018] The optimized segmentation module is used to acquire images within the urinary cavity using the optimized acquisition parameters, obtain optimized intracavitary images and optimized image acquisition parameter sets, and then perform re-identification and segmentation to obtain optimized segmentation results.

[0019] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0020] This application proposes a method and system for segmenting and extracting structural features within the urinary system cavity. By simultaneously acquiring acquisition settings and intracavitary images, mucosal texture features and image quality features are extracted for anomaly identification and preliminary segmentation. The texture deviation between abnormal and normal regions and the proportion of abnormalities are calculated to quantify the effectiveness of the segmentation results. Furthermore, the acquisition settings are dynamically optimized based on effectiveness parameters and image feature parameters. Finally, the optimized parameters are used to re-acquire images and perform re-segmentation, significantly improving the accuracy and stability of abnormal structure segmentation in intracavitary images. Compared to traditional methods, the technical solution provided in this application significantly enhances adaptability to different intracavitary environments and image quality fluctuations. Through self-verification of segmentation effectiveness and closed-loop feedback optimization of acquisition parameters, it effectively reduces the poor quality of segmentation extraction results caused by image blurring, insufficient contrast, or unclear texture features. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating a method for segmenting and extracting intracavitary structural features of the urinary system, provided in an embodiment of this application.

[0023] Figure 2 This is a schematic diagram of a system for segmenting and extracting intracavitary structural features of the urinary system, provided in an embodiment of this application.

[0024] The components represented by each number in the attached diagram are explained below:

[0025] Image acquisition module 100, feature extraction module 200, initial image segmentation module 300, validity analysis module 400, acquisition parameter optimization module 500, and optimized segmentation module 600. Detailed Implementation

[0026] This application provides a method and system for segmenting and extracting intraluminal structural features of the urinary system, which addresses the technical problem of poor image segmentation and extraction performance of the urinary system in existing technologies.

[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0028] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0029] Example 1, as Figure 1 As shown, this application provides a method for segmenting and extracting structural features within the urinary system, wherein the method includes:

[0030] S10: Acquire images of the urinary tract and simultaneously acquire acquisition settings parameters.

[0031] In the current process of acquiring images within the urinary cavity, the parameters of the imaging equipment, such as exposure gain, white balance, and shutter speed, are usually set to fixed preset values ​​or rely on manual adjustment by the operator. These parameters cannot be linked with subsequent image processing requirements, resulting in unstable image quality and loss of correlation information between parameters and images. This makes it difficult for subsequent segmentation algorithms to adapt to individual differences among different patients.

[0032] Step S10 in the method provided in this application embodiment includes:

[0033] Images of the urinary tract are acquired using an intraluminal imaging device. These images are of the mucosa and tubular structures within the urinary system.

[0034] The acquisition settings parameters of the imaging device when acquiring the image of the urinary cavity include exposure gain, white balance parameters, and shutter speed.

[0035] In this embodiment of the application, an image of the urinary cavity is acquired, and the acquisition setting parameters are acquired simultaneously.

[0036] Specifically, firstly, images of the urinary tract are acquired using an intra-urinary tract imaging device, which includes images of the mucosa and luminal structures within the urinary system. The images are acquired using an intra-urinary tract imaging device, such as a cystoscope, and output as digital images.

[0037] Furthermore, the imaging device's acquisition settings parameters are used when acquiring the images within the urinary cavity. These acquisition settings parameters include exposure gain, white balance parameters, and shutter speed. Specifically, while acquiring each frame of image, the current acquisition settings parameters are read from the imaging device's control interface. These parameters include exposure gain, white balance parameters, and shutter speed. Exposure gain controls the amplification factor of the light signal by the image sensor, with an exemplary range of 0.5-2.0; a higher value results in a brighter image. The white balance parameter corrects color deviations at different color temperatures and is expressed as a red-blue gain ratio, for example, a red gain of 1.2 and a blue gain of 0.8. Shutter speed controls the duration of light accumulation for each exposure, with an exemplary range of 5 milliseconds to 50 milliseconds.

[0038] By synchronously acquiring intracavitary images and corresponding acquisition settings, a basis for subsequent parameter optimization was provided, and a traceable relationship between image quality and acquisition conditions was established.

[0039] S20: Perform feature extraction on the image inside the urinary cavity to obtain mucosal texture feature parameters, and perform image feature extraction to obtain image clarity and image contrast as image feature parameters.

[0040] Raw images of the urinary tract contain a lot of noise and irrelevant information, and direct abnormal segmentation is easily affected by factors such as uneven illumination and mucosal surface reflection.

[0041] Step S20 in the method provided in this application embodiment includes:

[0042] The image inside the urinary cavity is converted to grayscale, the LBP value of the image inside the urinary cavity is calculated, and the mucosal texture feature parameters are obtained.

[0043] Image clarity is obtained based on the average grayscale change amplitude of adjacent pixels after the image inside the urinary cavity is grayscaled.

[0044] Calculate the grayscale deviation between the brightest and darkest pixels in the image of the urinary cavity to obtain the contrast.

[0045] By integrating the image sharpness and contrast, image feature parameters are obtained.

[0046] In this embodiment of the application, feature extraction is performed on the image inside the urinary cavity to obtain mucosal texture feature parameters, and image feature extraction is performed to obtain image clarity and image contrast as image feature parameters.

[0047] Specifically, firstly, the image within the urinary cavity is converted to grayscale, and the LBP value of the image is calculated to obtain mucosal texture feature parameters. For example, the color image within the urinary cavity acquired in step S10 is converted to a grayscale image. When converting the color image to grayscale, the red, green, and blue channel values ​​of each pixel are multiplied by weighting coefficients set according to the human eye's sensitivity to different colors. Then, the three products are summed to obtain the grayscale value of the pixel. The sum of the coefficients of the red, green, and blue channels is 1, with the green channel coefficient being the largest and the blue channel coefficient the smallest. On the obtained grayscale image, the Local Binary Pattern Value (LBP) is calculated. For each central pixel in the grayscale image, its eight neighboring pixels are taken as adjacent pixels. Using the grayscale value of the central pixel itself as a threshold, the grayscale value of each adjacent pixel is compared with this threshold: if the grayscale value of an adjacent pixel is greater than or equal to the threshold, the corresponding binary bit is recorded as 1; if it is less than the threshold, it is recorded as 0. Arrange the binary bits in a clockwise or counterclockwise direction to form a binary number, then convert it to a decimal number to obtain the LBP value of the center pixel. After traversing all pixels of the entire image, calculate the distribution histogram of all LBP values, and use this histogram as the mucosal texture feature vector, i.e., the mucosal texture feature parameter.

[0048] Furthermore, the image sharpness is obtained based on the average grayscale change amplitude of adjacent pixels after the image within the urinary cavity is grayscaled. For example, the absolute value of the grayscale difference between each pixel and its right-hand neighbor, and the absolute value of the grayscale difference between each pixel and its lower neighbor are calculated. The two absolute values ​​are added together to obtain the total change amplitude of that pixel. The arithmetic mean of the total change amplitudes of all pixels in the image within the urinary cavity is then calculated, and the result is used as the image sharpness.

[0049] Further, the grayscale deviation between the brightest and darkest pixels in the urinary cavity image is calculated to obtain the contrast. The maximum and minimum grayscale values ​​among all pixels in the grayscale image are obtained, and the grayscale difference between the maximum and minimum grayscale values ​​is calculated; this difference is used as the contrast.

[0050] Furthermore, the image sharpness and contrast are integrated to obtain image feature parameters. The calculated image sharpness and contrast are combined into a two-dimensional vector, which is then output as the image feature parameters.

[0051] By extracting mucosal texture features that describe local texture changes, as well as sharpness and contrast that reflect overall image quality, multi-dimensional quantitative indicators are provided for the validity verification of subsequent anomaly identification and segmentation results.

[0052] S30: Based on the mucosal texture feature parameters, perform anomaly identification, obtain the anomaly identification result, perform image segmentation on the intra-urinary cavity image, and obtain an initial segmentation result set, wherein the initial segmentation result set includes multiple initial abnormal segmentation images and initial normal segmentation images.

[0053] Traditional threshold segmentation struggles to distinguish subtle abnormalities in mucosal texture, easily misjudging normal folds as lesions or missing early lesion areas, resulting in poor segmentation and extraction performance.

[0054] Step S30 in the method provided in this application embodiment includes:

[0055] An anomaly detection model is constructed, wherein the anomaly detection model is trained using historical mucosal texture feature parameters and historical anomaly detection results;

[0056] The mucosal texture feature parameters are input into the anomaly recognition model, and the anomaly recognition result is output, wherein the anomaly recognition result includes the anomaly type and anomaly coordinates;

[0057] Based on the anomaly identification results, the image inside the urinary cavity is segmented, and the area with abnormal mucosal texture is segmented into an initial abnormal segmentation image, and the area with normal mucosal texture is segmented into an initial normal segmentation image.

[0058] All initial abnormal segmentation images are integrated with the initial normal segmentation images to form an initial segmentation result set.

[0059] In this embodiment of the application, anomaly identification is performed based on the mucosal texture feature parameters, and the anomaly identification results are used to segment the image of the urinary cavity to obtain an initial segmentation result set, wherein the initial segmentation result set includes multiple initial abnormal segmentation images and initial normal segmentation images.

[0060] Specifically, firstly, an anomaly recognition model is constructed, which is trained using historical mucosal texture feature parameters and historical anomaly recognition results. For example, an initial anomaly recognition model is constructed based on a neural network. The number of nodes in the input layer is set to 256, corresponding to the number of statistical intervals in the LBP histogram in step S20, i.e., the LBP value from 0 to 255 is evenly divided into 256 intervals, with the pixel count of each interval serving as an input feature. The hidden layer has one layer with 128 nodes, and the activation function is a linear rectified function. The output layer has two branches: the first branch is the classification output with three nodes, corresponding to "suspicious lesion area," "suspicious stone area," and "suspicious polyp area," respectively, and the activation function is softmax; the second branch is the regression output with four nodes, corresponding to the relative values ​​of the horizontal and vertical coordinates of the abnormal area's center point, the relative value of the area's width, and the relative value of the area's height (all ranging from 0 to 1), and the activation function is a linear function. The training method for the model is as follows: 1000 historical images of the urinary cavity were collected, and the relative values ​​of the x-coordinate, y-coordinate, width, and height of the center point of the abnormal region were labeled on each image in the form of bounding boxes, along with the suspected type of the region, such as "suspicious lesion area," "suspicious stone area," and "suspicious polyp area," serving as the historical abnormality identification results. Simultaneously, the mucosal texture feature parameters of each image were extracted. The labeled data were used as training samples, and the backpropagation algorithm was used to train the model. The learning rate was set to 0.001, the batch size to 32, and the model was iteratively trained for 100 epochs until the loss value converged, thus completing the construction of the abnormality identification model.

[0061] Further, the mucosal texture feature parameters are input into the anomaly recognition model, and the anomaly recognition result is output. The anomaly recognition result includes the anomaly type and anomaly coordinates. The mucosal texture feature parameters are input into the anomaly recognition model. After forward computation, the model outputs the anomaly recognition result: the anomaly type branch outputs three probability values, and the category with the highest probability is taken as the identified anomaly type, such as "suspicious lesion area"; the coordinate branch outputs four values, such as the center point's x-coordinate 0.6, y-coordinate 0.4, width 0.3, and height 0.2, indicating that in the image's relative coordinate system, the anomaly area is located slightly to the right of the image center, covering approximately 30% of the width and 20% of the height.

[0062] Further, based on the anomaly identification results, the image within the urinary cavity is segmented. Regions with abnormal mucosal texture are segmented into initial anomaly segmentation images, and regions with normal mucosal texture are segmented into initial normal segmentation images. For example, image segmentation is performed on the image within the urinary cavity based on the anomaly identification results. According to the output bounding box coordinates, pixels within the corresponding rectangular region of the original image are cropped and saved as an independent initial anomaly segmentation image, with the image size being the actual pixel size corresponding to the bounding box. The remaining region in the image, excluding the bounding box, is segmented into multiple connected components, with the size of each connected component matching the initial anomaly segmentation image size. Each connected component serves as an initial normal segmentation image. If the anomaly identification results contain multiple abnormal regions, the smallest initial anomaly segmentation image size is used as the segmentation size. Both regions with abnormal mucosal texture and regions with normal mucosal texture are segmented using this size to obtain initial anomaly segmentation images and initial normal segmentation images. If the remaining region size is insufficient, the last segment is taken as the segmentation image according to its actual size.

[0063] Furthermore, all initial abnormal segmentation images are integrated with the initial normal segmentation images to form an initial segmentation result set. All generated initial abnormal segmentation images and initial normal segmentation images are combined into one set as the initial segmentation result set.

[0064] By utilizing texture structure parameters for anomaly identification and image segmentation, abnormal mucosal texture regions can be initially separated from normal regions, obtaining an initial segmentation result set containing candidate abnormal and normal regions, providing a foundation for subsequent verification.

[0065] S40: Calculate the deviation between the mucosal texture features of the multiple initial abnormal segmented images and the mucosal texture features of the multiple normal segmented images, and obtain the validity parameter of the initial segmentation result.

[0066] The initial segmentation results may contain a large number of false positives or false negatives. The lack of quantitative means to assess whether the segmentation results are reliable leads to subsequent identification relying on manual verification, which is inefficient.

[0067] Step S40 in the method provided in this application embodiment includes:

[0068] The mean of the mucosal texture features of multiple initial abnormal segmentation images is calculated as the abnormal texture mean, and the mean of the mucosal texture features of multiple initial normal segmentation images is calculated as the normal texture mean.

[0069] The absolute difference between the mean value of the abnormal texture and the mean value of the normal texture is calculated as the mucosal texture feature deviation degree.

[0070] The ratio of the number of initially abnormal segmented images to the total number of images in the initial segmentation result set is obtained as the abnormality ratio.

[0071] The mucosal texture feature deviation and the anomaly ratio are jointly verified to obtain validity parameters;

[0072] The method involves jointly verifying the mucosal texture feature deviation and the anomaly ratio to obtain validity parameters, including:

[0073] Set a deviation threshold and a ratio threshold, wherein the deviation threshold is obtained based on historical mucosal texture feature deviation statistics, and the ratio threshold is obtained based on historical abnormal ratio statistics;

[0074] The deviation degree of mucosal texture features and the abnormality ratio are jointly verified. If the deviation degree of mucosal texture features is greater than the deviation degree threshold and the abnormality ratio is greater than the ratio threshold, it is determined to be doubly valid. The validity parameter is equal to the weighted sum of the deviation degree of mucosal texture features and the abnormality ratio.

[0075] If the mucosal texture feature deviation is greater than the deviation threshold and the abnormality ratio is less than or equal to the ratio threshold, or if the mucosal texture feature deviation is less than or equal to the deviation threshold and the abnormality ratio is greater than the ratio threshold, it is determined to be a single valid value. When the mucosal texture feature deviation is greater than the deviation threshold and the abnormality ratio is less than or equal to the ratio threshold, the validity parameter is equal to half of the texture feature deviation. When the mucosal texture feature deviation is less than or equal to the deviation threshold and the abnormality ratio is greater than the ratio threshold, the validity parameter is equal to half of the abnormality ratio.

[0076] If the deviation of the mucosal texture feature is less than or equal to the deviation threshold and the abnormality ratio is less than or equal to the ratio threshold, it is determined to be invalid, and the validity parameter is set to 0.

[0077] In this embodiment of the application, the deviation between the mucosal texture features of multiple initial abnormal segmented images and the mucosal texture features of multiple normal segmented images is calculated to obtain the validity parameters of the initial segmentation results.

[0078] Specifically, firstly, the mean of the mucosal texture features of multiple initial abnormal segmentation images is calculated as the abnormal texture mean, and the mean of the mucosal texture features of multiple initial normal segmentation images is calculated as the normal texture mean. For example, for each initial abnormal segmentation image, the method used in step S20 to calculate the mucosal texture features is employed to recalculate its local binary pattern histogram, which is then used as the mucosal texture feature vector for that image. If there are three initial abnormal segmentation images in the initial segmentation result set, each image corresponds to a 256-dimensional feature vector. When calculating the abnormal texture mean, the average value is taken for the first dimension of each of these three vectors, the average value for the second dimension, and so on up to the 256th dimension, ultimately resulting in a 256-dimensional average vector, which is used as the abnormal texture mean. The same method is used to calculate the mean of the mucosal texture features of the initial normal segmentation images, which is then used as the normal texture mean.

[0079] Further, the absolute difference between the mean of the abnormal texture and the mean of the normal texture is calculated as the mucosal texture feature deviation. For example, the first dimension value of the mean of the abnormal texture is subtracted from the first dimension value of the mean of the normal texture, and the absolute value is taken. This is repeated for the second dimension value, and so on, up to the 256th dimension. The absolute values ​​of the differences across all dimensions are summed to obtain the mucosal texture feature deviation. The larger this mucosal texture feature deviation, the more significant the texture difference between the abnormal and normal regions, and the more likely the segmentation result is to be reliable. If there is only one initial abnormal segmentation image in the initial segmentation result set, the mean of the abnormal texture is directly taken from the feature vector of that image. If there is only one initial normal segmentation image, the mean of the normal texture is directly taken from the feature vector of that image.

[0080] Further, the ratio of the number of initial abnormal segmented images to the total number of images in the initial segmentation result set is obtained as the abnormality ratio. The number of initial abnormal segmented images and the total number of all images in the initial segmentation result set are counted, and the abnormality ratio is calculated as: (Number of initial abnormal segmented images / Total number of all images in the initial segmentation result set).

[0081] Furthermore, the deviation degree of the mucosal texture feature and the abnormality ratio are jointly verified to obtain validity parameters.

[0082] Specifically, firstly, a deviation threshold and a proportion threshold are set. The deviation threshold is obtained based on historical mucosal texture feature deviation statistics, and the proportion threshold is obtained based on historical abnormal proportion statistics. For example, the deviation threshold is calculated by performing the same processing on historical intraurinary tract images, calculating the mucosal texture feature deviation of each image, and then taking the median of these deviations as the threshold; the proportion threshold is similarly taken as the median of the abnormal proportion in historical images.

[0083] Furthermore, the deviation degree and abnormality ratio of mucosal texture features are jointly verified. If the deviation degree of mucosal texture features is greater than the deviation threshold and the abnormality ratio is greater than the ratio threshold, it is determined to be doubly valid. The validity parameter is equal to the weighted sum of the deviation degree of mucosal texture features and the abnormality ratio. For example, the preset deviation threshold is 0.5 and the ratio threshold is 0.3. The obtained mucosal texture feature deviation degree is 0.8 and the abnormality ratio is 0.4. Since 0.8 is greater than the deviation threshold of 0.5 and 0.4 is greater than the ratio threshold of 0.3, it is determined to be doubly valid. The validity parameter is the weighted sum of the deviation degree of mucosal texture features and the abnormality ratio. Assuming that the weights of both are 0.5, then the validity parameter = 0.8 × 0.5 + 0.4 × 0.5 = 0.6. In practical applications, the weights of the deviation degree of mucosal texture features and the abnormality ratio can be obtained based on the actual scenario.

[0084] Further, if the mucosal texture feature deviation is greater than the deviation threshold and the abnormality ratio is less than or equal to the ratio threshold, or if the mucosal texture feature deviation is less than or equal to the deviation threshold and the abnormality ratio is greater than the ratio threshold, it is determined to be a single valid value. When the mucosal texture feature deviation is greater than the deviation threshold and the abnormality ratio is less than or equal to the ratio threshold, the validity parameter is equal to half of the texture feature deviation. When the mucosal texture feature deviation is less than or equal to the deviation threshold and the abnormality ratio is greater than the ratio threshold, the validity parameter is equal to half of the abnormality ratio. For example, if the current deviation is 0.8 and the abnormality ratio is 0.2, then the deviation is greater than 0.5 and the abnormality ratio is less than or equal to 0.3, which is determined to be a single valid value, i.e., the deviation is valid and the abnormality ratio is invalid, and the validity parameter is equal to half of the deviation, i.e., validity parameter = 0.8 / 2 = 0.4.

[0085] Furthermore, if the deviation of the mucosal texture feature is less than or equal to the deviation threshold and the anomaly ratio is less than or equal to the ratio threshold, it is determined to be invalid, and the validity parameter is set to 0. For example, if the current deviation is 0.3 and the anomaly ratio is 0.2, both of which are less than or equal to their respective thresholds, it is determined to be invalid, and the validity parameter is set to 0.

[0086] A large deviation indicates that the texture features of the abnormal region are significantly different from those of the normal region, and the segmented abnormal region has high discriminative power at the texture level, making the segmentation result more likely to be correct. A small deviation indicates that the abnormal texture is similar to the normal texture, and the segmentation may be misjudged. On the other hand, the abnormality ratio reflects the coverage of the abnormal region in the image. If the abnormality ratio is moderate or high, it indicates a higher probability of actual anomalies; if the abnormality ratio is too low, it may be noise or a false positive. When both the deviation and the abnormality ratio exceed their respective statistical thresholds, the evidence from the two dimensions corroborates each other, and is judged as doubly valid. In this case, the weighted sum of the two is directly used as the validity parameter to comprehensively reflect the overall credibility of the segmentation. When only one dimension exceeds the threshold, it indicates that there is only unilateral evidence, and the credibility is halved; therefore, the valid value is taken as half of the contribution value of that dimension. When neither dimension exceeds the threshold, it indicates a lack of any valid evidence, the segmentation result is unreliable, and the validity parameter is directly set to zero. This avoids misjudgment based on a single indicator and ensures reasonable quantification under different evidence strengths.

[0087] S50: Based on the validity parameters and combined with the image feature parameters, optimize the acquisition settings parameters to obtain optimized acquisition parameters.

[0088] When the initial segmentation results are of low effectiveness or the image quality is poor, existing technologies cannot automatically adjust the acquisition conditions to obtain better images. They can only rely on the operator's experience to make repeated manual adjustments, which is not only time-consuming but also makes it difficult to ensure that the parameter adjustment direction is correct.

[0089] Step S50 in the method provided in this application embodiment includes:

[0090] The acquisition settings parameters are used as initial particles, and the fluctuation range is obtained based on the variance of historical acquisition settings parameters.

[0091] Using the aforementioned fluctuation range, random fluctuations are performed based on the inherent constraints of the initial particles and the acquisition setting parameters to obtain multiple initially generated particles;

[0092] The initial particles and the initially generated particles are integrated to form an initial particle swarm;

[0093] Set a particle fitness function, with the optimization objective being the maximum value of the particle fitness function and the effectiveness parameter being greater than the average of the historical effectiveness parameters. Iterate and optimize the initial particle swarm to obtain the acquisition setting parameters corresponding to the particle with the maximum particle fitness function value, and use them as the optimized acquisition parameters.

[0094] The particle fitness function is set, including:

[0095] Based on the particles in the initial particle swarm, historical intra-urinary tract images obtained using similar acquisition settings are acquired.

[0096] The historical average sharpness and historical average contrast of the historical intra-urinary tract images are used as particle sharpness and particle contrast, and the validity parameters of the historical intra-urinary tract images are used as particle validity parameters.

[0097] The weighted sum of the particle sharpness, particle contrast, and particle effectiveness parameters is calculated and set as the particle adaptability function, wherein the weights of the weighted sum are obtained by mapping based on the anomaly identification results.

[0098] In this embodiment of the application, the acquisition setting parameters are optimized based on the validity parameters and the image feature parameters to obtain optimized acquisition parameters.

[0099] Specifically, firstly, the acquisition setting parameters are used as initial particles, and the fluctuation range is obtained based on the variance of historical acquisition setting parameters. For example, the acquisition setting parameters acquired synchronously in step S10 are used as an initial particle. The variance of historical acquisition setting parameters is statistically analyzed from the historical database, and the variances of exposure gain, red-blue gain in white balance parameters, and shutter speed are calculated respectively. These three variance values ​​are used as the fluctuation ranges of the three parameters in the acquisition setting parameters.

[0100] Furthermore, using the aforementioned fluctuation range, random fluctuations are performed based on the inherent constraints of the initial particles and the acquisition setting parameters to obtain multiple initially generated particles. For example, based on the initial particles, each acquisition setting parameter is randomly fluctuated within its fluctuation range while satisfying its inherent constraints, such as the exposure gain not being lower than the device's minimum gain value nor exceeding the maximum gain value, and the shutter speed being between the device's shortest and longest allowed times. During fluctuation, a random offset is independently generated for each parameter, the offset following a normal distribution with a mean of 0 and a standard deviation equal to the fluctuation range. The offset is added to the parameter value of the initial particles; if it exceeds the constraint boundary, the boundary value is used. The random fluctuation process is repeated to generate multiple new particles, thus obtaining the initial generated particles.

[0101] Furthermore, the initial particles and the initially generated particles are integrated to form an initial particle swarm.

[0102] Furthermore, a particle fitness function is set, with the optimization objective being the maximum value of the particle fitness function and an effectiveness parameter greater than the historical average effectiveness parameter. The initial particle swarm is iteratively optimized to obtain the acquisition setting parameters corresponding to the particle with the maximum particle fitness function value, which are then used as the optimized acquisition parameters.

[0103] Specifically, the particle fitness function is set, including:

[0104] First, based on the particles in the initial particle swarm, historical intraurological images acquired using similar acquisition settings are obtained. For each particle in the initial particle swarm, all intraurological images acquired using the same acquisition settings or all acquisition settings with a deviation of less than 3% are retrieved from the historical database.

[0105] Furthermore, the historical average sharpness and historical average contrast of the historical intra-urinary cavity images are used as the particle sharpness and particle contrast, and the validity parameters of the historical intra-urinary cavity images are used as the particle validity parameters. Specifically, the arithmetic mean of the sharpness of these historical images is calculated as the particle sharpness of the particle; the arithmetic mean of the contrast of these historical images is calculated as the particle contrast of the particle; and simultaneously, the arithmetic mean of the validity parameters obtained by these historical images in their respective processing flows is calculated as the particle validity parameters of the particle.

[0106] Further, a weighted sum of the particle sharpness, particle contrast, and particle effectiveness parameter is calculated and set as the particle adaptability function. The weights of the weighted sum are obtained by mapping based on the anomaly identification results. For example, the particle sharpness, particle contrast, and particle effectiveness parameter are weighted and summed to obtain the particle's adaptability function value. The three weights are not fixed values ​​but are obtained by mapping based on the type of anomaly identification result in step S30. For example, if the anomaly identification result is "suspicious lesion area," the sharpness weight is set to 0.4, the contrast weight to 0.3, and the effectiveness parameter weight to 0.3; if the anomaly identification result is "suspicious stone area," the sharpness weight is set to 0.3, the contrast weight to 0.4, and the effectiveness parameter weight to 0.3; if the anomaly identification result is "suspicious polyp area," the sharpness weight is set to 0.3, the contrast weight to 0.3, and the effectiveness parameter weight to 0.4. The mapping relationship between the weights and the anomaly identification results is determined in advance through expert evaluation.

[0107] Further, with the optimization objective of maximizing the particle fitness function value and ensuring the effectiveness parameter is greater than the historical average effectiveness parameter, the initial particle swarm is iteratively optimized to obtain the acquisition setting parameters corresponding to the particle with the largest particle fitness function value, which are then used as optimized acquisition parameters. For example, after obtaining the fitness function value of each particle, the initial particle swarm is iteratively optimized with the objective of maximizing the fitness function value and ensuring the particle's corresponding effectiveness parameter is greater than the historical average effectiveness parameter. In each iteration, the individual optimal position and global optimal position of each particle are updated based on its current fitness function value, and the position of each particle is adjusted. The iteration process continues until the improvement in the global optimal fitness function value is less than 0.1% in multiple consecutive iterations or the preset maximum number of iterations is reached. After the iteration ends, the particle with the largest fitness function value is selected from the final particle swarm, and its corresponding acquisition setting parameters are used as optimized acquisition parameters.

[0108] By using validity parameters and image feature parameters as feedback signals, closed-loop optimization is performed on acquisition parameters such as exposure gain and white balance, enabling the next acquisition to generate a clearer image based on the specific mucosal state of the current patient, thereby improving the reliability of subsequent segmentation and extraction.

[0109] S60: Using the optimized acquisition parameters, perform intra-urinary cavity image acquisition, obtain optimized intra-cavity image and optimized image acquisition parameter set, and perform re-identification and segmentation to obtain optimized segmentation results.

[0110] Initial acquisition and segmentation often result in unreliable results due to improper parameters, and re-acquiring with the same parameters will not improve the quality.

[0111] Step S60 in the method provided in this application embodiment includes:

[0112] The optimized acquisition parameters are input into the urinary system intracavitary imaging device to acquire images of the urinary cavity and obtain optimized intracavitary images;

[0113] The optimized intracavitary image is subjected to feature extraction, anomaly detection, image segmentation, and validity parameter calculation to obtain the re-identification segmentation result and the corresponding re-identification validity parameter.

[0114] The re-identification validity parameters are compared with the validity parameters of the initial segmentation result. At the same time, the image feature parameters of the optimized intracavitary image are compared with the image feature parameters of the initial intracavitary image. If the re-identification validity parameters are higher than the initial validity parameters, and the clarity and contrast of the optimized intracavitary image are higher than those of the initial image, the re-identification segmentation result is output as the optimized segmentation result.

[0115] In this embodiment of the application, the optimized acquisition parameters are used to acquire images inside the urinary cavity, obtain optimized intracavitary images and optimized image acquisition parameter sets, and then perform re-identification and segmentation to obtain optimized segmentation results.

[0116] Specifically, firstly, the optimized acquisition parameters are input into the urinary system intracavitary imaging device to acquire images within the urinary cavity, thus obtaining an optimized intracavitary image. The optimized acquisition parameters obtained in step S50 are then written into the urinary system intracavitary imaging device through its control interface. Image acquisition is restarted to acquire a new intracavitary image, which serves as the optimized intracavitary image.

[0117] Further, feature extraction, anomaly detection, image segmentation, and validity parameter calculation are performed on the optimized intracavitary image to obtain the re-identification segmentation result and the corresponding re-identification validity parameter. Specifically, steps S20, S30, and S40 are repeated to extract the mucosal texture features and image feature parameters of the optimized intracavitary image, perform anomaly detection and image segmentation, obtain the re-identification segmentation result, and calculate the re-identification validity parameter corresponding to the re-identification segmentation result.

[0118] Further, the re-identification validity parameter is compared with the validity parameter of the initial segmentation result, and the image feature parameters of the optimized intracavitary image are compared with the image feature parameters of the initial intracavitary image. If the re-identification validity parameter is higher than the initial validity parameter, and the clarity and contrast of the optimized intracavitary image are both higher than those of the initial image, the re-identification segmentation result is output as the optimized segmentation result. For example, the re-identification validity parameter is compared with the validity parameter of the initial segmentation result obtained in step S40, and the clarity and contrast of the optimized intracavitary image are compared with those of the intracavitary image in step S10. If the re-identification validity parameter is higher than the validity parameter, and the clarity and contrast of the optimized intracavitary image are both higher than those of the initial image, then the optimization is deemed effective, and the re-identification segmentation result is output as the final optimized segmentation result. If none of the above conditions are met, the initial segmentation result is retained, and steps S50 and S60 are repeated according to the system settings until the conditions are met or the maximum number of iterations is reached. For example, the maximum number of iterations can be set to 10.

[0119] By re-acquiring images with optimized parameters and re-executing the complete recognition and segmentation process, higher-quality segmentation results with clearer anomaly boundaries can be obtained. At the same time, the optimized parameters are recorded for subsequent reference, achieving iterative improvement from low-quality images to high-quality segmentation results.

[0120] Example 2, as Figure 2As shown, based on the same inventive concept as the method for segmenting and extracting intraluminal structural features of the urinary system provided in Embodiment 1, this embodiment of the invention also provides a system for segmenting and extracting intraluminal structural features of the urinary system, comprising:

[0121] The image acquisition module 100 is used to acquire images inside the urinary cavity and simultaneously acquire acquisition setting parameters;

[0122] The feature extraction module 200 is used to extract features from the image inside the urinary cavity, obtain mucosal texture feature parameters, and extract image features to obtain image clarity and image contrast as image feature parameters.

[0123] The initial image segmentation module 300 is used to perform anomaly identification based on the mucosal texture feature parameters, obtain anomaly identification results to perform image segmentation on the intra-urinary cavity image, and obtain an initial segmentation result set, wherein the initial segmentation result set includes multiple initial abnormal segmentation images and initial normal segmentation images;

[0124] The validity analysis module 400 is used to calculate the deviation between the mucosal texture features of multiple initial abnormal segmentation images and the mucosal texture features of multiple normal segmentation images, and to obtain the validity parameters of the initial segmentation results.

[0125] The acquisition parameter optimization module 500 is used to optimize the acquisition setting parameters based on the validity parameters and the image feature parameters to obtain optimized acquisition parameters;

[0126] The optimized segmentation module 600 is used to acquire images within the urinary cavity using the optimized acquisition parameters, obtain optimized intracavitary images and optimized image acquisition parameter sets, and then perform re-identification and segmentation to obtain optimized segmentation results.

[0127] In one embodiment, the image acquisition module 100 is further configured to:

[0128] Images of the urinary tract are acquired using an intraluminal imaging device. These images are of the mucosa and tubular structures within the urinary system.

[0129] The acquisition settings parameters of the imaging device when acquiring the image of the urinary cavity include exposure gain, white balance parameters, and shutter speed.

[0130] In one embodiment, the feature extraction module 200 is further configured to:

[0131] The image inside the urinary cavity is converted to grayscale, the LBP value of the image inside the urinary cavity is calculated, and the mucosal texture feature parameters are obtained.

[0132] Image clarity is obtained based on the average grayscale change amplitude of adjacent pixels after the image inside the urinary cavity is grayscaled.

[0133] Calculate the grayscale deviation between the brightest and darkest pixels in the image of the urinary cavity to obtain the contrast.

[0134] By integrating the image sharpness and contrast, image feature parameters are obtained.

[0135] In one embodiment, the initial image segmentation module 300 is further configured to:

[0136] An anomaly detection model is constructed, wherein the anomaly detection model is trained using historical mucosal texture feature parameters and historical anomaly detection results;

[0137] The mucosal texture feature parameters are input into the anomaly recognition model, and the anomaly recognition result is output, wherein the anomaly recognition result includes the anomaly type and anomaly coordinates;

[0138] Based on the anomaly identification results, the image inside the urinary cavity is segmented, and the area with abnormal mucosal texture is segmented into an initial abnormal segmentation image, and the area with normal mucosal texture is segmented into an initial normal segmentation image.

[0139] All initial abnormal segmentation images are integrated with the initial normal segmentation images to form an initial segmentation result set.

[0140] In one embodiment, the validity analysis module 400 is further configured to:

[0141] The mean of the mucosal texture features of multiple initial abnormal segmentation images is calculated as the abnormal texture mean, and the mean of the mucosal texture features of multiple initial normal segmentation images is calculated as the normal texture mean.

[0142] The absolute difference between the mean value of the abnormal texture and the mean value of the normal texture is calculated as the mucosal texture feature deviation degree.

[0143] The ratio of the number of initially abnormal segmented images to the total number of images in the initial segmentation result set is obtained as the abnormality ratio.

[0144] The mucosal texture feature deviation and the anomaly ratio are jointly verified to obtain validity parameters;

[0145] The method involves jointly verifying the mucosal texture feature deviation and the anomaly ratio to obtain validity parameters, including:

[0146] Set a deviation threshold and a ratio threshold, wherein the deviation threshold is obtained based on historical mucosal texture feature deviation statistics, and the ratio threshold is obtained based on historical abnormal ratio statistics;

[0147] The deviation degree of mucosal texture features and the abnormality ratio are jointly verified. If the deviation degree of mucosal texture features is greater than the deviation degree threshold and the abnormality ratio is greater than the ratio threshold, it is determined to be doubly valid. The validity parameter is equal to the weighted sum of the deviation degree of mucosal texture features and the abnormality ratio.

[0148] If the mucosal texture feature deviation is greater than the deviation threshold and the abnormality ratio is less than or equal to the ratio threshold, or if the mucosal texture feature deviation is less than or equal to the deviation threshold and the abnormality ratio is greater than the ratio threshold, it is determined to be a single valid value. When the mucosal texture feature deviation is greater than the deviation threshold and the abnormality ratio is less than or equal to the ratio threshold, the validity parameter is equal to half of the texture feature deviation. When the mucosal texture feature deviation is less than or equal to the deviation threshold and the abnormality ratio is greater than the ratio threshold, the validity parameter is equal to half of the abnormality ratio.

[0149] If the deviation of the mucosal texture feature is less than or equal to the deviation threshold and the abnormality ratio is less than or equal to the ratio threshold, it is determined to be invalid, and the validity parameter is set to 0.

[0150] In one embodiment, the parameter acquisition optimization module 500 is further configured to:

[0151] The acquisition settings parameters are used as initial particles, and the fluctuation range is obtained based on the variance of historical acquisition settings parameters.

[0152] Using the aforementioned fluctuation range, random fluctuations are performed based on the inherent constraints of the initial particles and the acquisition setting parameters to obtain multiple initially generated particles;

[0153] The initial particles and the initially generated particles are integrated to form an initial particle swarm;

[0154] Set a particle fitness function, with the optimization objective being the maximum value of the particle fitness function and the effectiveness parameter being greater than the average of the historical effectiveness parameters. Iterate and optimize the initial particle swarm to obtain the acquisition setting parameters corresponding to the particle with the maximum particle fitness function value, and use them as the optimized acquisition parameters.

[0155] The particle fitness function is set, including:

[0156] Based on the particles in the initial particle swarm, historical intra-urinary tract images obtained using similar acquisition settings are acquired.

[0157] The historical average sharpness and historical average contrast of the historical intra-urinary tract images are used as particle sharpness and particle contrast, and the validity parameters of the historical intra-urinary tract images are used as particle validity parameters.

[0158] The weighted sum of the particle sharpness, particle contrast, and particle effectiveness parameters is calculated and set as the particle adaptability function, wherein the weights of the weighted sum are obtained by mapping based on the anomaly identification results.

[0159] In one embodiment, the optimized segmentation module 600 is further configured to:

[0160] The optimized acquisition parameters are input into the urinary system intracavitary imaging device to acquire images of the urinary cavity and obtain optimized intracavitary images;

[0161] The optimized intracavitary image is subjected to feature extraction, anomaly detection, image segmentation, and validity parameter calculation to obtain the re-identification segmentation result and the corresponding re-identification validity parameter.

[0162] The re-identification validity parameters are compared with the validity parameters of the initial segmentation result. At the same time, the image feature parameters of the optimized intracavitary image are compared with the image feature parameters of the initial intracavitary image. If the re-identification validity parameters are higher than the initial validity parameters, and the clarity and contrast of the optimized intracavitary image are higher than those of the initial image, the re-identification segmentation result is output as the optimized segmentation result.

[0163] In summary, the embodiments of this application have at least the following technical effects:

[0164] This application proposes a method and system for segmenting and extracting structural features within the urinary system cavity. By simultaneously acquiring acquisition settings and intracavitary images, mucosal texture features and image quality features are extracted for anomaly identification and preliminary segmentation. The texture deviation between abnormal and normal regions and the proportion of abnormalities are calculated to quantify the effectiveness of the segmentation results. Furthermore, the acquisition settings are dynamically optimized based on effectiveness parameters and image feature parameters. Finally, the optimized parameters are used to re-acquire images and perform re-segmentation, significantly improving the accuracy and stability of abnormal structure segmentation in intracavitary images. Compared to traditional methods, the technical solution provided in this application significantly enhances adaptability to different intracavitary environments and image quality fluctuations. Through self-verification of segmentation effectiveness and closed-loop feedback optimization of acquisition parameters, it effectively reduces the poor quality of segmentation extraction results caused by image blurring, insufficient contrast, or unclear texture features.

[0165] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0166] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0167] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for segmenting and extracting structural features within the urinary system, characterized in that, include: Acquire images of the urinary tract and simultaneously acquire the acquisition settings parameters; Feature extraction is performed on the image inside the urinary cavity to obtain mucosal texture feature parameters, and image feature extraction is performed to obtain image clarity and image contrast as image feature parameters; Based on the mucosal texture feature parameters, anomaly identification is performed, and the anomaly identification results are used to segment the image of the urinary cavity to obtain an initial segmentation result set, wherein the initial segmentation result set includes multiple initial abnormal segmentation images and initial normal segmentation images. Calculate the deviation between the mucosal texture features of multiple initial abnormal segmentation images and the mucosal texture features of multiple normal segmentation images to obtain the validity parameters of the initial segmentation results; Based on the validity parameters and combined with the image feature parameters, the acquisition settings parameters are optimized to obtain optimized acquisition parameters; Using the optimized acquisition parameters, images of the urinary cavity are acquired, and optimized intracavitary images and optimized image acquisition parameter sets are obtained. Then, re-identification and segmentation are performed to obtain optimized segmentation results.

2. The method for segmenting and extracting intraluminal structural features of the urinary system according to claim 1, characterized in that, Acquire images of the urinary tract and simultaneously acquire acquisition settings parameters, including: Images of the urinary tract are acquired using an intraluminal imaging device. These images are of the mucosa and tubular structures within the urinary system. The acquisition settings parameters of the imaging device when acquiring the image of the urinary cavity include exposure gain, white balance parameters, and shutter speed.

3. The method for segmenting and extracting intraluminal structural features of the urinary system according to claim 1, characterized in that, Feature extraction is performed on the intra-urinary tract image to obtain mucosal texture feature parameters, and image feature extraction is performed to obtain image sharpness and image contrast as image feature parameters, including: The image inside the urinary cavity is converted to grayscale, the LBP value of the image inside the urinary cavity is calculated, and the mucosal texture feature parameters are obtained. Image clarity is obtained based on the average grayscale change amplitude of adjacent pixels after the image inside the urinary cavity is grayscaled. Calculate the grayscale deviation between the brightest and darkest pixels in the image of the urinary cavity to obtain the contrast. By integrating the image sharpness and contrast, image feature parameters are obtained.

4. The method for segmenting and extracting intraluminal structural features of the urinary system according to claim 1, characterized in that, Based on the mucosal texture feature parameters, anomaly identification is performed, and the anomaly identification results are used to segment the image of the urinary cavity to obtain an initial segmentation result set. The initial segmentation result set includes multiple initial abnormal segmentation images and initial normal segmentation images, including: An anomaly detection model is constructed, wherein the anomaly detection model is trained using historical mucosal texture feature parameters and historical anomaly detection results; The mucosal texture feature parameters are input into the anomaly recognition model, and the anomaly recognition result is output, wherein the anomaly recognition result includes the anomaly type and anomaly coordinates; Based on the anomaly identification results, the image inside the urinary cavity is segmented, and the area with abnormal mucosal texture is segmented into an initial abnormal segmentation image, and the area with normal mucosal texture is segmented into an initial normal segmentation image. All initial abnormal segmentation images are integrated with the initial normal segmentation images to form an initial segmentation result set.

5. The method for segmenting and extracting intraluminal structural features of the urinary system according to claim 1, characterized in that, Calculate the deviation between the mucosal texture features of multiple initial abnormal segmentation images and the mucosal texture features of multiple normal segmentation images to obtain the validity parameters of the initial segmentation results, including: The mean of the mucosal texture features of multiple initial abnormal segmentation images is calculated as the abnormal texture mean, and the mean of the mucosal texture features of multiple initial normal segmentation images is calculated as the normal texture mean. The absolute difference between the mean value of the abnormal texture and the mean value of the normal texture is calculated as the mucosal texture feature deviation degree. The ratio of the number of initially abnormal segmented images to the total number of images in the initial segmentation result set is obtained as the abnormality ratio. The deviation of the mucosal texture features and the proportion of anomalies are jointly verified to obtain validity parameters.

6. The method for segmenting and extracting intraluminal structural features of the urinary system according to claim 5, characterized in that, The joint verification of the mucosal texture feature deviation and the anomaly ratio yields validity parameters, including: Set a deviation threshold and a ratio threshold, wherein the deviation threshold is obtained based on historical mucosal texture feature deviation statistics, and the ratio threshold is obtained based on historical abnormal ratio statistics; The deviation degree of mucosal texture features and the abnormality ratio are jointly verified. If the deviation degree of mucosal texture features is greater than the deviation degree threshold and the abnormality ratio is greater than the ratio threshold, it is determined to be doubly valid. The validity parameter is equal to the weighted sum of the deviation degree of mucosal texture features and the abnormality ratio. If the mucosal texture feature deviation is greater than the deviation threshold and the abnormality ratio is less than or equal to the ratio threshold, or if the mucosal texture feature deviation is less than or equal to the deviation threshold and the abnormality ratio is greater than the ratio threshold, it is determined to be a single valid value. When the mucosal texture feature deviation is greater than the deviation threshold and the abnormality ratio is less than or equal to the ratio threshold, the validity parameter is equal to half of the texture feature deviation. When the mucosal texture feature deviation is less than or equal to the deviation threshold and the abnormality ratio is greater than the ratio threshold, the validity parameter is equal to half of the abnormality ratio. If the deviation of the mucosal texture feature is less than or equal to the deviation threshold and the abnormality ratio is less than or equal to the ratio threshold, it is determined to be invalid, and the validity parameter is set to 0.

7. The method for segmenting and extracting intraluminal structural features of the urinary system according to claim 1, characterized in that, Based on the validity parameters and the image feature parameters, the acquisition settings parameters are optimized to obtain optimized acquisition parameters, including: The acquisition settings parameters are used as initial particles, and the fluctuation range is obtained based on the variance of historical acquisition settings parameters. Using the aforementioned fluctuation range, random fluctuations are performed based on the inherent constraints of the initial particles and the acquisition setting parameters to obtain multiple initially generated particles; The initial particles and the initially generated particles are integrated to form an initial particle swarm; A particle fitness function is set, with the optimization objective being the maximum value of the particle fitness function and an effectiveness parameter greater than the historical average effectiveness parameter. The initial particle swarm is iteratively optimized to obtain the acquisition setting parameters corresponding to the particle with the maximum particle fitness function value, which are then used as the optimized acquisition parameters.

8. The method for segmenting and extracting intraluminal structural features of the urinary system according to claim 7, characterized in that, Set the particle fitness function, including: Based on the particles in the initial particle swarm, historical intra-urinary tract images obtained using similar acquisition settings are acquired. The historical average sharpness and historical average contrast of the historical intra-urinary tract images are used as particle sharpness and particle contrast, and the validity parameters of the historical intra-urinary tract images are used as particle validity parameters. The weighted sum of the particle sharpness, particle contrast, and particle effectiveness parameters is calculated and set as the particle adaptability function, wherein the weights of the weighted sum are obtained by mapping based on the anomaly identification results.

9. The method for segmenting and extracting intraluminal structural features of the urinary system according to claim 1, characterized in that, Using the optimized acquisition parameters, images of the urinary cavity are acquired to obtain optimized intracavitary images and optimized image acquisition parameter sets. These are then re-identified and segmented to obtain optimized segmentation results, including: The optimized acquisition parameters are input into the urinary system intracavitary imaging device to acquire images of the urinary cavity and obtain optimized intracavitary images; The optimized intracavitary image is subjected to feature extraction, anomaly detection, image segmentation, and validity parameter calculation to obtain the re-identification segmentation result and the corresponding re-identification validity parameter. The re-identification validity parameters are compared with the validity parameters of the initial segmentation result. At the same time, the image feature parameters of the optimized intracavitary image are compared with the image feature parameters of the initial intracavitary image. If the re-identification validity parameters are higher than the initial validity parameters, and the clarity and contrast of the optimized intracavitary image are higher than those of the initial image, the re-identification segmentation result is output as the optimized segmentation result.

10. A system for segmenting and extracting intraluminal structural features of the urinary system, characterized in that, A method for segmenting and extracting intraluminal structural features of the urinary system according to any one of claims 1-9, the system comprising: The image acquisition module is used to acquire images inside the urinary cavity and simultaneously acquire the acquisition settings parameters; The feature extraction module is used to extract features from the image inside the urinary cavity, obtain mucosal texture feature parameters, and perform image feature extraction to obtain image clarity and image contrast as image feature parameters. The initial image segmentation module is used to perform anomaly identification based on the mucosal texture feature parameters, obtain the anomaly identification result, perform image segmentation on the urinary cavity image, and obtain an initial segmentation result set, wherein the initial segmentation result set includes multiple initial abnormal segmentation images and initial normal segmentation images; The validity analysis module is used to calculate the deviation between the mucosal texture features of multiple initial abnormal segmentation images and the mucosal texture features of multiple normal segmentation images, and to obtain the validity parameters of the initial segmentation results. The acquisition parameter optimization module is used to optimize the acquisition setting parameters based on the validity parameters and the image feature parameters to obtain optimized acquisition parameters; The optimized segmentation module is used to acquire images within the urinary cavity using the optimized acquisition parameters, obtain optimized intracavitary images and optimized image acquisition parameter sets, and then perform re-identification and segmentation to obtain optimized segmentation results.

Images