A device for determining colon polyps based on a parameterized algorithm

Abstract

The application relates to the technical field of image processing, in particular to a colon polyp image determination device based on a parameterized algorithm. A processor in the device is used to execute the following steps: obtaining a plurality of initial polyp images in a complex plane parameterized image of the inner wall surface of a colon; the complex plane parameterized image is an image of the inner wall surface of the colon after conformal parameterization; intercepting a three-dimensional polyp curved surface piece of each initial polyp image in the three-dimensional curved surface of the inner wall surface of the colon; and determining a target polyp image from the plurality of initial polyp images according to the geometric features of the three-dimensional polyp curved surface piece corresponding to each initial polyp image. The application can determine the colon polyp image, considers the global curvature of the inner wall surface of the colon to be detected, avoids the problems of detection errors caused by uneven CT density in the polyp and missed detection of the polyp image caused by the shielding of the wrinkle, and improves the accuracy of determining the colon polyp image.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and more specifically, to a device for determining colon polyp images based on a parametric algorithm. Background Technology

[0002] Computed tomography (CTC) uses a computer-aided detection (CAD) system to assist in the detection of colonic polyps. CTC often interprets images of areas with high curvature as potentially indicative of polyps.

[0003] However, curvature measurement is often hampered by the uneven density distribution of CT (Computed Tomography) within the polyp. That is, the CT density of the polyp center image is similar to that of normal tissue, while the CT density of the polyp boundary image is similar to that of the surrounding material (such as air and labeled fluid). Moreover, curvature measurement may be biased by spurious enhancement between the polyp and the labeled material. In addition, only high-curvature images are considered as images of possible polyps, without taking into account the curvature around the high-curvature region. Therefore, this method has low accuracy in identifying colonic polyps. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide a device for determining colon polyp images based on a parametric algorithm, which can determine colon polyp images and takes into account the global curvature of the colonic wall surface, avoiding problems such as detection errors caused by uneven CT density within the polyp and missed detection of polyp images due to fold occlusion, thereby improving the accuracy of determining colon polyp images.

[0005] In a first aspect, embodiments of this application provide an apparatus for determining colon polyp images based on a parametric algorithm. The apparatus includes a processor, a storage medium, and a bus. The storage medium stores machine-readable instructions executable by the processor. When the determining apparatus is running, the processor communicates with the storage medium via the bus, and the processor performs the following steps:

[0006] Multiple initial polyp images are obtained from a complex-plane parametric image of the inner wall surface of the colon; the complex-plane parametric image is an image of the inner wall surface of the colon after conformal parametricization.

[0007] Extract three-dimensional polyp surface patches from the three-dimensional curved surface of the inner wall surface of the colon for each initial polyp image;

[0008] The target polyp image is determined from multiple initial polyp images based on the geometric features of the three-dimensional polyp surface patches corresponding to each initial polyp image.

[0009] In one possible implementation, when the processor performs the acquisition of multiple initial polyp images in a complex-plane parametric image of the inner wall surface of the colon, it performs the following steps:

[0010] Obtain a complex-plane parametric image of the inner wall surface of the colon;

[0011] Detecting polyp-connected components in complex plane parameterized images;

[0012] Based on the shape parameters of each polyp's connected regions, multiple initial polyp images are obtained by filtering each polyp's connected regions.

[0013] In one possible implementation, when the processor acquires a complex-plane parametric image of the inner wall surface of the colon, it performs the following steps:

[0014] Obtain the three-dimensional surface of the inner wall of the colon;

[0015] Remove topological noise from the 3D surface to obtain the first surface of the inner wall of the colon; the first surface is a zero-genus surface.

[0016] By performing a double-layer covering construction on the first surface, a second surface with a genus of 1 is obtained; the dimension of the homology group basis of the second surface is 2;

[0017] Calculate the holomorphic 1-forms of the second surface corresponding to each homology group basis;

[0018] Integrating the holomorphic 1-form of any second surface yields a rectangular complex field;

[0019] Based on the curvature of each point in the first surface, the rectangular complex domain is rendered to obtain a parametric image of the complex plane.

[0020] In one possible implementation, when the processor performs the computation of the holomorphic 1-forms of the second surface corresponding to the basis of each homology group, it executes the following steps:

[0021] Calculate the basis of each homology group of the second surface;

[0022] Based on the homology group basis of the second surface, calculate the harmonic 1-form corresponding to each homology group basis;

[0023] Based on the harmonic 1-forms of the second surface corresponding to the basis of each homology group, calculate the holomorphic 1-forms corresponding to the basis of each homology group.

[0024] In one possible implementation, when the processor renders the rectangular complex domain based on the curvature of points in the first surface, it performs the following steps:

[0025] Color each point according to the curvature of each point in the first surface;

[0026] The rectangular complex domain is rendered based on the first colored surface.

[0027] In one possible implementation, when the processor performs the detection of polyp connected components in a complex plane parameterized image, it performs the following steps:

[0028] Binarize each pixel in the complex plane parameterized image according to the preset polyp threshold range;

[0029] Determine the polyp connected components in the binarized complex plane parameterized image.

[0030] In one possible implementation, when the processor filters each polyp connected region according to the shape parameters of each polyp connected region to obtain multiple initial polyp images, it performs the following steps:

[0031] Based on the roundness, eccentricity, and the ratio of the length to the width of the circumscribed rectangle of each polyp connected region in the complex plane parametric image, determine whether each polyp connected region is a polyp region.

[0032] If the polyp connected component is a polyp region, then the image at the location of the polyp connected component is determined as the initial polyp image.

[0033] In one possible implementation, when the processor determines the target polyp image from multiple initial polyp images based on the geometric features of the three-dimensional polyp surface patches corresponding to each initial polyp image, it performs the following steps:

[0034] The target polyp image is determined from multiple initial polyp images based on the area of ​​the 3D polyp surface patch and the ratio of the longest side to the second longest side of the directional bounding box of the 3D polyp surface patch.

[0035] In one possible implementation, after determining the target polyp image, the processor is further configured to perform the following steps:

[0036] Extract polyp texture features from each target polyp image in the complex plane parameterized image;

[0037] The polyp texture features of each target polyp image are input into the polyp classification model to obtain the polyp image detection results for each target polyp image; the polyp classification model is trained using polyp texture sample features and corresponding polyp image detection result labels;

[0038] The target polyp image that is detected as a polyp image is determined as the final polyp image.

[0039] Secondly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which is executed by a processor such as in the first aspect to perform steps executed by the processor.

[0040] This application provides a device for determining colonic polyp images based on a parametric algorithm. The device includes a processor, a storage medium, and a bus. The storage medium stores machine-readable instructions executable by the processor. When the device is running, the processor communicates with the storage medium via the bus. The processor performs the following steps: acquiring multiple initial polyp images from a complex-plane parametric image of the colon's inner wall surface; the complex-plane parametric image is a conformally parametric image of the colon's inner wall surface; extracting three-dimensional polyp surface patches from each initial polyp image on the three-dimensional curved surface of the colon's inner wall surface; and determining a target polyp image from the multiple initial polyp images based on the geometric features of the three-dimensional polyp surface patches corresponding to each initial polyp image. This application determines colonic polyp images by using multiple initial polyp images from a complex-plane parametric image of the colon's inner wall surface and the geometric features of the three-dimensional polyp surface patches on the three-dimensional curved surface of the colon's inner wall surface. This avoids detection errors caused by uneven CT density within the polyp and missed detections due to fold occlusion, thus improving the accuracy of determining colonic polyp images. Attached Figure Description

[0041] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This illustration shows a schematic diagram of a device for determining colon polyp images based on a parametric algorithm, according to an embodiment of this application.

[0043] Figure 2 This illustration shows a structural schematic diagram of a method for determining colon polyp images based on a parametric algorithm, provided in an embodiment of this application.

[0044] Figure 3 The image shown is a binarized complex plane parametric image provided in an embodiment of this application;

[0045] Figure 4 The illustration shows a flowchart of another method for determining colon polyp images based on a parametric algorithm provided in an embodiment of this application. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0047] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0048] To enable those skilled in the art to utilize the content of this application, and in conjunction with the specific application scenario of "image processing technology," the following embodiments are provided. For those skilled in the art, the general principles defined herein can be applied to other embodiments and application scenarios without departing from the spirit and scope of this application. Although this application is primarily described within the "image processing technology field," it should be understood that this is merely an exemplary embodiment.

[0049] It should be noted that the term "comprising" will be used in the embodiments of this application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0050] Reference Figure 1 The diagram shown is a structural schematic of a device for determining colon polyp images based on a parametric algorithm according to an embodiment of this application. The device includes a processor 101, a storage medium 102, and a bus 103. The storage medium 102 stores machine-readable instructions executable by the processor 101. When the detection device is running, the processor 101 and the storage medium 102 communicate via the bus 103. The storage medium 102 and the processor 101 can be general-purpose memories and processors, and are not specifically limited here.

[0051] Reference Figure 2The diagram shown is a structural schematic of a method for determining colon polyps based on a parametric algorithm, provided in an embodiment of this application. The method comprises... Figure 1 The processor 101 in the process executes the following exemplary steps according to embodiments of this application:

[0052] S201. Obtain multiple initial polyp images from a complex plane parametric image of the inner wall surface of the colon.

[0053] The complex plane parametric image is an image of the inner wall surface of the colon after conformal parametricization. Because this application performs global conformal parametricization on the image of the inner wall surface of the colon—that is, it considers the global curvature of the inner wall surface—and the conformal properties of the parametricization can preserve the shape of the fold and polyp images in the image of the inner wall surface of the colon, the accuracy of polyp image determination in this application is higher than that in the prior art.

[0054] Here, when the processor 101 executes the acquisition of multiple initial polyp images in a complex-plane parametric image of the inner wall surface of the colon, it performs the following steps:

[0055] Step 1: Obtain a parametric image of the inner wall surface of the colon in a complex plane.

[0056] a. Obtain the three-dimensional curved surface of the inner wall of the colon.

[0057] In this embodiment, the inner wall surface of the colon in an abdominal CT image is reconstructed in 3D to obtain a three-dimensional curved surface of the colon's inner wall. Commonly used reconstruction methods include 3DMAX, Maya, AutoCAD, and UG.

[0058] b. Remove topological noise from the three-dimensional surface to obtain the first surface of the inner wall of the colon; the first surface is a zero-genus surface.

[0059] In this embodiment, if there is no topological noise in the three-dimensional surface, the surface M of the inner wall surface of the colon is determined as the first surface, and the first surface M is a zero-genus surface; if there is topological noise in the three-dimensional surface, all homology group bases of surface M that are not equivalent to the boundary are calculated; the topological noise in surface M is removed according to all homology group bases that are not equivalent to the boundary to obtain the first surface, and the first surface M is a zero-genus surface.

[0060] Homology is an important concept in algebraic topology, describing the relationship between the local and global properties of a topological space. A homology is defined as a sequence of vector spaces and linear mappings, where each vector space is a chain of n-dimensional fields encircling a topological space. It describes the relationship between the local and global properties of a topological space. Homology can be used to study properties such as the connectivity and dimension of a topological space. Fundamental properties of homology include the length and exactness of homology groups, the contractility of homology groups, and the heterogeneity of homology groups.

[0061] Genus is one of the most fundamental concepts in algebraic geometry and algebraic topology. It is defined as follows: if a surface can be drawn with at most n closed curves without dividing the surface, then the genus of that surface is called n. A zero-genus surface is a surface without any genus.

[0062] c. Perform a double-layer covering construction on the first surface to obtain a second surface with a genus of 1; the dimension of the homology group basis of the second surface is 2.

[0063] d. Calculate the holomorphic 1-forms of the second surface corresponding to the basis of each homology group.

[0064] In this embodiment, the basis of each homology group of the second surface is calculated; based on the basis of each homology group of the second surface, the harmonic 1-form corresponding to each homology group basis is calculated; based on the harmonic 1-form corresponding to each homology group basis of the second surface, the holomorphic 1-form corresponding to each homology group basis is calculated.

[0065] Here, the homology group basis r1 and r2 for each dimension of the second surface are calculated. Based on homology group basis r1, the harmonic 1-form w1 corresponding to homology group basis r1 is calculated; based on homology group basis r2, the harmonic 1-form w2 corresponding to homology group basis r2 is calculated; the holomorphic 1-form corresponding to the homology group basis for each dimension is calculated through the following steps: w n +i*w n Among them, w n For the harmonic 1-form of the nth dimension, *w n Let i be the conjugate harmonic 1-form of the harmonic 1-form of the nth dimension, where i is an imaginary number.

[0066] e. Integrate the holomorphic 1-form of any second surface to obtain the rectangular complex field.

[0067] In this embodiment, the holomorphic 1-form of any second surface is integrated to obtain a mapping. Let be a rectangular complex field.

[0068] f. Based on the curvature of each point in the first surface, render the rectangular complex domain to obtain a parametric image of the complex plane.

[0069] In this embodiment, each point is colored according to its curvature in the first surface; the rectangular complex domain is then rendered based on the colored first surface. Since the grayscale value of the complex plane parametric image is determined by its curvature, the complex plane parametric image can reflect the global distribution of curvature.

[0070] Here, the curvature of each point in the first surface is mapped onto a color bar to color each point in the first surface; based on the colored first surface, the rectangular complex domain of the second surface is rendered to obtain a parametric image of the complex plane.

[0071] Step 2: Detect polyp connected components in the complex plane parameterized image.

[0072] Specifically, each pixel in the complex plane parameterized image is binarized according to a preset polyp threshold range.

[0073] In this embodiment, pixels within a preset polyp threshold range are considered polyps, and their values ​​are set to 1; pixels outside the preset polyp threshold range are considered non-polyps, and their values ​​are set to 0.

[0074] Reference Figure 3 As shown, this is a binarized parametric image of the complex plane provided in an embodiment of this application.

[0075] Specifically, polyp connected components are determined in the complex plane parameterized image after binarization.

[0076] Step 3: Filter the connected regions of each polyp according to their shape parameters to obtain multiple initial polyp images.

[0077] In this embodiment, a portion of the polyp connected regions obtained based on a preset polyp threshold range are not polyp image regions, but rather wrinkled image regions. Therefore, by using the shape parameters of each polyp connected region, wrinkled images are filtered out, thereby avoiding the influence of wrinkled images on the polyp image determination result.

[0078] Specifically, based on the roundness (roundness refers to the degree to which the shape of the polyp connected region is close to the theoretical circle), eccentricity, and the ratio of the length to the width of the circumscribed rectangle of each polyp connected region in the complex plane parameterized image, it is determined whether each polyp connected region is a polyp region; if the polyp connected region is a polyp region, the image at the location of the polyp connected region is determined as the initial polyp image.

[0079] Here, based on the roundness, eccentricity, and the length-to-width ratio of the circumscribed rectangle of each polyp connected region in the complex plane parametric image, it is determined whether each polyp connected region is a polyp region. This includes: if the number of polyp connected regions satisfying the corresponding threshold conditions in roundness, eccentricity, and the length-to-width ratio of the circumscribed rectangle is greater than a preset number, then the polyp connected region is a polyp region; otherwise, the polyp connected region is a wrinkled region. The roundness, eccentricity, and the length-to-width ratio of the circumscribed rectangle each have their own corresponding threshold conditions.

[0080] S202. Extract the three-dimensional polyp surface patch from the three-dimensional curved surface of the inner wall surface of the colon for each initial polyp image.

[0081] S203. Based on the geometric features of the three-dimensional polyp surface patches corresponding to each initial polyp image, determine the target polyp image from multiple initial polyp images.

[0082] Specifically, the target polyp image is determined from multiple initial polyp images based on the area of ​​the three-dimensional polyp surface patch and the ratio of the longest side to the second longest side of the directional bounding box of the three-dimensional polyp surface patch.

[0083] In this embodiment, the Principal Component Analysis (PCA) method is used to calculate the orientation bounding box (OOBB) of the three-dimensional polyp surface patch. The smaller the ratio of the longest side to the second longest side of the OOBB of the three-dimensional polyp surface patch, that is, the closer the longest side is to the second longest side, the greater the probability that the initial polyp image is the target polyp image.

[0084] Bounding boxes are used in computer graphics to provide a simplified representation of the spatial extent of an object or set of objects, enabling efficient computation of various geometric operations. The basic idea is to approximate complex geometric objects with a slightly larger, simpler geometric shape (called a bounding box). Out-of-Box (OOBB) is a commonly used type of bounding box. It is the smallest rectangular prism containing the object and arbitrarily oriented relative to the coordinate axes. The most significant characteristic of an OOBB is its arbitrary orientation. The main advantage of using an OOBB is that it can enclose the object as tightly as possible based on its shape. This is particularly useful for objects not aligned with the coordinate axes, such as complex 3D models or irregularly shaped point clouds. To define an OOBB, we need its position, size, and orientation. Position represents the center point of the rectangular prism, size determines the dimensions (width, height, and depth), and orientation specifies the rotation matrix or quaternion that transforms the rectangular prism from its normal orientation to its current orientation. OOBBs can be computed using various algorithms, such as iterative algorithms known as "minimum bounding rectangles" or methods based on Principal Component Analysis (PCA). In summary, directional bounding boxes provide a flexible and efficient way to enclose and approximate the shape of complex objects, allowing for a variety of optimizations and calculations in computer graphics and computational geometry applications.

[0085] Furthermore, after determining the target polyp image, the method further includes: extracting polyp texture features of each target polyp image in the complex plane parameterized image; inputting the polyp texture features of each target polyp image into the polyp classification model to obtain the polyp image detection result of each target polyp image; the polyp classification model is trained by polyp texture sample features and corresponding polyp image detection result labels; and determining the target polyp image whose polyp image detection result is a polyp image as the final polyp image.

[0086] This application provides a method for determining colonic polyp images based on a parametric algorithm. The method is applied to a processor and includes: acquiring multiple initial polyp images from a complex-plane parametric image of the colon's inner wall surface; the complex-plane parametric image is a conformally parametric image of the colon's inner wall surface; extracting three-dimensional polyp surface patches from each initial polyp image on a three-dimensional curved surface of the colon's inner wall surface; and determining a target polyp image from the multiple initial polyp images based on the geometric features of the three-dimensional polyp surface patches corresponding to each initial polyp image. This application can determine colonic polyp images and considers the global curvature of the colon's inner wall surface to be detected, avoiding detection errors caused by uneven CT density within the polyp and missed detections due to fold occlusion, thus improving the accuracy of determining colonic polyp images.

[0087] Reference Figure 4 The diagram shown is a flowchart illustrating another method for determining colon polyp images based on a parametric algorithm provided in this application. This method comprises... Figure 1 The processor 101 in the process executes the following exemplary steps according to embodiments of this application:

[0088] S401. Extract the polyp texture features of each target polyp image in the complex plane parameterized image.

[0089] In this embodiment of the application, for two-dimensional slices of each target polyp image in the complex plane parameterized image, 14 texture measures of the two-dimensional slices are calculated in four directions based on the gray-level co-occurrence matrix, resulting in 28 texture features (14 means and 14 ranges).

[0090] Here, the gray-level co-occurrence matrix is ​​generally defined as a comparison matrix (also called a comparison gray-level matrix) used to represent the relationship between the gray values ​​of a pixel and other pixels, or it can be viewed as a comparison matrix between a finite number of adjacent pixels. Texture is usually defined as a certain local property of an image, or a measure of the relationship between pixels in a local image.

[0091] S402. Input the polyp texture features of each target polyp image into the polyp classification model to obtain the polyp image detection results of each target polyp image; the polyp classification model is trained by polyp texture sample features and corresponding polyp image detection result labels.

[0092] In this embodiment, the polyp classification model is constructed using multiple decision trees (commonly referred to as "random forests"). Each decision tree randomly selects a number of growth samples equal to the total number of training samples, and the number of sampled features at each node is equal to the square root of all features. For each classification experiment, the data is divided into a training set and a test set. After building the polyp classification model using the training set information, the polyp classification model is used to classify the test set. Evaluation metrics for the polyp classification model include classification accuracy, recall, precision, and accuracy.

[0093] The polyp classification model is trained using polyp texture sample features and corresponding polyp image detection result labels. The polyp image detection results include polyp images and non-polyp images.

[0094] S403. The target polyp image whose polyp image detection result is a polyp image is determined as the final target polyp image.

[0095] This application provides another method for detecting colonic polyps. This method is applied to a processor and can determine the polyp image by the geometric features of the three-dimensional polyp surface patch in the three-dimensional curved surface of the inner wall of the colon based on the initial polyp image. Then, it can further detect the target polyp image that has been initially identified as a polyp image by using a polyp classification model. This avoids problems such as detection errors caused by uneven CT density within the polyp and missed detection of polyps due to folds, thereby improving the accuracy of polyp image determination.

[0096] This application provides a device for determining colonic polyp images based on a parametric algorithm. The device includes a processor, a storage medium, and a bus. The storage medium stores machine-readable instructions executable by the processor. When the device is running, the processor communicates with the storage medium via the bus. The processor performs the following steps: acquiring multiple initial polyp images from a complex-plane parametric image of the colon's inner wall surface; the complex-plane parametric image is a conformally parametric image of the colon's inner wall surface; extracting three-dimensional polyp surface patches from each initial polyp image on the three-dimensional curved surface of the colon's inner wall surface; and determining a target polyp image from the multiple initial polyp images based on the geometric features of the three-dimensional polyp surface patches corresponding to each initial polyp image. This application can determine colonic polyp images and considers the global curvature of the colon's inner wall surface to be detected, avoiding detection errors caused by uneven CT density within the polyp and missed detections due to fold occlusion, thus improving the accuracy of determining colonic polyp images.

[0097] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the method embodiments, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection can be through some communication interfaces; the indirect coupling or communication connection of devices or modules can be electrical, mechanical, or other forms.

[0098] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0099] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0100] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the information processing methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0101] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A device for determining colon polyp images based on a parametric algorithm, characterized in that, The device includes a processor, a storage medium, and a bus. The storage medium stores machine-readable instructions executable by the processor. When the device is running, the processor communicates with the storage medium via the bus. The processor is used to perform the following steps: Acquiring multiple initial polyp images from a complex plane parametric image of the inner wall surface of the colon includes: acquiring a three-dimensional surface of the inner wall surface of the colon; removing topological noise from the three-dimensional surface to obtain a first surface of the inner wall surface of the colon; the first surface is a zero-genus surface; performing a double-layer overlay construction on the first surface to obtain a second surface of genus 1; the dimension of the homology group basis of the second surface is 2; calculating the holomorphic 1-form of the second surface corresponding to each homology group basis; and converting any holomorphic 1-form of the second surface... The equation is integrated to obtain a rectangular complex domain; the rectangular complex domain is rendered according to the curvature of each point in the first surface to obtain the complex plane parameterized image; the polyp connected components in the complex plane parameterized image are detected, including: binarizing each pixel in the complex plane parameterized image according to a preset polyp threshold range; determining the polyp connected components in the binarized complex plane parameterized image; filtering each polyp connected component according to its shape parameters to obtain multiple initial polyp images; the complex plane parameterized image is an image of the inner wall surface of the colon after conformal parameterization; Extract three-dimensional polyp surface patches from the three-dimensional curved surface of the inner wall surface of the colon for each initial polyp image; Determining a target polyp image from multiple initial polyp images based on the geometric features of the three-dimensional polyp surface corresponding to each initial polyp image includes: determining the target polyp image from multiple initial polyp images based on the area of ​​the three-dimensional polyp surface and the ratio of the longest side to the second longest side of the directional bounding box of the three-dimensional polyp surface.

2. The apparatus for determining colon polyp images based on a parameterized algorithm according to claim 1, characterized in that, When the processor performs the calculation of the holomorphic 1-forms of the second surface corresponding to each homology group basis, it performs the following steps: Calculate the basis of each homology group of the second surface; Based on the homology group basis of the second surface, calculate the harmonic 1-form corresponding to each homology group basis; Based on the harmonic 1-forms of the second surface corresponding to the basis of each homology group, calculate the holomorphic 1-forms corresponding to the basis of each homology group.

3. The apparatus for determining colon polyp images based on a parametric algorithm according to claim 1, characterized in that, When the processor renders the rectangular complex domain based on the curvature of each point in the first surface, it performs the following steps: Each point is colored according to its curvature in the first surface; The rectangular complex domain is rendered based on the first colored surface.

4. The apparatus for determining colon polyp images based on a parametric algorithm according to claim 1, characterized in that, When the processor performs the step of filtering each polyp connected region according to the shape parameters of each polyp connected region to obtain multiple initial polyp images, it performs the following steps: Based on the roundness, eccentricity, and the ratio of the length to the width of the circumscribed rectangle of each polyp connected region in the complex plane parameterized image, it is determined whether each polyp connected region is a polyp region. If the polyp connected region is a polyp area, then the image at the location of the polyp connected region is determined as the initial polyp image.

5. The apparatus for determining colon polyp images based on a parametric algorithm according to any one of claims 1 to 3, characterized in that, After determining the target polyp image, the processor is further configured to perform the following steps: Extract the polyp texture features of each target polyp image in the complex plane parameterized image; The polyp texture features of each target polyp image are input into the polyp classification model to obtain the polyp image detection results of each target polyp image; the polyp classification model is trained by polyp texture sample features and corresponding polyp image detection result labels; The target polyp image whose polyp image detection result is a polyp image is determined as the final polyp image.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor as described in any one of claims 1 to 5, performs the steps executed by the processor.

Images