Fluorescence metabolism enhancement imaging method and system based on three-mode fusion

Abstract

The invention relates to the technical field of enhanced imaging, in particular to a fluorescence metabolism enhanced imaging method and system based on three-mode fusion, and the method comprises the steps: confirming three-mode colposcope imaging equipment, carrying out the illumination uniformity calibration operation on the three-mode colposcope imaging equipment, obtaining calibration imaging equipment, obtaining a three-mode registration image set, and carrying out the illumination uniformity calibration operation on the three-mode colposcope imaging equipment; performing dynamic monitoring based on a plurality of preset acetic acid action time points to obtain a lesion evolution dynamic video, performing lesion area identification analysis operation on the registered fluorescence metabolism image to obtain a metabolism characteristic parameter set, performing time sequence dynamic characteristic analysis operation on the lesion evolution dynamic video to obtain a time sequence dynamic characteristic parameter set, and performing dynamic analysis on the time sequence dynamic characteristic parameter set; and obtaining a lesion risk grading result based on the metabolism characteristic parameter set and the time sequence dynamic characteristic parameter set, and completing fluorescence metabolism enhancement imaging based on three-mode fusion based on the lesion risk grading result. According to the invention, the imaging effect and the lesion detection sensitivity can be improved.

Description

Technical Field

[0001] This invention relates to the field of enhanced imaging technology, and in particular to a fluorescence metabolism enhanced imaging method and system based on trimodal fusion. Background Technology

[0002] Modality refers to the imaging mode or method. Trimodality refers to the fusion of three different imaging modalities. Fluorescence imaging utilizes the property of fluorescent substances (fluorescent probes) emitting fluorescence when excited by light of a specific wavelength. Metabolic enhancement refers to enhancing the metabolic processes of an organism or tissue in a certain way, making the relevant metabolic signals more clearly detectable during imaging.

[0003] Despite significant advancements in existing imaging technologies, single-modal imaging methods often have limitations. First, while optical imaging offers advantages such as high resolution and real-time imaging capabilities, its imaging depth is limited, and its penetration into tissues is weak, making it difficult to acquire information from deep tissues. Second, conventional magnetic resonance imaging (MRI) has a clear advantage in imaging depth, clearly revealing soft tissue anatomy, but its imaging speed is relatively slow, and its sensitivity for detecting certain early lesions (such as early tumors) remains insufficient. Therefore, improving imaging quality and lesion detection sensitivity are urgent technical challenges that need to be addressed. Summary of the Invention

[0004] This invention provides a fluorescence metabolism-enhanced imaging method based on trimodal fusion and a computer-readable storage medium, the main purpose of which is to improve imaging effect and lesion detection sensitivity.

[0005] To achieve the above objectives, the present invention provides a fluorescence metabolism enhancement imaging method based on three-modal fusion, comprising:

[0006] A three-modal colposcope imaging device was identified, which includes: white light imaging mode, polarized light imaging mode, fluorescence imaging mode, three-link automatic image acquisition mode, and laser ranging module.

[0007] An illumination uniformity calibration operation is performed on the three-modal colposcopy imaging device to obtain a calibrated imaging device. Based on the calibrated imaging device, a three-modal registration image set is obtained, which includes: a reference image, a registration polarized light image, and a registration fluorescence metabolism image.

[0008] Dynamic monitoring is performed based on multiple preset acetic acid treatment time points to obtain dynamic video of lesion evolution, which includes multiple lesion evolution image frames.

[0009] A lesion region identification and analysis operation is performed on the registered fluorescence metabolic image to obtain a set of metabolic feature parameters, which includes one or more metabolic feature parameters.

[0010] Perform temporal dynamic feature analysis on the dynamic video of lesion evolution to obtain a set of temporal dynamic feature parameters, wherein the set of temporal dynamic feature parameters includes one or more temporal dynamic feature parameters;

[0011] Based on the metabolic feature parameter set and the time-series dynamic feature parameter set, the lesion risk grading results are obtained, and based on the lesion risk grading results, fluorescence metabolic enhancement imaging based on three-modal fusion is completed.

[0012] Optionally, acquiring the three-modal registration image set based on the calibrated imaging device includes:

[0013] The laser ranging module and white light imaging mode are used to automatically focus the calibrated imaging device to obtain the focused imaging device.

[0014] An initial white light cervical region image is obtained based on a focusing imaging device and a pre-built AI image recognition algorithm. Reflectance analysis is performed on the initial white light cervical region image to obtain the analysis results, which indicate whether reflection exists or not.

[0015] If the analysis result indicates the presence of reflection, then the reflection area image is obtained based on the initial white light cervical area image. The reflection area image in the initial white light cervical area image is then subjected to intelligent reflection elimination operation using the polarized light imaging mode to obtain the white light cervical area image. The reflection interference index is calculated based on the initial white light cervical area image and the white light cervical area image. The white light cervical area image is then labeled using the reflection interference index to obtain the labeled cervical area image.

[0016] If the analysis result shows that there is no reflection, then the initial white light cervical region image is used as the cervical region image.

[0017] A three-modal registration image set was obtained based on the image of the cervical region and the focusing imaging device.

[0018] Optionally, the step of calculating the reflection interference index based on the initial white light cervical region image and the white light cervical region image includes:

[0019] The number of initial pixels and the number of reflected pixels in the initial white light cervical region image and the reflected region image were counted respectively, and the proportion of reflected area was calculated based on the number of initial pixels and the number of reflected pixels.

[0020] Obtain the pixel set of the reflective area in the reflective area image, and determine whether the reflective area image is a preset gray image;

[0021] If the image of the reflective area is not a preset gray image, then each pixel in the reflective area pixel set is converted to grayscale to obtain a reflective grayscale pixel set.

[0022] A set of reflective grayscale pixel values ​​is obtained from the set of reflective grayscale pixels, wherein the set of reflective grayscale pixels includes multiple reflective grayscale pixels, the set of reflective grayscale pixel values ​​includes multiple reflective grayscale pixel values, and there is a one-to-one correspondence between reflective grayscale pixels and reflective grayscale pixel values.

[0023] The number of reflected grayscale pixel values ​​and the total number of grayscale pixel values ​​are calculated based on the set of reflected grayscale pixel values. The average brightness value is calculated based on the number of reflected grayscale pixel values ​​and the total number of grayscale pixel values. The average brightness value is obtained by dividing the total number of grayscale pixel values ​​by the number of reflected grayscale pixel values.

[0024] The reflection interference index is obtained by weighting and summing the reflective area ratio and the average brightness value.

[0025] Optionally, the step of acquiring a three-modal registration image set based on the identified cervical region image and the focusing imaging device includes:

[0026] A pseudo-color metabolic overlay image is obtained based on the fluorescence imaging mode and the image of the marked cervical region. A positioning command is received, and the focusing imaging device is adjusted according to the positioning command and the pseudo-color metabolic overlay image to obtain the positioning imaging device.

[0027] The positioning imaging device is set using a preset acquisition cycle to obtain the set imaging device. The three-link automatic image acquisition mode in the set imaging device is then activated. The set imaging device after activating the three-link automatic image acquisition mode is used to acquire white light morphology images, polarized light detail images, and fluorescence metabolism images.

[0028] Multimodal spatiotemporal registration was performed on white light morphology images, polarized light detail images, and fluorescence metabolism images to obtain a three-modal registered image set.

[0029] Optionally, the multimodal image spatiotemporal registration operation is performed on the white light morphology image, polarized light detail image, and fluorescence metabolism image to obtain a three-modal registered image set, including:

[0030] Using the white light morphology image as the reference image, feature points were extracted from the reference image, the polarized light detail image, and the fluorescence metabolism image respectively, resulting in the feature point group of the reference image, the feature point group of the polarized light image, and the feature point group of the fluorescence metabolism image.

[0031] Polarized light image feature points are extracted sequentially from the polarized light image feature point group. The polarized light image feature points are then matched with the reference image feature point group to obtain multiple matching feature point pairs. These multiple matching feature point pairs are then summarized to obtain a matching feature point pair group.

[0032] The optimal spatial transformation matrix is ​​obtained based on the matching feature point pair group. The registered polarized light image is obtained based on the optimal spatial transformation matrix and the polarized light detail image. The registered fluorescence metabolism image is obtained based on the fluorescent metabolism image feature point group and the reference image feature point group.

[0033] A three-modal registration image set was identified based on the reference image, the registered polarized light image, and the registered fluorescence metabolic image.

[0034] Optionally, obtaining the registered polarized light image based on the optimal spatial transformation matrix and the polarized light detail image includes:

[0035] A blank canvas is created based on the reference image, and the blank canvas is used as the target polarization light registration image, wherein the target polarization light registration image includes multiple target pixels;

[0036] The target pixels are extracted sequentially from the target polarization registration image. The target pixel coordinates are determined based on the target pixels. The inverse transformation matrix is ​​obtained based on the optimal spatial transformation matrix. The source pixel coordinates are calculated based on the inverse transformation matrix and the target pixel coordinates. The source pixel coordinates include the horizontal and vertical coordinates.

[0037] If the source pixel coordinates are preset integer coordinates, the target source pixel value is obtained from the polarized light detail image based on the source pixel coordinates, and the target pixel is assigned a value using the target source pixel value to obtain the assigned target pixel;

[0038] If the source pixel coordinates are preset floating-point coordinates, then the x-coordinate of the source pixel coordinates is rounded down to obtain the x-coordinate rounded down and the x-coordinate rounded up. The y-coordinate of the source pixel coordinates is rounded down to obtain the y-coordinate rounded down and the y-coordinate rounded up.

[0039] The filling pixel value is calculated based on the rounded down x-coordinate, rounded up x-coordinate, rounded down y-coordinate, and rounded up y-coordinate. The filling pixel value is then used to assign a value to the target pixel to obtain the assigned target pixel.

[0040] The target pixels are summed and assigned values ​​to obtain the registered polarized light image.

[0041] Optionally, calculating the source pixel coordinates based on the inverse transformation matrix and the target pixel coordinates includes:

[0042] The target pixel coordinates are transformed to obtain homogeneous pixel coordinates. The source pixel coordinates are then calculated based on the homogeneous pixel coordinates and the inverse transformation matrix. The formula for calculating the source pixel coordinates is shown below:

[0043] ,

[0044] in, Represents homogeneous pixel coordinates. This represents the x-coordinate in homogeneous pixel coordinates. This represents the ordinate in homogeneous pixel coordinates. Represents the inverse transformation matrix. This represents the x-coordinate of the source pixel coordinates. The ordinate represents the source pixel coordinates.

[0045] Optionally, calculating the filled pixel value based on the rounded-down x-coordinate, rounded-up x-coordinate, rounded-down y-coordinate, and rounded-up y-coordinate includes:

[0046] Combine the down-rounded x-coordinate with the down-rounded y-coordinate to obtain the first down-combined coordinate; combine the down-rounded x-coordinate with the up-rounded y-coordinate to obtain the second down-combined coordinate.

[0047] Based on the first downward combined coordinates and the second downward combined coordinates, the first downward pixel value and the second downward pixel value are determined in the polarized light detail image.

[0048] The horizontal coordinate of the source pixel coordinates is interpolated using the first downward pixel value and the second downward pixel value to obtain the downward interpolation;

[0049] Combine the upward-rounded x-coordinate with the downward-rounded y-coordinate to obtain the first upward combined coordinate; combine the upward-rounded x-coordinate with the upward-rounded y-coordinate to obtain the second upward combined coordinate.

[0050] Based on the first upward combined coordinates and the second upward combined coordinates, the first upward pixel value and the second upward pixel value are determined in the polarized light detail image;

[0051] The x-coordinate of the source pixel coordinates is interpolated using the first upward pixel value and the second upward pixel value to obtain the upward interpolation;

[0052] The fill pixel value is calculated based on down-interpolation, up-interpolation, and the ordinate of the source pixel coordinates. The formula for calculating the fill pixel value is as follows:

[0053] ,

[0054] in, Indicates filling in pixel values, Indicates downward interpolation. This indicates the ordinate is rounded up. The ordinate represents the source pixel coordinates. Indicates upward interpolation. This indicates the vertical coordinate being rounded down.

[0055] Optionally, the step of performing temporal dynamic feature analysis on the dynamic video of lesion evolution to obtain a set of temporal dynamic feature parameters includes:

[0056] Extract disease evolution image frames sequentially from the dynamic video of disease evolution, and perform image difference operation between the disease evolution image frames and the reference image to obtain the difference map;

[0057] Based on the difference map, a set of connected regions is obtained. Connected regions are extracted sequentially from the set of connected regions. The extracted connected regions are used as the acetic whitening reaction regions. The area of ​​the acetic whitening reaction region and the average pixel intensity of the region are calculated.

[0058] The area of ​​the acetic acid whitening reaction region and the average pixel intensity of the region are summarized to obtain the set of the acetic acid whitening reaction region area and the set of the average pixel intensity of the region.

[0059] By sorting the area set of the acetic acid whitening reaction region in chronological order using multiple acetic acid reaction time points, a sequence of acetic acid whitening reaction region areas is obtained.

[0060] The occurrence time of the acetic whitening reaction is determined based on the area sequence of the acetic whitening reaction region and the preset minimum area threshold, and the expansion rate of the acetic whitening reaction region is obtained based on the area sequence of the acetic whitening reaction region.

[0061] The peak time of the acetic acid whitening reaction and the rate of change of the intensity of the acetic acid whitening reaction are obtained based on the average pixel intensity set of the region. The temporal dynamic characteristic parameters are determined based on the occurrence time of the acetic acid whitening reaction, the peak time of the acetic acid whitening reaction, the rate of change of the intensity of the acetic acid whitening reaction and the expansion speed of the acetic acid whitening reaction region.

[0062] By summarizing the time-series dynamic feature parameters, a set of time-series dynamic feature parameters is obtained.

[0063] To achieve the above objectives, the present invention also provides a fluorescence metabolism enhancement imaging system based on trimodal fusion, comprising:

[0064] The imaging equipment calibration module is used to identify the trimodal colposcope imaging equipment, which includes: white light imaging mode, polarized light imaging mode, fluorescence imaging mode, triple automatic image acquisition mode, and laser ranging module. The module performs an illumination uniformity calibration operation on the trimodal colposcope imaging equipment to obtain a calibrated imaging equipment. Based on the calibrated imaging equipment, a trimodal registration image set is acquired, which includes: a reference image, a registration polarized light image, and a registration fluorescence metabolism image.

[0065] The dynamic monitoring module is used to perform dynamic monitoring based on multiple preset acetic acid treatment time points to obtain dynamic video of lesion evolution, wherein the dynamic video of lesion evolution includes multiple lesion evolution image frames.

[0066] The lesion region identification module is used to perform lesion region identification and analysis on the registered fluorescence metabolic image to obtain a set of metabolic feature parameters, wherein the set of metabolic feature parameters includes one or more metabolic feature parameters.

[0067] The temporal dynamic feature analysis module is used to perform temporal dynamic feature analysis on dynamic videos of lesion evolution to obtain a temporal dynamic feature parameter set. The temporal dynamic feature parameter set includes one or more temporal dynamic feature parameters. Based on the metabolic feature parameter set and the temporal dynamic feature parameter set, the lesion risk classification result is obtained. Based on the lesion risk classification result, fluorescence metabolic enhancement imaging based on three-modal fusion is completed.

[0068] To address the above problems, the present invention also provides an electronic device, the electronic device comprising:

[0069] Memory, storing at least one instruction;

[0070] The processor executes the instructions stored in the memory to implement the fluorescence metabolism enhancement imaging method based on trimodal fusion described above.

[0071] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing at least one instruction, which is executed by a processor in an electronic device to implement the aforementioned fluorescence metabolism enhancement imaging method based on trimodal fusion.

[0072] To address the problems described in the background art, this invention provides a three-modal colposcopy imaging device. This device includes: a white light imaging mode, a polarized light imaging mode, a fluorescence imaging mode, a triple automatic image acquisition mode, and a laser ranging module. These multiple imaging modes complement each other, providing more comprehensive and richer information for subsequent lesion diagnosis. Illumination uniformity calibration is performed on the three-modal colposcopy imaging device to obtain a calibrated imaging device. Based on the calibrated imaging device, a three-modal registration image set is acquired, comprising: a reference image, a registration polarized light image, and a registration fluorescence image. In metabolic imaging, the illumination uniformity calibration operation of this invention can eliminate illumination differences at different positions and angles of the device, ensuring that the acquired image has uniform brightness and contrast throughout the entire field of view. This avoids image distortion or loss of detail caused by uneven illumination, improving image clarity and quality, and facilitating more accurate observation and analysis of lesion characteristics. Based on multiple preset acetic acid treatment time points, dynamic monitoring is performed to obtain a dynamic video of lesion evolution. This dynamic video includes multiple lesion evolution image frames. The acetic acid of this invention can change the color and morphology of lesion tissue. By performing dynamic monitoring at multiple acetic acid treatment time points… This invention employs dynamic monitoring to capture the evolution of lesions under the influence of acetic acid. The dynamic video of lesion evolution contains multiple image frames, allowing observation of the entire process from the initial state to gradual changes, revealing the development trend and characteristics of the lesion. Lesion region identification and analysis are performed on the registered fluorescence metabolic images to obtain a set of metabolic feature parameters, which includes one or more metabolic feature parameters. These metabolic feature parameters provide important basis for the diagnosis and assessment of lesions. Different types and stages of lesions may have different metabolic characteristics. By analyzing these parameters, doctors can distinguish between normal and diseased tissues, determine the benign or malignant nature and development stage of the lesion, and provide a reference for developing treatment plans. Temporal dynamic feature analysis is performed on the dynamic video of lesion evolution to obtain a set of temporal dynamic feature parameters, which also includes one or more temporal dynamic feature parameters. This invention combines metabolic feature parameters and temporal dynamic feature parameters to assess lesions from different perspectives, improving the accuracy and reliability of diagnosis. Based on the metabolic feature parameter set and the temporal dynamic feature parameter set, lesion risk grading results are obtained. Based on the lesion risk grading results, fluorescence metabolic enhancement imaging based on three-modal fusion is completed. Therefore, the present invention can improve imaging results and the sensitivity of lesion detection. Attached Figure Description

[0073] Figure 1 This is a schematic flowchart of a fluorescence metabolism enhancement imaging method based on three-modal fusion provided in an embodiment of the present invention;

[0074] Figure 2A functional block diagram of a fluorescence metabolism enhancement imaging system based on three-modal fusion provided in an embodiment of the present invention;

[0075] Figure 3 This is a schematic diagram of the structure of an electronic device for implementing the fluorescence metabolism enhancement imaging method based on three-modal fusion, according to an embodiment of the present invention.

[0076] Explanation of reference numerals in the attached figures:

[0077] 10. Electronic device; 11. Processor; 12. Memory; 13. Bus.

[0078] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0079] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0080] This application provides a fluorescence metabolism enhancement imaging method based on trimodal fusion. The executing entity of the trimodal fusion-based fluorescence metabolism enhancement imaging method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the trimodal fusion-based fluorescence metabolism enhancement imaging method can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0081] Reference Figure 1 The diagram shown is a flowchart illustrating a fluorescence metabolism enhancement imaging method based on trimodal fusion according to an embodiment of the present invention. In this embodiment, the fluorescence metabolism enhancement imaging method based on trimodal fusion includes:

[0082] S1. The three-modal colposcopy imaging device is identified, which includes: white light imaging mode, polarized light imaging mode, fluorescence imaging mode, three-link automatic image acquisition mode and laser ranging module.

[0083] It should be explained that the trimodal colposcopy imaging device is an imaging device composed of multiple LED ring light sources, capable of operating simultaneously or separately in three different imaging modes to acquire images of the cervix or other biological tissues. For example, the trimodal colposcopy imaging device consists of 16 LED ring light sources. The laser ranging module is a unit that uses lasers to measure distance. By emitting a laser beam and receiving the reflected laser signal, it calculates the distance between the calibrated imaging device and the target object (cervix). White light imaging mode is a method of imaging using white light (i.e., light containing all visible wavelengths) to acquire preliminary images of the cervical region for subsequent analysis and processing. Polarized light imaging mode is a technique that uses polarized light for imaging. Polarized light refers to light whose vibration direction is confined to a specific direction. By using polarizing filters, the polarization direction of light can be controlled, thereby reducing or eliminating reflections. Fluorescence imaging mode is a technique that uses fluorescent markers for imaging. Fluorescent markers emit fluorescence when excited by light of a specific wavelength; information within the organism is obtained by detecting the fluorescence signal. In this invention, the fluorescence imaging mode is used to acquire metabolic information of the cervical region and generate a pseudo-color metabolic overlay map. The Triple Automatic Image Acquisition Mode is an automated image acquisition mode that can simultaneously acquire three different modalities of images (such as white light morphology images, polarized light detail images, and fluorescence metabolism images) in a single operation, ensuring the spatiotemporal consistency of the three modalities and facilitating subsequent multimodal image spatiotemporal registration operations.

[0084] S2. Perform illumination uniformity calibration on the three-modal colposcopy imaging device to obtain a calibrated imaging device. Based on the calibrated imaging device, acquire a three-modal registration image set, which includes: a reference image, a registration polarized light image, and a registration fluorescence metabolism image.

[0085] Specifically, the acquisition of the three-modal registration image set based on the calibrated imaging device includes:

[0086] The laser ranging module and white light imaging mode are used to automatically focus the calibrated imaging device to obtain the focused imaging device.

[0087] An initial white light cervical region image is obtained based on a focusing imaging device and a pre-built AI image recognition algorithm. Reflectance analysis is performed on the initial white light cervical region image to obtain the analysis results, which indicate whether reflection exists or not.

[0088] If the analysis result indicates the presence of reflection, then the reflection area image is obtained based on the initial white light cervical area image. The reflection area image in the initial white light cervical area image is then subjected to intelligent reflection elimination operation using the polarized light imaging mode to obtain the white light cervical area image. The reflection interference index is calculated based on the initial white light cervical area image and the white light cervical area image. The white light cervical area image is then labeled using the reflection interference index to obtain the labeled cervical area image.

[0089] If the analysis result shows that there is no reflection, then the initial white light cervical region image is used as the cervical region image.

[0090] A three-modal registration image set was obtained based on the image of the cervical region and the focusing imaging device.

[0091] It should be explained that the illumination uniformity calibration operation involves using a uniform white calibration plate, placing it within the imaging area of ​​the trimodal colposcopy imaging device, acquiring an image of the calibration plate in white light imaging mode, calculating the brightness value of each pixel in the calibration image, generating a brightness distribution map, and adjusting the brightness and angle of the LED ring light source based on the brightness distribution map to ensure uniform illumination within the imaging area. A calibrated imaging device refers to an imaging device that provides uniform illumination after the illumination uniformity calibration operation. Illumination uniformity calibration eliminates illumination differences at different positions and angles, ensuring uniform brightness and contrast in the acquired image across the entire field of view. This avoids image distortion or loss of detail caused by uneven illumination, improves image clarity and quality, and helps in more accurately observing and analyzing lesion characteristics. A focusing imaging device refers to a calibrated imaging device that, after autofocusing, can accurately align with a target object (such as the cervix). AI image recognition algorithms are technologies that utilize artificial intelligence, particularly machine learning and deep learning algorithms, to analyze and recognize images. In this invention, an AI image recognition algorithm is used to identify the cervical region from an image acquired in a white light imaging mode and generate an initial white light cervical region image.

[0092] Understandably, the initial white-light cervical region image is an image of the cervical region extracted from an image acquired using an AI image recognition algorithm in white-light imaging mode. This image is used for subsequent reflection analysis and reflection removal. Reflection analysis is an image processing technique used to detect reflective areas in the initial white-light cervical region image. Reflective areas are bright areas formed when light directly reflects off the imaging device. Reflection analysis determines whether reflective areas exist in the initial white-light cervical region image, thus deciding whether reflection removal is necessary. Intelligent reflection removal is an automated image processing technique that uses polarized light imaging mode to eliminate reflective areas in the initial white-light cervical region image, resulting in a clearer and more accurate cervical region image and improved image quality. The white-light cervical region image is the final image of the cervical region after reflection removal. The labeled cervical region image is the cervical region image labeled with the reflection interference index and is used for subsequent three-modal registration. A trimodal registration image set refers to a collection of trimodal registration images obtained by aligning and fusing images of three different modalities (such as white light imaging, fluorescence imaging, and polarized light imaging). This invention achieves an objective quantitative assessment of colposcopy image quality by calculating and recording a reflection interference index. This index, by quantifying the area and intensity of reflective regions, provides traceable and objective quality evidence for diagnostic images. When a diagnostic report includes a low reflection interference index, it indicates that the conclusion is based on high-quality image evidence, significantly enhancing the reliability of the diagnostic results. Furthermore, during operation, this index can be fed back in real time to guide the operator in adjusting the lens angle or cleaning procedures to ensure optimal image acquisition.

[0093] Specifically, the calculation of the reflectivity interference index based on the initial white light cervical region image and the white light cervical region image includes:

[0094] The number of initial pixels and the number of reflected pixels in the initial white light cervical region image and the reflected region image were counted respectively, and the proportion of reflected area was calculated based on the number of initial pixels and the number of reflected pixels.

[0095] Obtain the pixel set of the reflective area in the reflective area image, and determine whether the reflective area image is a preset gray image;

[0096] If the image of the reflective area is not a preset gray image, then each pixel in the reflective area pixel set is converted to grayscale to obtain a reflective grayscale pixel set.

[0097] A set of reflective grayscale pixel values ​​is obtained from the set of reflective grayscale pixels, wherein the set of reflective grayscale pixels includes multiple reflective grayscale pixels, the set of reflective grayscale pixel values ​​includes multiple reflective grayscale pixel values, and there is a one-to-one correspondence between reflective grayscale pixels and reflective grayscale pixel values.

[0098] The number of reflected grayscale pixel values ​​and the total number of grayscale pixel values ​​are calculated based on the set of reflected grayscale pixel values. The average brightness value is calculated based on the number of reflected grayscale pixel values ​​and the total number of grayscale pixel values. The average brightness value is obtained by dividing the total number of grayscale pixel values ​​by the number of reflected grayscale pixel values.

[0099] The reflection interference index is obtained by weighting and summing the reflective area ratio and the average brightness value.

[0100] It needs to be explained that: the initial pixel count refers to the total number of pixels in the initial white light cervical region image. The reflective pixel count refers to the number of pixels in the reflective region image that are identified as reflective regions. The reflective area ratio refers to the ratio of the reflective pixel count to the initial pixel count. The reflective region pixel set refers to the set of all reflective region pixels in the reflective region image. Reflective region pixels refer to the pixels in the reflective region image. A grayscale image means that all pixel values ​​in the image are grayscale values, that is, the image has no color information, only brightness information. Grayscale conversion is an image processing technique; converting a color image to a grayscale image is an existing technology and will not be elaborated on here. The reflective grayscale pixel set refers to the set of reflective grayscale pixels of the reflective region pixels after grayscale conversion. The reflective grayscale pixel value set refers to the set of grayscale values ​​of each reflective grayscale pixel in the reflective grayscale pixel set. The reflective grayscale pixel value refers to the grayscale value of the reflective grayscale pixel. The number of reflective grayscale pixel values ​​refers to the number of reflective grayscale pixel values ​​in the reflective grayscale pixel value set. The sum of grayscale pixel values ​​refers to the sum of all reflective grayscale pixel values ​​in the reflective grayscale pixel value set. In the step of weighted summation of the reflective area ratio and average brightness value to obtain the reflective interference index, the reflective interference index = reflective area ratio. Preset area weight + average brightness value Preset brightness value weights. Area weights are pre-set coefficients used to adjust the relative importance of the reflective area ratio when calculating the glare interference index. Brightness value weights are pre-set coefficients used to adjust the relative importance of the average brightness value when calculating the glare interference index. In this invention, the sum of area weights and brightness value weights is 1.

[0101] Specifically, the acquisition of a three-modal registered image set based on the identified cervical region image and the focusing imaging device includes:

[0102] A pseudo-color metabolic overlay image is obtained based on the fluorescence imaging mode and the image of the marked cervical region. A positioning command is received, and the focusing imaging device is adjusted according to the positioning command and the pseudo-color metabolic overlay image to obtain the positioning imaging device.

[0103] The positioning imaging device is set using a preset acquisition cycle to obtain the set imaging device. The three-link automatic image acquisition mode in the set imaging device is then activated. The set imaging device after activating the three-link automatic image acquisition mode is used to acquire white light morphology images, polarized light detail images, and fluorescence metabolism images.

[0104] Multimodal spatiotemporal registration was performed on white light morphology images, polarized light detail images, and fluorescence metabolism images to obtain a three-modal registered image set.

[0105] It should be explained that pseudo-color metabolic overlay is a technique that converts fluorescence imaging data into color images. By mapping fluorescence signals of different intensities to different colors, it visually displays the distribution of metabolic activity. In this invention, pseudo-color metabolic overlay is used to visually display the metabolic activity of the cervical region, aiding in imaging localization and adjustment. Localization instructions are system-generated commands that guide the focusing imaging device to adjust its position and angle to ensure accurate alignment with the target area in the pseudo-color metabolic overlay. Localization imaging device refers to the device that, after adjustment by the localization instructions, can accurately align with the target area. Acquisition cycle refers to the time interval during which the localization imaging device acquires images. For example, acquiring a white light image in seconds 0-1, turning off the white light imaging mode and acquiring a polarized light image in seconds 1-2, and turning off the polarized light imaging mode and acquiring a fluorescence image in seconds 2-3.5. Pre-set imaging device refers to the imaging device after the acquisition cycle has been set. White light morphological images are images acquired using the white light imaging mode, used to provide anatomical information about the cervical region, helping doctors or researchers understand the morphological characteristics of the cervix. Polarized detail images, acquired using polarized light imaging mode, provide detailed information about the cervical region, especially revealing the fine structure of tissues more clearly after reducing reflective interference. Fluorescence metabolic images, captured by adding a polarizer before fluorescence imaging mode, provide information on metabolic activity in the cervical region, helping doctors and researchers understand the metabolic state of the cervix, particularly important for detecting early lesions or disease progression. It should be noted that adding a polarizer before fluorescence imaging mode aims to suppress unwanted specular reflections generated during fluorescence excitation, thereby capturing a purer and more realistic autofluorescence signal. This autofluorescence signal is light emitted at a different wavelength when the tissue's own natural substances are excited by light of a specific wavelength.

[0106] Specifically, the multimodal spatiotemporal registration operation is performed on the white light morphology image, the polarized light detail image, and the fluorescence metabolism image to obtain a three-modal registered image set, including:

[0107] Using the white light morphology image as the reference image, feature points were extracted from the reference image, the polarized light detail image, and the fluorescence metabolism image respectively, resulting in the feature point group of the reference image, the feature point group of the polarized light image, and the feature point group of the fluorescence metabolism image.

[0108] Polarized light image feature points are extracted sequentially from the polarized light image feature point group. The polarized light image feature points are then matched with the reference image feature point group to obtain multiple matching feature point pairs. These multiple matching feature point pairs are then summarized to obtain a matching feature point pair group.

[0109] The optimal spatial transformation matrix is ​​obtained based on the matching feature point pair group. The registered polarized light image is obtained based on the optimal spatial transformation matrix and the polarized light detail image. The registered fluorescence metabolism image is obtained based on the fluorescent metabolism image feature point group and the reference image feature point group.

[0110] A three-modal registration image set was identified based on the reference image, the registered polarized light image, and the registered fluorescence metabolic image.

[0111] It should be explained that the reference image is the image used as a reference in multimodal image registration. This invention uses the white light morphology image as the reference image because it provides clear anatomical information, facilitating alignment with other images. Feature point extraction refers to the operation of extracting feature points from the reference image, polarized light detail image, and fluorescence metabolism image using feature point extraction algorithms (such as SIFT, SURF, etc.). The reference image feature point set is the collection of all reference image feature points extracted from the reference image. The fluorescence metabolism image feature point set is the collection of all fluorescence metabolism image feature points extracted from the fluorescence metabolism image. The polarized light image feature point refers to the collection of all polarized light image feature points extracted from the polarized light detail image. A polarized light image feature point is a single feature point within the polarized light image feature point set. A matched feature point pair refers to a pair of feature points found in the reference image and the polarized light image that correspond in spatial location. The feature point matching operation refers to the operation of finding matching feature point pairs using a feature point matching algorithm (such as FLANN or BFMatcher). The matched feature point pair set is the collection of all matching feature point pairs. The optimal spatial transformation matrix is ​​calculated from the matching feature point pairs and is used to transform the polarized light image into the coordinate system of the reference image. This invention uses the RANSAC algorithm to robustly estimate the optimal transformation matrix. The registered polarized light image is obtained by aligning the polarized light detail image to the reference image using the optimal spatial transformation matrix. The registered fluorescence metabolism image is obtained by aligning the fluorescence metabolism image to the reference image. The method for obtaining the registered fluorescence metabolism image based on the feature point pairs of the fluorescence metabolism image and the reference image is the same as the method for obtaining the registered polarized light image based on the feature point pairs of the polarized light image and the reference image, and will not be repeated here. The registration operation of this invention spatially aligns images of different modalities, accurately matching information from different imaging modes. For example, after registering the fluorescence metabolism image with the reference image or the polarized light image, metabolic information can be intuitively combined with anatomical structures, providing a more comprehensive understanding of the location, extent, and metabolic characteristics of lesions, and offering a more accurate basis for subsequent lesion analysis.

[0112] Specifically, the step of obtaining the registered polarized light image based on the optimal spatial transformation matrix and the polarized light detail image includes:

[0113] A blank canvas is created based on the reference image, and the blank canvas is used as the target polarization light registration image, wherein the target polarization light registration image includes multiple target pixels;

[0114] The target pixels are extracted sequentially from the target polarization registration image. The target pixel coordinates are determined based on the target pixels. The inverse transformation matrix is ​​obtained based on the optimal spatial transformation matrix. The source pixel coordinates are calculated based on the inverse transformation matrix and the target pixel coordinates. The source pixel coordinates include the horizontal and vertical coordinates.

[0115] If the source pixel coordinates are preset integer coordinates, the target source pixel value is obtained from the polarized light detail image based on the source pixel coordinates, and the target pixel is assigned a value using the target source pixel value to obtain the assigned target pixel;

[0116] If the source pixel coordinates are preset floating-point coordinates, then the x-coordinate of the source pixel coordinates is rounded down to obtain the x-coordinate rounded down and the x-coordinate rounded up. The y-coordinate of the source pixel coordinates is rounded down to obtain the y-coordinate rounded down and the y-coordinate rounded up.

[0117] The filling pixel value is calculated based on the rounded down x-coordinate, rounded up x-coordinate, rounded down y-coordinate, and rounded up y-coordinate. The filling pixel value is then used to assign a value to the target pixel to obtain the assigned target pixel.

[0118] The target pixels are summed and assigned values ​​to obtain the registered polarized light image.

[0119] It should be explained that the blank canvas is an image with the same size as the reference image, with an initial pixel value of 0 (completely black), used to store the registered image data. The target pixel is a pixel point in the target polarized light registration image. The target pixel coordinates are the position of the target pixel in the target polarized light registration image. These target pixel coordinates are located in the image coordinate system of the target polarized light registration image. The horizontal coordinate rounding operation converts the horizontal coordinate of a floating-point number to an integer, used to determine integer coordinate points near the source pixel coordinates for interpolation calculations. The down-rounding horizontal coordinate is the integer obtained by rounding the horizontal coordinate of a floating-point number down. The up-rounding horizontal coordinate is the integer obtained by rounding the horizontal coordinate of a floating-point number up. The source pixel coordinates are floating-point coordinates calculated using the inverse transformation matrix, representing the position of the target pixel in the polarized light detail image. The vertical coordinate rounding operation converts the vertical coordinate of a floating-point number to an integer, used to determine integer coordinate points near the source pixel coordinates for interpolation calculations. The down-rounding vertical coordinate is the integer obtained by rounding the vertical coordinate of a floating-point number down. The rounded-up ordinate is the integer obtained by rounding the ordinate of a floating-point number up. The padded pixel value is the pixel value calculated using an interpolation algorithm and used to fill the target pixel. Assigning the padded pixel value to the target pixel is the operation of assigning the calculated padded pixel value to the target pixel. The assigned target pixel is the target pixel after the assignment operation. The registered polarized image is the final generated image, spatially aligned with the reference image. Integer coordinates refer to coordinates where both the x-coordinate and y-coordinate are integers. Floating-point coordinates refer to coordinates where both the x-coordinate and y-coordinate are floating-point numbers. The target source pixel value refers to the pixel value in the polarized detail image when the source pixel coordinates are preset integer coordinates.

[0120] Specifically, the step of calculating the source pixel coordinates based on the inverse transformation matrix and the target pixel coordinates includes:

[0121] The target pixel coordinates are transformed to obtain homogeneous pixel coordinates. The source pixel coordinates are then calculated based on the homogeneous pixel coordinates and the inverse transformation matrix. The formula for calculating the source pixel coordinates is shown below:

[0122] ,

[0123] in, Represents homogeneous pixel coordinates. This represents the x-coordinate in homogeneous pixel coordinates. This represents the ordinate in homogeneous pixel coordinates. Represents the inverse transformation matrix. This represents the x-coordinate of the source pixel coordinates. The ordinate represents the source pixel coordinates.

[0124] It should be explained that the inverse transformation matrix is ​​the inverse of the optimal spatial transformation matrix, used to map the target pixel coordinates back to the source coordinates in the polarized detail image. The transformation of the target pixel coordinates refers to an extended coordinate representation method, which adds an extra dimension to the two-dimensional coordinates of the target pixel, forming a three-dimensional vector. Homogeneous pixel coordinates refer to the coordinates obtained after the transformation of the target pixel coordinates.

[0125] Specifically, the calculation of the filled pixel value based on the rounded-down x-coordinate, rounded-up x-coordinate, rounded-down y-coordinate, and rounded-up y-coordinate includes:

[0126] Combine the down-rounded x-coordinate and the down-rounded y-coordinate to obtain the first down-combined coordinate; combine the up-rounded x-coordinate and the down-rounded y-coordinate to obtain the second down-combined coordinate.

[0127] Based on the first downward combined coordinates and the second downward combined coordinates, the first downward pixel value and the second downward pixel value are determined in the polarized light detail image.

[0128] The horizontal coordinate of the source pixel coordinates is interpolated using the first downward pixel value and the second downward pixel value to obtain the downward interpolation;

[0129] Combine the upward-rounded x-coordinate with the upward-rounded y-coordinate to obtain the first upward combined coordinate; combine the downward-rounded x-coordinate with the upward-rounded y-coordinate to obtain the second upward combined coordinate.

[0130] Based on the first upward combined coordinates and the second upward combined coordinates, the first upward pixel value and the second upward pixel value are determined in the polarized light detail image;

[0131] The x-coordinate of the source pixel coordinates is interpolated using the first upward pixel value and the second upward pixel value to obtain the upward interpolation;

[0132] The fill pixel value is calculated based on down-interpolation, up-interpolation, and the ordinate of the source pixel coordinates. The formula for calculating the fill pixel value is as follows:

[0133] ,

[0134] in, Indicates filling in pixel values, Indicates downward interpolation. This indicates the ordinate is rounded up. The ordinate represents the source pixel coordinates. Indicates upward interpolation. This indicates the vertical coordinate being rounded down.

[0135] It should be explained that the first downward combined coordinate is a coordinate composed of a down-rounded x-coordinate and a down-rounded y-coordinate. The second downward combined coordinate is a coordinate composed of an up-rounded x-coordinate and a down-rounded y-coordinate. The first downward pixel value is the pixel value of the first downward combined coordinate in the polarized light detail image, used to calculate the interpolation result of the source pixel coordinate in the x-axis direction. The second downward pixel value is the pixel value of the second downward combined coordinate in the polarized light detail image. The first upward combined coordinate is a coordinate composed of an up-rounded x-coordinate and an up-rounded y-coordinate. The second upward combined coordinate is a coordinate composed of a down-rounded x-coordinate and an up-rounded y-coordinate. The first upward pixel value is the pixel value of the first upward combined coordinate in the polarized light detail image. The second upward pixel value is the pixel value of the second upward combined coordinate in the polarized light detail image. The calculation formula for downward interpolation in the step of interpolating the x-coordinate of the source pixel coordinate using the first downward pixel value and the second downward pixel value is as follows:

[0136] ,

[0137] in, Indicates downward interpolation. Indicates the first downward pixel value. This indicates rounding up the x-coordinate. This represents the x-coordinate in the target pixel coordinates. Indicates the second downward pixel value. This indicates rounding down to the nearest integer on the x-coordinate. The formula for calculating the upward interpolation in the step of interpolating the x-coordinate of the source pixel coordinate using the first and second upward pixel values ​​is as follows:

[0138] ,

[0139] in, Indicates downward interpolation. Indicates the first upward pixel value. This indicates rounding up the x-coordinate. This represents the x-coordinate in the target pixel coordinates. Indicates the second upward pixel value. This indicates that the x-coordinate is rounded down to the nearest integer.

[0140] S3. Dynamic monitoring is performed based on multiple preset acetic acid treatment time points to obtain dynamic video of lesion evolution, wherein the dynamic video of lesion evolution includes multiple lesion evolution image frames.

[0141] It should be explained that the acetic acid treatment time points refer to the pre-set time points at which images are acquired after acetic acid treatment of biological tissue. For example, multiple acetic acid treatment time points may be 60 seconds, 90 seconds, 120 seconds, 150 seconds, and 180 seconds. The lesion evolution image frame refers to an image captured at each time point, showing the state of the lesion tissue at that specific time. By capturing images at multiple time points, the dynamic changes of the lesion tissue under the action of acetic acid can be recorded, helping doctors or researchers observe the evolution of the lesion. Simultaneously, by analyzing these image frames, the severity and risk level of the lesion can be more accurately assessed. The steps for dynamic monitoring based on multiple pre-set acetic acid treatment time points are as follows: a cotton ball (diameter > 2 cm) saturated with 5% diluted glacial acetic acid solution is used to cover the entire cervical surface, and a timer is started simultaneously. The timer is activated after confirming the timing mark is displayed and continues until 50 seconds have elapsed. The cotton ball is then removed, and the surface is dried with a dry cotton ball. Dynamic monitoring of the cervix is ​​then performed at each of the multiple acetic acid treatment time points.

[0142] S4. Perform lesion region identification and analysis on the registered fluorescence metabolic image to obtain a set of metabolic feature parameters, which includes one or more metabolic feature parameters.

[0143] It should be explained that the steps of performing lesion region identification and analysis on the registered fluorescence metabolic image are as follows: Registered fluorescence pixels are extracted sequentially from the registered fluorescence metabolic image; pixel intensity values ​​are obtained based on the extracted registered fluorescence pixels; if the pixel intensity value is less than or equal to a preset global fluorescence intensity threshold, the registered fluorescence pixel is considered a suspected abnormal metabolic pixel; otherwise, the registered fluorescence pixel is considered a normal metabolic pixel. The suspected abnormal metabolic pixels and normal metabolic pixels are summarized to obtain an initial abnormal region mask; an abnormal connected region set is obtained based on the initial abnormal region mask; and a morphological opening operation is performed on each abnormal connected region in the abnormal connected region set. The process involves obtaining an optimized set of abnormal connected regions, extracting optimized abnormal connected regions sequentially from this set, calculating the actual pixel area of ​​each optimized abnormal connected region, and if the actual pixel area is greater than a preset area threshold, then the abnormal connected region corresponding to the actual pixel area is considered a candidate region for suspected lesions. The metabolic region area of ​​the candidate region for suspected lesions is obtained, and the cervical region area and maximum fluorescence intensity are obtained based on the registered fluorescence metabolic image. The metabolic region area ratio is calculated based on the metabolic region area and the cervical region area, and the maximum fluorescence intensity and the metabolic region area ratio are used as metabolic feature parameters. The metabolic feature parameters are then summarized to obtain a set of metabolic feature parameters.

[0144] It should be explained that a registered fluorescent pixel is a single pixel in the registered fluorescent metabolic image. The pixel intensity value is the fluorescence intensity value of the registered fluorescent pixel. The global fluorescence intensity threshold is a pre-set value used to distinguish between normal metabolic pixels and suspected abnormal metabolic pixels. Suspected abnormal metabolic pixels are registered fluorescent pixels whose pixel intensity value is less than or equal to the global fluorescence intensity threshold. Normal metabolic pixels are registered fluorescent pixels whose pixel intensity value is greater than the global fluorescence intensity threshold. The initial abnormal region mask is a binary image in which suspected abnormal metabolic pixels are marked as 1 and normal metabolic pixels are marked as 0, used for preliminary identification of abnormal metabolic regions. The abnormal connected region set is the set of all connected abnormal regions in the initial abnormal region mask. This invention uses a connected component analysis algorithm to identify and label connected regions. For example, the connected component analysis algorithm is cv2.connectedComponents. Morphological opening is an image processing operation used to remove small connected regions and smooth region boundaries. Opening is prior art and will not be elaborated here. The optimized abnormal connected region set is the set of abnormal connected regions after morphological opening processing. An optimized abnormal connected region is a single connected region within the optimized abnormal connected region set. The actual pixel area is the number of pixels within the optimized abnormal connected region. The area threshold is a pre-set value used to determine whether an optimized abnormal connected region is a candidate region for a suspected lesion. A candidate region for a suspected lesion refers to an optimized abnormal connected region whose actual pixel area is greater than the area threshold; this region is considered a possible lesion region and requires further analysis. The metabolic region area is the number of pixels in the candidate region for a suspected lesion. The cervical region area is the number of pixels in the entire cervical region. The maximum fluorescence intensity is the highest pixel intensity value in the registered fluorescence metabolic image. The metabolic region area ratio is the value obtained by dividing the metabolic region area by the cervical region area, used to assess the relative size of the lesion region. The metabolic feature parameter set refers to the collection consisting of all metabolic feature parameters.

[0145] S5. Perform time-series dynamic feature analysis on the dynamic video of lesion evolution to obtain a set of time-series dynamic feature parameters, wherein the set of time-series dynamic feature parameters includes one or more time-series dynamic feature parameters.

[0146] In detail, the temporal dynamic feature analysis operation performed on the dynamic video of lesion evolution to obtain a set of temporal dynamic feature parameters includes:

[0147] Extract disease evolution image frames sequentially from the dynamic video of disease evolution, and perform image difference operation between the disease evolution image frames and the reference image to obtain the difference map;

[0148] Based on the difference map, a set of connected regions is obtained. Connected regions are extracted sequentially from the set of connected regions. The extracted connected regions are used as the acetic whitening reaction regions. The area of ​​the acetic whitening reaction region and the average pixel intensity of the region are calculated.

[0149] The area of ​​the acetic acid whitening reaction region and the average pixel intensity of the region are summarized to obtain the set of the acetic acid whitening reaction region area and the set of the average pixel intensity of the region.

[0150] By sorting the area set of the acetic acid whitening reaction region in chronological order using multiple acetic acid reaction time points, a sequence of acetic acid whitening reaction region areas is obtained.

[0151] The occurrence time of the acetic whitening reaction is determined based on the area sequence of the acetic whitening reaction region and the preset minimum area threshold, and the expansion rate of the acetic whitening reaction region is obtained based on the area sequence of the acetic whitening reaction region.

[0152] The peak time of the acetic acid whitening reaction and the rate of change of the intensity of the acetic acid whitening reaction are obtained based on the average pixel intensity set of the region. The temporal dynamic characteristic parameters are determined based on the occurrence time of the acetic acid whitening reaction, the peak time of the acetic acid whitening reaction, the rate of change of the intensity of the acetic acid whitening reaction and the expansion speed of the acetic acid whitening reaction region.

[0153] By summarizing the time-series dynamic feature parameters, a set of time-series dynamic feature parameters is obtained.

[0154] It should be explained that image differencing refers to calculating the pixel differences between two images, typically used to detect changes between images. Image differencing highlights the differences between a lesion evolution image frame and a baseline image, facilitating subsequent analysis. A difference map is the image obtained after image differencing, used to identify and locate changed regions in the image, such as changes in lesion areas. A connected region set is the collection of all connected difference regions in the difference map. The method for obtaining the connected region set based on the difference map is the same as the method for obtaining the abnormal connected region set based on registered fluorescence metabolic images, and will not be repeated here. An acetic acid whitening reaction region refers to the area where the lesion tissue changes color under the action of acetic acid, used to assess the evolution of the lesion. The area of ​​the acetic acid whitening reaction region refers to the number of pixels in the acetic acid whitening reaction region. The average pixel intensity of the region refers to the average intensity value of all pixels in the acetic acid whitening reaction region. The set of acetic acid whitening reaction region areas is the collection of all acetic acid whitening reaction region areas. The set of average pixel intensity of the region is the collection of all average pixel intensities of the region.

[0155] Importantly, the step of determining the occurrence time of the acetic whitening reaction based on the acetic whitening reaction area sequence and a preset minimum area threshold is as follows: the time corresponding to the first occurrence of the acetic whitening reaction area in the acetic whitening reaction area sequence that exceeds the minimum area threshold is taken as the occurrence time of the acetic whitening reaction. The minimum area threshold is a preset value, which is used to determine whether the area is a true lesion area by comparing it with the acetic whitening reaction area area. The occurrence time of the acetic whitening reaction refers to the time point when the area of ​​the acetic whitening reaction area first exceeds the minimum area threshold. The step of obtaining the expansion rate of the acetic whitening reaction area based on the acetic whitening reaction area sequence is as follows: according to the acetic whitening reaction area sequence and the time point corresponding to each acetic whitening reaction area in the acetic whitening reaction area sequence, the slope of the area-time change curve is obtained by linear regression fitting, and the slope of the area-time change curve is taken as the expansion rate of the acetic whitening reaction area. The steps for obtaining the peak time of the acetic acid whitening reaction and the rate of change of acetic acid whitening reaction intensity based on the regional average pixel intensity set are as follows: Based on the regional average pixel intensity set, the slope of the intensity-time change curve is obtained through linear regression fitting; the slope of the intensity-time change curve is used as the rate of change of acetic acid whitening reaction intensity; and the regional average pixel intensity with the largest value in the regional average pixel intensity set is used as the peak time of the acetic acid whitening reaction. The temporal dynamic feature parameter set refers to the set composed of all temporal dynamic feature parameters.

[0156] S6. Obtain lesion risk grading results based on metabolic feature parameter set and time-series dynamic feature parameter set, and complete fluorescence metabolic enhancement imaging based on three-modal fusion based on lesion risk grading results.

[0157] Importantly, the acquisition of lesion risk grading results based on metabolic feature parameter sets and temporal dynamic feature parameter sets involves inputting these sets into a pre-constructed multi-parameter risk classification model to obtain lesion risk grading results. These results include low, medium, and high risk. The multi-parameter risk classification model is a model pre-trained with a large number of past metabolic feature parameter sets and temporal dynamic feature parameter sets. It integrates multiple objective quantitative parameters to arrive at a qualitative risk level assessment that can be used for clinical decision-making. By comprehensively utilizing information from trimodal imaging and combining metabolic and temporal dynamic features for analysis, trimodal fusion fluorescence metabolic enhancement imaging is achieved. This fusion imaging method fully leverages the advantages of different imaging modalities, providing more comprehensive and accurate information for the diagnosis of vaginal and cervical lesions, improving the detection rate and diagnostic accuracy of early lesions, and possessing significant clinical application value.

[0158] To address the problems described in the background art, this invention provides a three-modal colposcopy imaging device. This device includes: a white light imaging mode, a polarized light imaging mode, a fluorescence imaging mode, a triple automatic image acquisition mode, and a laser ranging module. These multiple imaging modes complement each other, providing more comprehensive and richer information for subsequent lesion diagnosis. Illumination uniformity calibration is performed on the three-modal colposcopy imaging device to obtain a calibrated imaging device. Based on the calibrated imaging device, a three-modal registration image set is acquired, comprising: a reference image, a registration polarized light image, and a registration fluorescence image. In metabolic imaging, the illumination uniformity calibration operation of this invention can eliminate illumination differences at different positions and angles of the device, ensuring that the acquired image has uniform brightness and contrast throughout the entire field of view. This avoids image distortion or loss of detail caused by uneven illumination, improving image clarity and quality, and facilitating more accurate observation and analysis of lesion characteristics. Based on multiple preset acetic acid treatment time points, dynamic monitoring is performed to obtain a dynamic video of lesion evolution. This dynamic video includes multiple lesion evolution image frames. The acetic acid of this invention can change the color and morphology of lesion tissue. By performing dynamic monitoring at multiple acetic acid treatment time points… This invention employs dynamic monitoring to capture the evolution of lesions under the influence of acetic acid. The dynamic video of lesion evolution contains multiple image frames, allowing observation of the entire process from the initial state to gradual changes, revealing the development trend and characteristics of the lesion. Lesion region identification and analysis are performed on the registered fluorescence metabolic images to obtain a set of metabolic feature parameters, which includes one or more metabolic feature parameters. These metabolic feature parameters provide important basis for the diagnosis and assessment of lesions. Different types and stages of lesions may have different metabolic characteristics. By analyzing these parameters, doctors can distinguish between normal and diseased tissues, determine the benign or malignant nature and development stage of the lesion, and provide a reference for developing treatment plans. Temporal dynamic feature analysis is performed on the dynamic video of lesion evolution to obtain a set of temporal dynamic feature parameters, which also includes one or more temporal dynamic feature parameters. This invention combines metabolic feature parameters and temporal dynamic feature parameters to assess lesions from different perspectives, improving the accuracy and reliability of diagnosis. Based on the metabolic feature parameter set and the temporal dynamic feature parameter set, lesion risk grading results are obtained. Based on the lesion risk grading results, fluorescence metabolic enhancement imaging based on three-modal fusion is completed. Therefore, the present invention can improve imaging results and the sensitivity of lesion detection.

[0159] like Figure 2 The diagram shown is a functional block diagram of a fluorescence metabolism enhancement imaging system based on three-modal fusion provided in an embodiment of the present invention.

[0160] The fluorescence metabolism enhancement imaging system 100 based on trimodal fusion described in this invention can be installed in an electronic device. Depending on the functions implemented, the fluorescence metabolism enhancement imaging system 100 based on trimodal fusion may include an imaging device calibration module 101, a dynamic monitoring module 102, a lesion area identification module 103, and a time-series dynamic feature analysis module 104. The module described in this invention can also be referred to as a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and which are stored in the memory of the electronic device.

[0161] The imaging device calibration module 101 is used to identify the three-modal colposcopy imaging device, which includes: white light imaging mode, polarized light imaging mode, fluorescence imaging mode, three-link automatic image acquisition mode and laser ranging module. The module performs an illumination uniformity calibration operation on the three-modal colposcopy imaging device to obtain a calibrated imaging device. Based on the calibrated imaging device, a three-modal registration image set is obtained, which includes: a reference image, a registration polarized light image and a registration fluorescence metabolism image.

[0162] The dynamic monitoring module 102 is used to perform dynamic monitoring based on multiple preset acetic acid action time points to obtain a dynamic video of lesion evolution, wherein the dynamic video of lesion evolution includes multiple lesion evolution image frames.

[0163] The lesion region identification module 103 is used to perform lesion region identification and analysis on the registered fluorescence metabolic image to obtain a set of metabolic feature parameters, wherein the set of metabolic feature parameters includes one or more metabolic feature parameters.

[0164] The temporal dynamic feature analysis module 104 is used to perform temporal dynamic feature analysis on the dynamic video of lesion evolution to obtain a temporal dynamic feature parameter set, wherein the temporal dynamic feature parameter set includes one or more temporal dynamic feature parameters. Based on the metabolic feature parameter set and the temporal dynamic feature parameter set, the lesion risk classification result is obtained, and based on the lesion risk classification result, fluorescence metabolic enhancement imaging based on three-modal fusion is completed.

[0165] In detail, the modules in the fluorescence metabolism enhancement imaging system 100 based on trimodal fusion described in this embodiment of the invention employ the same methods as described above. Figure 1 The method uses the same techniques as the three-modal fusion-based fluorescence metabolism enhancement imaging method described in the article and can produce the same technical effects, so it will not be repeated here.

[0166] like Figure 3 The diagram shown is a schematic representation of an electronic device for implementing a fluorescence metabolism enhancement imaging method based on three-modal fusion, according to an embodiment of the present invention.

[0167] The electronic device 1 may include a processor 10, a memory 11 and a bus 12, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a program for a fluorescence metabolism enhancement imaging method based on trimodal fusion.

[0168] The memory 11 includes at least one type of readable storage medium, such as flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as a portable hard drive. In other embodiments, the memory 11 can be an external storage device of the electronic device 1, such as a plug-in portable hard drive, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device 1. Furthermore, the memory 11 includes both internal storage units and external storage devices of the electronic device 1. The memory 11 can be used not only to store application software and various types of data installed on the electronic device 1, such as the code of a fluorescence metabolism enhancement imaging method program based on trimodal fusion, but also to temporarily store data that has been output or will be output.

[0169] In some embodiments, the processor 10 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control unit of the electronic device, connecting various components of the entire electronic device through various interfaces and lines. It executes programs or modules stored in the memory 11 (e.g., a fluorescence metabolism enhancement imaging method program based on trimodal fusion) and calls data stored in the memory 11 to perform various functions of the electronic device 1 and process data.

[0170] The bus 12 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 can be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize the connection and communication between the memory 11 and at least one processor 10, etc.

[0171] Figure 3 Only electronic devices with components are shown; it will be understood by those skilled in the art that... Figure 3 The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0172] For example, although not shown, the electronic device 1 may also include a power supply (such as a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, thereby enabling functions such as charging management, discharging management, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 1 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0173] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a Wi-Fi interface, a Bluetooth interface, etc.), which is typically used to establish communication connections between the electronic device 1 and other electronic devices.

[0174] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), and optionally, a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the electronic device 1 and to display a visual user interface.

[0175] The program for the fluorescence metabolism enhancement imaging method based on trimodal fusion, stored in the memory 11 of the electronic device 1, is a combination of multiple instructions. When run in the processor 10, it can achieve the following:

[0176] Upon receiving a high-temperature performance analysis command, the titanium rod testing device is activated according to the high-temperature performance analysis command. The titanium rod testing device consists of a titanium rod fixing unit, a data acquisition unit, a surface friction unit, and a high-temperature testing unit.

[0177] Multiple sets of titanium rods to be tested were obtained. The common feature of each set of titanium rods is that they are all alloy coating substrates, and the difference is that they have coating surfaces with different materials.

[0178] The following operations were performed on each group of titanium rods to be tested in sequence:

[0179] The titanium rod to be tested is fixed to the titanium rod fixing unit. After the fixing operation is completed, the monitoring equipment of the data acquisition unit is started at the same time. The titanium rod fixing unit is in a semi-closed state with three openings on the top, namely the first opening, the second opening and the third opening. The first opening is directly opposite the monitoring equipment, and a filter is set between the first opening and the monitoring equipment.

[0180] The original titanium rod image is obtained by using monitoring equipment to capture the titanium rod under test. After the original titanium rod image is transmitted back to the data acquisition unit, the time and force of the surface friction unit are set, and surface friction is performed on the titanium rod under test to obtain the friction titanium rod.

[0181] The friction titanium rod is photographed using monitoring equipment to obtain a friction titanium rod image. After the friction titanium rod image is transmitted back to the data acquisition unit, the high temperature test unit is started. The high temperature test unit consists of a laser generator and an airflow generator.

[0182] After aligning the laser generator with the second opening and the airflow generator with the third opening, the laser generator emits a laser through the second opening to the friction titanium rod to obtain a high-temperature titanium rod. The high-temperature titanium rod is then photographed using monitoring equipment to obtain an image of the high-temperature titanium rod.

[0183] After the high-temperature titanium rod image is transmitted back to the data acquisition unit, an airflow generator is used to generate airflow through the third opening to the titanium rod fixing unit, and a monitoring device is used to capture the high-temperature titanium rod affected by the airflow to obtain the airflow titanium rod image.

[0184] The original titanium rod image, friction titanium rod image, high temperature titanium rod image, and airflow titanium rod image are used as input data for a pre-built surface detection model. The surface damage degree is obtained using the surface detection model, which is built based on a convolutional neural network.

[0185] The optimal titanium rod was selected from multiple groups of titanium rods to be tested based on the degree of surface damage.

[0186] Specifically, the processor 10's implementation method for the above instructions can be found in [reference needed]. Figures 1 to 3 The descriptions of the relevant steps in the corresponding embodiments are not repeated here.

[0187] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0188] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor of an electronic device, can perform the following:

[0189] Upon receiving a high-temperature performance analysis command, the titanium rod testing device is activated according to the high-temperature performance analysis command. The titanium rod testing device consists of a titanium rod fixing unit, a data acquisition unit, a surface friction unit, and a high-temperature testing unit.

[0190] Multiple sets of titanium rods to be tested were obtained. The common feature of each set of titanium rods is that they are all alloy coating substrates, and the difference is that they have coating surfaces with different materials.

[0191] The following operations were performed on each group of titanium rods to be tested in sequence:

[0192] The titanium rod to be tested is fixed to the titanium rod fixing unit. After the fixing operation is completed, the monitoring equipment of the data acquisition unit is started at the same time. The titanium rod fixing unit is in a semi-closed state with three openings on the top, namely the first opening, the second opening and the third opening. The first opening is directly opposite the monitoring equipment, and a filter is set between the first opening and the monitoring equipment.

[0193] The original titanium rod image is obtained by using monitoring equipment to capture the titanium rod under test. After the original titanium rod image is transmitted back to the data acquisition unit, the time and force of the surface friction unit are set, and surface friction is performed on the titanium rod under test to obtain the friction titanium rod.

[0194] The friction titanium rod is photographed using monitoring equipment to obtain a friction titanium rod image. After the friction titanium rod image is transmitted back to the data acquisition unit, the high temperature test unit is started. The high temperature test unit consists of a laser generator and an airflow generator.

[0195] After aligning the laser generator with the second opening and the airflow generator with the third opening, the laser generator emits a laser through the second opening to the friction titanium rod to obtain a high-temperature titanium rod. The high-temperature titanium rod is then photographed using monitoring equipment to obtain an image of the high-temperature titanium rod.

[0196] After the high-temperature titanium rod image is transmitted back to the data acquisition unit, an airflow generator is used to generate airflow through the third opening to the titanium rod fixing unit, and a monitoring device is used to capture the high-temperature titanium rod affected by the airflow to obtain the airflow titanium rod image.

[0197] The original titanium rod image, friction titanium rod image, high temperature titanium rod image, and airflow titanium rod image are used as input data for a pre-built surface detection model. The surface damage degree is obtained using the surface detection model, which is built based on a convolutional neural network.

[0198] The optimal titanium rod was selected from multiple groups of titanium rods to be tested based on the degree of surface damage.

[0199] In the embodiments provided by this invention, it should be understood that the disclosed devices, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and actual implementations may have other classification methods.

[0200] 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 modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0201] Furthermore, the functional modules in the various embodiments of the present invention 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. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0202] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0203] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A fluorescence metabolic enhanced imaging method based on three-modal fusion, characterized in that, The method comprises: Confirming a three-modal colposcope imaging device, wherein the three-modal colposcope imaging device comprises a white light imaging mode, a polarized light imaging mode, a fluorescence imaging mode, a triple automatic image acquisition mode, and a laser ranging module; Performing an illumination uniformity calibration operation on the three-modal colposcope imaging device to obtain a calibrated imaging device, and obtaining a three-modal registration image set based on the calibrated imaging device, wherein the three-modal registration image set comprises a reference image, a registration polarized light image, and a registration fluorescence metabolic image; Performing dynamic monitoring based on a plurality of preset acetic acid action time points to obtain a lesion evolution dynamic video, wherein the lesion evolution dynamic video comprises a plurality of lesion evolution image frames; Performing lesion region identification analysis on the registration fluorescence metabolic image to obtain a metabolic feature parameter set, wherein the metabolic feature parameter set comprises one or more metabolic feature parameters; Performing time sequence dynamic feature analysis on the lesion evolution dynamic video to obtain a time sequence dynamic feature parameter set, wherein the time sequence dynamic feature parameter set comprises one or more time sequence dynamic feature parameters; Obtaining a lesion risk grading result based on the metabolic feature parameter set and the time sequence dynamic feature parameter set, and completing fluorescence metabolic enhanced imaging based on three-modal fusion based on the lesion risk grading result.

2. The method of claim 1, wherein the method is a three-modality fusion based fluorescence metabolic enhancement imaging method. The method comprises: Using the laser ranging module and the white light imaging mode to automatically focus the calibrated imaging device to obtain a focused imaging device; Obtaining an initial white light cervical region image according to the focused imaging device and a pre-constructed AI image recognition algorithm, performing reflection analysis on the initial white light cervical region image to obtain an analysis result, wherein the analysis result is whether reflection exists or not; If the analysis result is that reflection exists, obtaining a reflection region image according to the initial white light cervical region image, performing intelligent reflection elimination on the reflection region image in the initial white light cervical region image using the polarized light imaging mode to obtain a white light cervical region image, calculating a reflection interference index according to the initial white light cervical region image and the white light cervical region image, and identifying the white light cervical region image using the reflection interference index to obtain an identified cervical region image; If the analysis result is that reflection does not exist, using the initial white light cervical region image as the identified cervical region image; Obtaining the three-modal registration image set according to the identified cervical region image and the focused imaging device.

3. The method of claim 2, wherein the three-modality fusion based fluorescence metabolic enhancement imaging is characterized by, The method comprises: Statistically obtaining the initial pixel number and the reflection pixel number of the initial white light cervical region image and the reflection region image respectively, and calculating a reflection area proportion according to the initial pixel number and the reflection pixel number; Obtaining a reflection region pixel set of the reflection region image, and determining whether the reflection region image is a preset gray image; If the reflection region image is not the preset gray image, performing gray scale conversion on each reflection region pixel in the reflection region pixel set to obtain a reflection gray scale pixel set; According to the reflection gray pixel set, a reflection gray pixel value set is obtained, wherein the reflection gray pixel set includes a plurality of reflection gray pixels, the reflection gray pixel value set includes a plurality of reflection gray pixel values, and the reflection gray pixels and the reflection gray pixel values correspond one by one; According to the reflection gray pixel value set, a reflection gray pixel value quantity and a gray pixel value total sum are calculated, and an average brightness value is calculated according to the reflection gray pixel value quantity and the gray pixel value total sum, wherein the average brightness value is a value obtained by dividing the gray pixel value total sum by the reflection gray pixel value quantity; The reflection area proportion and the average brightness value are weighted and summed to obtain a reflection interference index.

4. The fluorescence metabolism enhancement imaging method based on three-modal fusion as described in claim 3, characterized in that, The three-modality registration image set is obtained according to the identified cervical region image and the focusing imaging device, and includes: A pseudo-color metabolic superposition image is obtained according to the fluorescence imaging mode and the identified cervical region image, a positioning instruction is received, the focusing imaging device is adjusted according to the positioning instruction and the pseudo-color metabolic superposition image, and a positioned imaging device is obtained; The positioned imaging device is set by using a preset acquisition cycle to obtain a set imaging device, a three-in-one automatic image acquisition mode in the set imaging device is started, and white light form images, polarized light detail images and fluorescence metabolic images are acquired by using the set imaging device after the three-in-one automatic image acquisition mode is started; The white light form images, the polarized light detail images and the fluorescence metabolic images are subjected to multi-modality image space-time registration operation to obtain the three-modality registration image set.

5. The fluorescence metabolism enhancement imaging method based on three-modal fusion as described in claim 4, characterized in that, The three-modality registration image set is obtained by performing multi-modality image space-time registration operation on the white light form images, the polarized light detail images and the fluorescence metabolic images, and includes: The white light form image is taken as a reference image, feature points are extracted from the reference image, the polarized light detail image and the fluorescence metabolic image respectively to obtain a reference image feature point group, a polarized light image feature point group and a fluorescence metabolic image feature point group; Polarized light image feature points are sequentially extracted from the polarized light image feature point group, the polarized light image feature points are subjected to feature point matching operation with the reference image feature point group to obtain a plurality of matching feature point pairs, and the plurality of matching feature point pairs are summarized to obtain a matching feature point pair group; An optimal space transformation matrix is obtained according to the matching feature point pair group, a registration polarized light image is obtained according to the optimal space transformation matrix and the polarized light detail image, and a registration fluorescence metabolic image is obtained based on the fluorescence metabolic image feature point group and the reference image feature point group; The three-modality registration image set is confirmed based on the reference image, the registration polarized light image and the registration fluorescence metabolic image.

6. The fluorescence metabolism enhancement imaging method based on three-modal fusion as described in claim 5, characterized in that, The registration polarized light image is obtained according to the optimal space transformation matrix and the polarized light detail image, and includes: A blank canvas is created according to the reference image, and the blank canvas is taken as a target polarized light registration image, wherein the target polarized light registration image includes a plurality of target pixels; Target pixels are sequentially extracted from the target polarized light registration image, target pixel coordinates are confirmed according to the target pixels, an inverse transformation matrix is obtained according to the optimal space transformation matrix, and source pixel coordinates are calculated according to the inverse transformation matrix and the target pixel coordinates, wherein the source pixel coordinates include an abscissa and an ordinate. If the source pixel coordinate is a preset integer coordinate, a target source pixel value is obtained from the polarized light detail image according to the source pixel coordinate, the target pixel is valued by using the target source pixel value, and a valued target pixel is obtained; If the source pixel coordinate is a preset floating point coordinate, a horizontal coordinate in the source pixel coordinate is subjected to a horizontal coordinate rounding operation to obtain a down-rounded horizontal coordinate and an up-rounded horizontal coordinate, and a vertical coordinate in the source pixel coordinate is subjected to a vertical coordinate rounding operation to obtain a down-rounded vertical coordinate and an up-rounded vertical coordinate; The padding pixel value is calculated based on the down-rounded horizontal coordinate, the up-rounded horizontal coordinate, the down-rounded vertical coordinate and the up-rounded vertical coordinate, the target pixel is valued by using the padding pixel value, and a valued target pixel is obtained; The valued target pixels are summarized to obtain the registered polarized light image.

7. The fluorescence metabolism enhancement imaging method based on three-modal fusion as described in claim 6, characterized in that, The source pixel coordinate is calculated according to the inverse transformation matrix and the target pixel coordinate, and includes: The target pixel coordinate is converted to obtain a homogeneous pixel coordinate, and the source pixel coordinate is calculated according to the homogeneous pixel coordinate and the inverse transformation matrix, wherein the calculation formula of the source pixel coordinate is as follows: ， wherein denotes the homogeneous pixel coordinate, denotes the horizontal coordinate in the homogeneous pixel coordinate, denotes the vertical coordinate in the homogeneous pixel coordinate, denotes the inverse transformation matrix, denotes the horizontal coordinate of the source pixel coordinate, denotes the vertical coordinate of the source pixel coordinate.

8. The fluorescence metabolism enhancement imaging method based on three-modal fusion as described in claim 7, characterized in that, The padding pixel value is calculated based on the down-rounded horizontal coordinate, the up-rounded horizontal coordinate, the down-rounded vertical coordinate and the up-rounded vertical coordinate, and includes: The down-rounded horizontal coordinate and the down-rounded vertical coordinate are combined to obtain a first down combination coordinate, and the up-rounded horizontal coordinate and the down-rounded vertical coordinate are combined to obtain a second down combination coordinate; The first down pixel value and the second down pixel value are confirmed from the polarized light detail image according to the first down combination coordinate and the second down combination coordinate; The horizontal coordinate of the source pixel coordinate is interpolated by using the first down pixel value and the second down pixel value to obtain a down interpolation; The up-rounded horizontal coordinate and the up-rounded vertical coordinate are combined to obtain a first up combination coordinate, and the down-rounded horizontal coordinate and the up-rounded vertical coordinate are combined to obtain a second up combination coordinate; The first up pixel value and the second up pixel value are confirmed from the polarized light detail image according to the first up combination coordinate and the second up combination coordinate; The horizontal coordinate of the source pixel coordinate is interpolated by using the first up pixel value and the second up pixel value to obtain an up interpolation; The padding pixel value is calculated according to the down interpolation, the up interpolation and the vertical coordinate of the source pixel coordinate, wherein the calculation formula of the padding pixel value is as follows: ， wherein, represents a padding pixel value, represents a downward interpolation, represents a rounding up of a vertical coordinate, represents a vertical coordinate of a source pixel coordinate, represents an upward interpolation, represents a rounding down of a vertical coordinate.

9. The fluorescence metabolism enhancement imaging method based on three-modal fusion as described in claim 8, characterized in that, The time sequence dynamic feature analysis operation is performed on the lesion evolution dynamic video to obtain a time sequence dynamic feature parameter set, and includes: The lesion evolution image frames are sequentially extracted from the lesion evolution dynamic video, and the lesion evolution image frames are subjected to image difference operation with the reference image to obtain difference images; The connected region set is obtained based on the difference images, and the connected regions are sequentially extracted from the connected region set, the extracted connected regions are taken as aceto-white reaction regions, and the aceto-white reaction region area and the region average pixel intensity of the aceto-white reaction regions are calculated; The aceto-white reaction region area and the region average pixel intensity are respectively summarized to obtain an aceto-white reaction region area set and a region average pixel intensity set; The aceto-white reaction region area set is sorted according to the time sequence by using the plurality of acetic acid action time points to obtain an aceto-white reaction region area sequence; Confirm the appearance time of the acetowhite reaction based on the sequence of the area of the acetowhite reaction region and a preset minimum area threshold, and obtain the expansion speed of the acetowhite reaction region based on the sequence of the area of the acetowhite reaction region; Obtain the peak time of the acetowhite reaction and the intensity change rate of the acetowhite reaction based on the set of average pixel intensities of the region, and confirm the time sequence dynamic characteristic parameters according to the appearance time of the acetowhite reaction, the peak time of the acetowhite reaction, the intensity change rate of the acetowhite reaction, and the expansion speed of the acetowhite reaction region. The time sequence dynamic characteristic parameter set is obtained by summarizing the time sequence dynamic characteristic parameters.

10. A fluorescence metabolic enhanced imaging system based on three-modal fusion, characterized in that, The system comprises: An imaging device calibration module is configured to confirm a three-modal colposcope imaging device, wherein the three-modal colposcope imaging device comprises a white light imaging mode, a polarized light imaging mode, a fluorescence imaging mode, a triple automatic image acquisition mode, and a laser ranging module. The illumination uniformity calibration operation is performed on the three-modal colposcope imaging device to obtain a calibrated imaging device. The three-modal registration image set is obtained based on the calibrated imaging device, wherein the three-modal registration image set comprises a reference image, a registered polarized light image, and a registered fluorescence metabolic image. A dynamic monitoring module is configured to perform dynamic monitoring based on a plurality of preset acetic acid action time points to obtain a lesion evolution dynamic video, wherein the lesion evolution dynamic video comprises a plurality of lesion evolution image frames. A lesion region identification module is configured to perform lesion region identification analysis operation on the registered fluorescence metabolic image to obtain a metabolic characteristic parameter set, wherein the metabolic characteristic parameter set comprises one or more metabolic characteristic parameters. A time sequence dynamic characteristic analysis module is configured to perform time sequence dynamic characteristic analysis operation on the lesion evolution dynamic video to obtain a time sequence dynamic characteristic parameter set, wherein the time sequence dynamic characteristic parameter set comprises one or more time sequence dynamic characteristic parameters. The lesion risk grading result is obtained based on the metabolic characteristic parameter set and the time sequence dynamic characteristic parameter set. The fluorescence metabolic enhanced imaging based on three-modal fusion is completed based on the lesion risk grading result.

Images