AI-based intestinal preparation quality evaluation system

Abstract

The invention discloses an AI-based intestinal preparation quality evaluation system, and belongs to the technical field of enteroscopy, and the system comprises an acquisition and processing module which is configured to acquire an excrement fluorescence image of a patient, carry out the preprocessing and feature extraction of the excrement fluorescence image, and determine excrement feature data; the model building module is configured to build an intestinal tract preparation quality evaluation model; the quality evaluation module is configured to perform feature recognition and analysis on the excrement feature data, perform quality evaluation on intestinal preparation of the patient and determine an intestinal preparation quality evaluation result; and the output management module is configured to visually output the intestinal preparation quality evaluation report of the patient and perform subsequent enteroscopy management on the patient. The problems that existing intestinal preparation quality evaluation has certain inaccuracy and subjectivity, and the qualification rate of intestinal preparation is as low as 70% are solved. According to the method, a more objective and efficient intestinal preparation quality evaluation result can be provided, so that the accuracy and efficiency of enteroscopy are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of colonoscopy, and in particular to an AI-based intestinal preparation quality assessment system. Background Art

[0002] Bowel preparation before colonoscopy is a key step to ensure the smooth progress of the examination and improve the accuracy of diagnosis. Poor bowel preparation can lead to the following situations: 1) unclear vision, easy to miss the diagnosis, and literature shows that the missed diagnosis rate is as high as 45.7% with poor bowel preparation; because feces and food residues in the intestines will block the intestinal mucosa, affecting the doctor's observation of subtle lesions in the intestinal wall, such as polyps, ulcers, tumors, etc., especially small polyps or early cancers, delaying the treatment time; 2) reduced examination efficiency and safety. A poorly prepared intestine is not conducive to the smooth passage of the colonoscope, increasing operation time and patient discomfort, and increasing the occurrence of complications. Intestinal residues may make it difficult to advance the colonoscope and increase the risk of intestinal wall damage; 3) increased treatment costs. If polyps or bleeding spots are found, the doctor can directly remove or stop bleeding during the examination to avoid secondary preparation and operation. However, insufficient preparation may require the examination to be rescheduled, increasing time and economic costs.

[0003] Therefore, the importance of bowel preparation is self-evident. However, the cleanliness of bowel preparation is currently judged subjectively by medical staff, which is somewhat inaccurate and subjective. Moreover, patients who do bowel preparation at home often have inadequate bowel preparation during colonoscopy, and the pass rate of bowel preparation is as low as 70%. Summary of the Invention

[0004] The purpose of the present invention is to provide an AI-based intestinal preparation quality assessment system that can provide more objective and efficient intestinal preparation quality assessment results, thereby improving the accuracy and efficiency of colonoscopy and solving the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: An AI-based bowel preparation quality assessment system, including: an acquisition and processing module configured to acquire a fluorescent image of a patient's excrement, perform preprocessing and feature extraction on the fluorescent image, and determine characteristic data of the excrement; a model building module configured to build a bowel preparation quality assessment model according to bowel preparation quality assessment requirements; a quality assessment module configured to perform feature recognition and analysis on fecal characteristic data according to a bowel preparation quality assessment model, perform a quality assessment on the patient's bowel preparation, and determine a bowel preparation quality assessment result; The output management module is configured to visually output the patient's bowel preparation quality assessment report and conduct subsequent colonoscopy management for the patient.

[0006] Preferably, collecting a fluorescent image of the patient's excrement includes: Collecting the patient's fecal sample in a sterile container, and diluting or centrifuging the patient's fecal sample to remove impurities in the fecal sample and concentrate the target component; Based on fluorescence imaging technology, a fluorescent dye is selected according to the characteristics of the target molecule in the fecal sample, and the fluorescent dye is mixed with the fecal sample to ensure full contact between the two; Select reaction conditions, including pH, temperature, and time, based on the target molecule in the fecal sample. Wash the fecal sample with a buffer to remove unbound dye. The labeled sample is purified by centrifugation or filtration, placed in a fluorescence microscope, and an excitation wavelength and filter combination are selected to obtain a fluorescent image of the patient's excrement.

[0007] Preferably, preprocessing the patient's fecal fluorescence image includes: Denoising the patient's fecal fluorescence image based on a filter to remove noise and reduce background interference in the fecal fluorescence image; Performing contrast enhancement processing on the patient's fecal fluorescence image based on histogram equalization to enhance details in the fecal fluorescence image and highlight the fluorescence signal of the target area in the fecal fluorescence image; Image segmentation processing is performed on the patient's fecal fluorescence image based on the threshold segmentation method to separate the background and target areas in the fecal fluorescence image and eliminate nonspecific fluorescence interference.

[0008] Preferably, setting a threshold for image segmentation processing includes: extracting the contrast value before and after contrast enhancement processing; Obtaining a contrast enhancement amplitude according to a contrast value before and after the contrast enhancement process; Performing a ratio processing on the contrast enhancement amplitude and the contrast value before the contrast enhancement processing to obtain a contrast enhancement ratio; comparing the contrast enhancement ratio with a preset contrast enhancement ratio threshold; When the contrast enhancement ratio does not exceed a preset contrast enhancement ratio threshold, a preset initial threshold parameter for image segmentation processing is retrieved; and image segmentation processing is performed on the patient's fecal fluorescence image using the preset initial threshold parameter for image segmentation processing based on a threshold segmentation method; When the contrast enhancement ratio exceeds a preset contrast enhancement ratio threshold, the initial threshold parameter is adjusted using the grayscale values of the fecal fluorescence images before and after contrast enhancement; Based on the threshold segmentation method, the patient's fecal fluorescence image is segmented using the adjusted threshold parameters.

[0009] Preferably, when the contrast enhancement ratio exceeds a preset contrast enhancement ratio threshold, the initial threshold parameter is adjusted using the grayscale values of the fecal fluorescence images before and after contrast enhancement, including: When the contrast enhancement ratio exceeds a preset contrast enhancement ratio threshold, retrieving the fecal fluorescence image before contrast enhancement processing; performing grayscale processing on the feces fluorescence image before contrast enhancement processing to obtain a first grayscale image; Retrieve the contrast-enhanced fluorescence image of the excrement; performing grayscale processing on the excrement fluorescence image after contrast enhancement processing to obtain a second grayscale image; Extracting each pixel in the first grayscale image; Extracting each pixel in the second grayscale image; For a pixel point at the same position in the first grayscale image and the second grayscale image, obtaining a grayscale difference between the pixel point at the same position; performing ratio processing on the grayscale difference of the pixel at the same position and the grayscale value of the pixel at the same position in the first grayscale image to obtain the grayscale change rate of the pixel at the same position; The initial threshold parameters are adjusted using the grayscale change rate and contrast enhancement ratio.

[0010] Preferably, feature extraction is performed on the patient's fecal fluorescence image, including: Based on the convolutional neural network (CNN), feature extraction is performed on the patient's fecal fluorescence images. Semantic features valuable for bowel preparation quality assessment are extracted from the fecal fluorescence images. The extracted features are then subjected to dimensionality reduction processing based on principal component analysis to determine the fecal characteristic data, including fecal composition, concentration, color, contrast, and texture.

[0011] Preferably, a bowel preparation quality assessment model is constructed, including: According to the needs of AI-based bowel preparation quality assessment, historical excrement data is collected and divided into training sets and test sets; Based on deep learning technology, a training set is used to train the deep learning model, allowing the deep learning model to autonomously learn bowel preparation quality assessment behaviors from the training set. This is used to assess the quality of the patient's bowel preparation, determine the cleanliness of the patient's intestines, and determine the bowel preparation quality assessment model based on deep learning; Use the test set to test the deep learning-based bowel preparation quality assessment model, evaluate its performance, determine whether it can achieve the expected effect of assessing the quality of patients' bowel preparation and judging the cleanliness of the patients' intestines, and determine the model test evaluation results; According to the model test evaluation results, the parameters of the intestinal preparation quality assessment model based on deep learning are adjusted and optimized until the intestinal preparation quality assessment model based on deep learning can achieve the expected effect of evaluating the quality of the patient's intestinal preparation and judging the cleanliness level of the patient's intestine, and then determine the optimal intestinal preparation quality assessment model.

[0012] Preferably, the patient's bowel preparation should be assessed for quality, including: Deploy the optimal bowel preparation quality assessment model and deploy it in the actual bowel preparation quality assessment environment; The fecal characteristic data is input into the intestinal preparation quality assessment model, and the fecal characteristic data is characterized and analyzed according to the intestinal preparation quality assessment model. The quality of the patient's intestinal preparation is assessed and the cleanliness of the patient's intestine is judged to determine the intestinal preparation quality assessment result.

[0013] Preferably, a visual output of the patient's bowel preparation quality assessment report is provided, and the patient is managed for subsequent colonoscopy examinations, including: Based on the intestinal preparation quality assessment results and combined with the patient's fecal fluorescent image, the patient's intestinal preparation quality assessment report is formed, and the patient's intestinal preparation quality assessment report is displayed to the doctor in real time in a visual form, so that the doctor can conduct subsequent colonoscopy management on the patient based on the patient's intestinal preparation quality assessment report.

[0014] Preferably, the cleanliness of the patient's intestine is divided into four levels based on the bowel preparation quality assessment results, including: Grade 1: The intestinal mucosa is not clearly visible and there is a lot of fecal residue in the intestine; Grade II, that is, the intestinal mucosa is clearly displayed, with a moderate amount of fecal residue in the intestine; Grade 3: The intestinal mucosa is clearly visible, with a small amount of fecal residue in the intestine; Grade 4: The intestinal mucosa is clearly visible and there is no fecal residue in the intestine; Among them, the higher the level, the better the intestinal preparation, and the more conducive it is to colonoscopy.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention collects the patient's fecal fluorescent image through fluorescence imaging technology, and preprocesses and extracts features of the patient's fecal fluorescent image to determine the fecal characteristic data, and constructs an intestinal preparation quality assessment model based on AI-based intestinal preparation quality assessment requirements. The fecal characteristic data is feature identified and analyzed according to the intestinal preparation quality assessment model, the patient's intestinal preparation quality is assessed and the cleanliness of the patient's intestine is judged to determine the intestinal preparation quality assessment result, and the patient's fecal fluorescent image is combined to form the patient's intestinal preparation quality assessment report, and the patient's intestinal preparation quality assessment report is displayed to the doctor in real time in a visual form, so that the doctor can perform subsequent colonoscopy management on the patient according to the patient's intestinal preparation quality assessment report, and can provide more objective and efficient intestinal preparation quality assessment results, thereby improving the accuracy and efficiency of colonoscopy. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a module diagram of the AI-based bowel preparation quality assessment system of the present invention; Figure 2 Flowchart of the AI-based bowel preparation quality assessment system of the present invention. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0018] In order to solve the problem that the cleanliness of bowel preparation is subjectively judged by medical staff, which is somewhat inaccurate and subjective, and patients who do bowel preparation at home often have inadequate bowel preparation during colonoscopy, with the bowel preparation pass rate as low as 70%, please refer to Figure 1-Figure 2 , this embodiment provides the following technical solutions: An AI-based bowel preparation quality assessment system includes: an acquisition and processing module, a model building module, a quality assessment module and an output management module.

[0019] Specifically, through the interaction between the acquisition and processing module, the model building module, the quality assessment module and the output management module, more objective and efficient intestinal preparation quality assessment results can be provided, thereby improving the accuracy and efficiency of colonoscopy.

[0020] The acquisition and processing module is used to collect the patient's excrement fluorescence image, perform preprocessing and feature extraction on it, and determine the excrement feature data; In this embodiment, collecting a fluorescent image of the patient's excrement includes: Collecting the patient's fecal sample in a sterile container, and diluting or centrifuging the patient's fecal sample to remove impurities in the fecal sample and concentrate the target component; A fluorescent dye is selected based on the characteristics of the target molecule in the fecal sample, and the fluorescent dye is mixed with the fecal sample to ensure that the two are in full contact; Select reaction conditions, including pH, temperature, and time, based on the target molecule in the fecal sample. Wash the fecal sample with a buffer to remove unbound dye. The labeled sample is purified by centrifugation or filtration, placed in a fluorescence microscope, and an excitation wavelength and filter combination are selected to obtain a fluorescent image of the patient's excrement.

[0021] It should be noted that fluorescence imaging can be combined to improve the sensitivity of the intestinal preparation quality assessment model to subtle residues.

[0022] In this embodiment, the patient's feces fluorescence image is preprocessed, including: Denoising the patient's fecal fluorescence image based on a filter to remove noise and reduce background interference in the fecal fluorescence image; Performing contrast enhancement processing on the patient's fecal fluorescence image based on histogram equalization to enhance details in the fecal fluorescence image and highlight the fluorescence signal of the target area in the fecal fluorescence image; Image segmentation processing is performed on the patient's fecal fluorescence image based on the threshold segmentation method to separate the background and target areas in the fecal fluorescence image and eliminate nonspecific fluorescence interference.

[0023] Specifically, setting the threshold for image segmentation processing includes: extracting the contrast value before and after contrast enhancement processing; Obtaining a contrast enhancement amplitude according to a contrast value before and after the contrast enhancement process; Performing a ratio processing on the contrast enhancement amplitude and the contrast value before the contrast enhancement processing to obtain a contrast enhancement ratio; comparing the contrast enhancement ratio with a preset contrast enhancement ratio threshold; When the contrast enhancement ratio does not exceed a preset contrast enhancement ratio threshold, a preset initial threshold parameter for image segmentation processing is retrieved; and image segmentation processing is performed on the patient's fecal fluorescence image using the preset initial threshold parameter for image segmentation processing based on a threshold segmentation method; When the contrast enhancement ratio exceeds a preset contrast enhancement ratio threshold, the initial threshold parameter is adjusted using the grayscale values of the fecal fluorescence images before and after contrast enhancement; Based on the threshold segmentation method, the patient's fecal fluorescence image is segmented using the adjusted threshold parameters.

[0024] The technical effect of the above technical solution is that contrast enhancement typically stretches the grayscale range of an image, making the grayscale difference between the target area (such as the fecal fluorescence signal) and the background more pronounced. In this case, the original threshold may no longer be applicable to the new grayscale distribution. By analyzing the grayscale values before and after enhancement, the threshold can be adjusted using the following method: The grayscale histogram of the enhanced image is calculated, the grayscale peaks of the target area and the background are found, and the intermediate value is selected as the new threshold. The initial threshold is proportionally adjusted based on the enhancement ratio and grayscale change trend. The optimal threshold is recalculated based on the local grayscale characteristics of the enhanced image. Conversely, by comparing the contrast enhancement values before and after, the contrast enhancement amplitude and ratio are calculated, and the threshold adjustment is determined accordingly. If the contrast enhancement ratio does not exceed the threshold, the initial threshold is used. This threshold is a reasonable value determined through experience or preliminary research. If it exceeds the threshold, the threshold is adjusted based on the grayscale value. This method provides a more appropriate threshold for image segmentation based on the actual contrast enhancement of the image, enabling the segmentation result to more accurately delineate different regions in the fecal fluorescence image, improving segmentation accuracy and reducing mis-segmentation. This solution can flexibly select the threshold processing method according to the actual degree of image contrast enhancement. For fecal fluorescence images with different contrast enhancement effects, appropriate strategies can be adopted. The initial threshold can be used, or the threshold can be adjusted according to the grayscale value to adapt it to a variety of imaging conditions and image features, enhancing the versatility and adaptability of the method in different scenarios. When the contrast enhancement ratio does not exceed the threshold, the preset initial threshold is directly used to avoid unnecessary threshold adjustment calculations and improve processing efficiency; when the ratio exceeds the threshold, the threshold is adjusted based on the grayscale value, which can ensure good segmentation effects when the image contrast changes greatly. This method achieves a good balance between processing efficiency and segmentation effect, improving the overall performance of image processing.

[0025] Specifically, when the contrast enhancement ratio exceeds a preset contrast enhancement ratio threshold, the initial threshold parameter is adjusted using the grayscale values of the fecal fluorescence images before and after contrast enhancement, including: When the contrast enhancement ratio exceeds a preset contrast enhancement ratio threshold, retrieving the fecal fluorescence image before contrast enhancement processing; performing grayscale processing on the feces fluorescence image before contrast enhancement processing to obtain a first grayscale image; Retrieve the contrast-enhanced fluorescence image of the excrement; performing grayscale processing on the excrement fluorescence image after contrast enhancement processing to obtain a second grayscale image; Extracting each pixel in the first grayscale image; Extracting each pixel in the second grayscale image; For a pixel point at the same position in the first grayscale image and the second grayscale image, obtaining a grayscale difference between the pixel point at the same position; performing ratio processing on the grayscale difference of the pixel at the same position and the grayscale value of the pixel at the same position in the first grayscale image to obtain the grayscale change rate of the pixel at the same position; Adjusting the initial threshold parameters using the grayscale change rate and contrast enhancement ratio; The adjusted threshold parameter is obtained by the following formula: in, represents the adjusted threshold parameter; represents the initial threshold parameter; Indicates the number of pixels contained in the excrement fluorescence image; Indicates the Grayscale change rate corresponding to each pixel; represents the contrast enhancement ratio. Specifically, Subtracting the two results in the deviation of the grayscale change of a single pixel relative to the contrast enhancement of the entire image, which is used to measure the difference between local and overall changes. use The difference divided by and This is a normalization operation. Its purpose is to convert the deviations mentioned above into relative proportions, mapping the numerical range to the interval [-1, 1]. This allows the changes in different pixels to be compared on the same scale, facilitating subsequent calculations and analysis of the performance of all pixels, and accurately reflecting the contribution of individual pixel changes to threshold adjustment. right The normalized difference values of all pixels are summed up to combine the local and overall change differences of all pixels. Then divided by the number of pixels Averaging eliminates the influence of the number of pixels, resulting in a value that represents the combined difference in pixel grayscale variation and contrast enhancement across the entire image. This value reflects the overall image characteristics in these two aspects and provides a basis for threshold adjustment. An adjustment factor is then constructed by adding 1 to this average difference value. When this factor is greater than 1, the adjusted threshold increases; when it is less than 1, the adjusted threshold decreases. In this way, the direction and magnitude of threshold adjustment are determined based on the combined grayscale variation and contrast enhancement of the image pixels. This formula comprehensively considers the grayscale variation rate of each pixel and the contrast enhancement ratio of the entire image. Through a series of calculations, the adjustment factor is derived, allowing the threshold to be accurately adjusted based on the actual image changes. For fecal fluorescence images with varying grayscale variation and contrast enhancement, an appropriate adjustment magnitude can be found to ensure that the adjusted threshold closely matches the image characteristics, improving image segmentation accuracy. Because the formula is based on the image's pixel grayscale variation and overall contrast enhancement, it can dynamically generate appropriate adjustment factors to adjust the threshold for fecal fluorescence images with various imaging conditions and grayscale distribution characteristics. This method effectively adapts to varying image scenarios, regardless of whether pixel grayscale changes are simple or complex, or whether the contrast enhancement is large or small. The normalization and averaging operations in the formula effectively prevent abnormal changes in individual pixels or fluctuations in contrast enhancement ratios from excessively interfering with threshold adjustment. This ensures a stable and reliable threshold adjustment process, preventing significant fluctuations or unreasonable changes in the adjusted threshold, thus ensuring the stability and reliability of the image segmentation results.

[0026] The technical effect of the above technical solution is as follows: First, grayscale processing is performed on the fecal fluorescence images before and after contrast enhancement, respectively, to obtain a first grayscale image and a second grayscale image. Pixels at the same location in the two images are then extracted, the grayscale difference is calculated, and the grayscale change rate is further calculated. The grayscale change rate reflects the relative change in the grayscale value of each pixel during the contrast enhancement process and is a key indicator for determining the degree of change in image content. The contrast enhancement ratio measures the degree of contrast enhancement for the entire image. The grayscale change rate and contrast enhancement ratio are combined because they reflect image changes from different perspectives. The grayscale change rate reflects local pixel changes, while the contrast enhancement ratio reflects overall contrast changes. Combining these two information provides a more comprehensive understanding of the changes in image characteristics after contrast enhancement. Based on these changes, the initial threshold parameters are adjusted to better match the actual grayscale distribution of the enhanced image, thereby meeting the requirements for accurate image segmentation. By adjusting the threshold based on both the grayscale change rate and the contrast enhancement ratio, the grayscale characteristics of the contrast-enhanced image can be more accurately adapted. During image segmentation, different areas in the image can be distinguished more accurately, mis-segmentation caused by unreasonable thresholds can be reduced, the consistency between the segmentation results and the actual image content can be improved, and the accuracy of image segmentation can be improved. This technical solution can dynamically adjust the threshold according to the actual situation of contrast enhancement of different fecal fluorescence images (reflected by the grayscale change rate and contrast enhancement ratio). Regardless of the magnitude of image contrast enhancement or the grayscale change characteristics, a suitable threshold can be found, which enhances the adaptability of the method to various imaging conditions and image features, allowing it to work effectively in different scenarios. Compared with fixed thresholds or adjusting thresholds based on a single indicator, adjusting the threshold based on the grayscale change rate and contrast enhancement ratio can avoid the problem of poor segmentation results caused by ignoring local or overall image change characteristics. The threshold can be comprehensively optimized from multiple dimensions to make the effect of image segmentation more in line with actual needs and improve the overall quality of image processing.

[0027] In this embodiment, feature extraction is performed on the patient's feces fluorescence image, including: Based on the convolutional neural network (CNN), feature extraction is performed on the patient's fecal fluorescence images. Semantic features valuable for bowel preparation quality assessment are extracted from the fecal fluorescence images. The extracted features are then subjected to dimensionality reduction processing based on principal component analysis to determine the fecal characteristic data, including fecal composition, concentration, color, contrast, and texture.

[0028] Among them, the model building module is used to build a bowel preparation quality assessment model based on the bowel preparation quality assessment requirements; In this embodiment, a bowel preparation quality assessment model is constructed, including: According to the needs of AI-based bowel preparation quality assessment, historical excrement data is collected and divided into training sets and test sets; Based on deep learning technology, a training set is used to train the deep learning model, allowing the deep learning model to autonomously learn bowel preparation quality assessment behaviors from the training set. This is used to assess the quality of the patient's bowel preparation, determine the cleanliness of the patient's intestines, and determine the bowel preparation quality assessment model based on deep learning; Use the test set to test the deep learning-based bowel preparation quality assessment model, evaluate its performance, determine whether it can achieve the expected effect of assessing the quality of patients' bowel preparation and judging the cleanliness of the patients' intestines, and determine the model test evaluation results; According to the model test evaluation results, the parameters of the intestinal preparation quality assessment model based on deep learning are adjusted and optimized until the intestinal preparation quality assessment model based on deep learning can achieve the expected effect of evaluating the quality of the patient's intestinal preparation and judging the cleanliness level of the patient's intestine, and then determine the optimal intestinal preparation quality assessment model.

[0029] The quality assessment module is used to identify and analyze the fecal characteristic data according to the bowel preparation quality assessment model, conduct a quality assessment of the patient's bowel preparation, and determine the bowel preparation quality assessment result; In this example, the quality of the patient's bowel preparation is assessed, including: Deploy the optimal bowel preparation quality assessment model and deploy it in the actual bowel preparation quality assessment environment; The fecal characteristic data is input into the intestinal preparation quality assessment model, and the fecal characteristic data is characterized and analyzed according to the intestinal preparation quality assessment model. The quality of the patient's intestinal preparation is assessed and the cleanliness of the patient's intestine is judged to determine the intestinal preparation quality assessment result.

[0030] Among them, the output management module is used to visualize the output of the patient's intestinal preparation quality assessment report and conduct subsequent colonoscopy management for the patient.

[0031] In this embodiment, a patient's bowel preparation quality assessment report is visually output, and subsequent colonoscopy management is performed on the patient, including: Based on the intestinal preparation quality assessment results and combined with the patient's fecal fluorescent image, the patient's intestinal preparation quality assessment report is formed, and the patient's intestinal preparation quality assessment report is displayed to the doctor in real time in a visual form, so that the doctor can conduct subsequent colonoscopy management on the patient based on the patient's intestinal preparation quality assessment report.

[0032] In this embodiment, the cleanliness of the patient's intestine is divided into four levels based on the bowel preparation quality assessment results, including: Grade 1: The intestinal mucosa is not clearly visible and there is a lot of fecal residue in the intestine; Grade II, that is, the intestinal mucosa is clearly displayed, with a moderate amount of fecal residue in the intestine; Grade 3: The intestinal mucosa is clearly visible, with a small amount of fecal residue in the intestine; Grade 4: The intestinal mucosa is clearly visible and there is no fecal residue in the intestine; Among them, the higher the level, the better the intestinal preparation, and the more conducive it is to colonoscopy.

[0033] Specifically, the fecal characteristic data is identified and analyzed according to the bowel preparation quality assessment model, the quality of the patient's bowel preparation is assessed, and the cleanliness of the patient's intestine is determined to determine the bowel preparation quality assessment result. The bowel preparation quality assessment result is shown in Table 1: Therefore, the fecal characteristic data is characterized and analyzed according to the intestinal preparation quality assessment model, the patient's intestinal preparation quality is assessed and the cleanliness of the patient's intestine is judged, the intestinal preparation quality assessment result is determined, and the patient's fecal fluorescence image is combined to form the patient's intestinal preparation quality assessment report, and the patient's intestinal preparation quality assessment report is displayed to the doctor in real time in a visual form, so that the doctor can conduct subsequent colonoscopy management of the patient according to the patient's intestinal preparation quality assessment report, which can provide more objective and efficient intestinal preparation quality assessment results, thereby improving the accuracy and efficiency of colonoscopy.

[0034] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0035] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An AI-based bowel preparation quality assessment system, characterized in that: include: an acquisition and processing module configured to acquire a fluorescent image of a patient's excrement, perform preprocessing and feature extraction on the fluorescent image, and determine characteristic data of the excrement; a model building module configured to build a bowel preparation quality assessment model according to bowel preparation quality assessment requirements; a quality assessment module configured to perform feature recognition and analysis on fecal characteristic data according to a bowel preparation quality assessment model, perform a quality assessment on the patient's bowel preparation, and determine a bowel preparation quality assessment result; The output management module is configured to visually output the patient's bowel preparation quality assessment report and conduct subsequent colonoscopy management for the patient.

2. The AI-based bowel preparation quality assessment system according to claim 1, characterized in that: Acquire fluorescent images of the patient's feces, including: Collecting the patient's fecal sample in a sterile container, and diluting or centrifuging the patient's fecal sample to remove impurities in the fecal sample and concentrate the target component; Based on fluorescence imaging technology, a fluorescent dye is selected according to the characteristics of the target molecule in the fecal sample, and the fluorescent dye is mixed with the fecal sample to ensure full contact between the two; Select reaction conditions, including pH, temperature, and time, based on the target molecule in the fecal sample. Wash the fecal sample with a buffer to remove unbound dye. The labeled sample is purified by centrifugation or filtration, placed in a fluorescence microscope, and an excitation wavelength and filter combination are selected to obtain a fluorescent image of the patient's excrement.

3. The AI-based bowel preparation quality assessment system according to claim 1, characterized in that: Preprocessing of the patient's fecal fluorescence image includes: Denoising the patient's fecal fluorescence image based on a filter to remove noise and reduce background interference in the fecal fluorescence image; Performing contrast enhancement processing on the patient's fecal fluorescence image based on histogram equalization to enhance details in the fecal fluorescence image and highlight the fluorescence signal of the target area in the fecal fluorescence image; Image segmentation processing is performed on the patient's fecal fluorescence image based on the threshold segmentation method to separate the background and target areas in the fecal fluorescence image and eliminate nonspecific fluorescence interference.

4. The AI-based bowel preparation quality assessment system according to claim 3, characterized in that: Set the threshold for image segmentation, including: extracting the contrast value before and after contrast enhancement processing; Obtaining a contrast enhancement amplitude according to a contrast value before and after the contrast enhancement process; Performing a ratio processing on the contrast enhancement amplitude and the contrast value before the contrast enhancement processing to obtain a contrast enhancement ratio; comparing the contrast enhancement ratio with a preset contrast enhancement ratio threshold; When the contrast enhancement ratio does not exceed a preset contrast enhancement ratio threshold, a preset initial threshold parameter for image segmentation processing is retrieved; and image segmentation processing is performed on the patient's fecal fluorescence image using the preset initial threshold parameter for image segmentation processing based on a threshold segmentation method; When the contrast enhancement ratio exceeds a preset contrast enhancement ratio threshold, the initial threshold parameter is adjusted using the grayscale values of the fecal fluorescence images before and after contrast enhancement; Based on the threshold segmentation method, the patient's fecal fluorescence image is segmented using the adjusted threshold parameters.

5. The AI-based bowel preparation quality assessment system according to claim 4, characterized in that: When the contrast enhancement ratio exceeds a preset contrast enhancement ratio threshold, the initial threshold parameter is adjusted using the grayscale values of the fecal fluorescence images before and after contrast enhancement, including: When the contrast enhancement ratio exceeds a preset contrast enhancement ratio threshold, retrieving the fecal fluorescence image before contrast enhancement processing; performing grayscale processing on the feces fluorescence image before contrast enhancement processing to obtain a first grayscale image; Retrieve the contrast-enhanced fluorescence image of the excrement; performing grayscale processing on the excrement fluorescence image after contrast enhancement processing to obtain a second grayscale image; Extracting each pixel in the first grayscale image; Extracting each pixel in the second grayscale image; For a pixel point at the same position in the first grayscale image and the second grayscale image, obtaining a grayscale difference between the pixel point at the same position; performing ratio processing on the grayscale difference of the pixel at the same position and the grayscale value of the pixel at the same position in the first grayscale image to obtain the grayscale change rate of the pixel at the same position; The initial threshold parameters are adjusted using the grayscale change rate and contrast enhancement ratio.

6. The AI-based bowel preparation quality assessment system according to claim 3, characterized in that: Feature extraction is performed on the patient's fecal fluorescence image, including: Based on the convolutional neural network (CNN), feature extraction is performed on the patient's fecal fluorescence images. Semantic features valuable for bowel preparation quality assessment are extracted from the fecal fluorescence images. The extracted features are then subjected to dimensionality reduction processing based on principal component analysis to determine the fecal characteristic data, including fecal composition, concentration, color, contrast, and texture.

7. The AI-based bowel preparation quality assessment system according to claim 1, characterized in that: Construct a bowel preparation quality assessment model, including: According to the needs of AI-based bowel preparation quality assessment, historical excrement data is collected and divided into training sets and test sets; Based on deep learning technology, a training set is used to train the deep learning model, allowing the deep learning model to autonomously learn bowel preparation quality assessment behaviors from the training set. This is used to assess the quality of the patient's bowel preparation, determine the cleanliness of the patient's intestines, and determine the bowel preparation quality assessment model based on deep learning; Use the test set to test the deep learning-based bowel preparation quality assessment model, evaluate its performance, determine whether it can achieve the expected effect of assessing the quality of patients' bowel preparation and judging the cleanliness of the patients' intestines, and determine the model test evaluation results; According to the model test evaluation results, the parameters of the intestinal preparation quality assessment model based on deep learning are adjusted and optimized until the intestinal preparation quality assessment model based on deep learning can achieve the expected effect of evaluating the quality of the patient's intestinal preparation and judging the cleanliness level of the patient's intestine, and then determine the optimal intestinal preparation quality assessment model.

8. The AI-based bowel preparation quality assessment system according to claim 7, characterized in that: Assess the quality of the patient's bowel preparation, including: Deploy the optimal bowel preparation quality assessment model and deploy it in the actual bowel preparation quality assessment environment; The fecal characteristic data is input into the intestinal preparation quality assessment model, and the fecal characteristic data is characterized and analyzed according to the intestinal preparation quality assessment model. The quality of the patient's intestinal preparation is assessed and the cleanliness of the patient's intestine is judged to determine the intestinal preparation quality assessment result.

9. The AI-based bowel preparation quality assessment system according to claim 1, wherein: Visually output the patient's bowel preparation quality assessment report and conduct subsequent colonoscopy management for the patient, including: Based on the intestinal preparation quality assessment results and combined with the patient's fecal fluorescent image, the patient's intestinal preparation quality assessment report is formed, and the patient's intestinal preparation quality assessment report is displayed to the doctor in real time in a visual form, so that the doctor can conduct subsequent colonoscopy management on the patient based on the patient's intestinal preparation quality assessment report.

10. The AI-based bowel preparation quality assessment system according to claim 9, characterized in that: The degree of cleanliness in the patient's intestine is divided into four levels based on the bowel preparation quality assessment results.

Images