Treatment effect comparison method and system based on LED treatment instrument

Abstract

The invention relates to an image data processing technology, and provides a treatment effect comparison method and system based on an LED treatment instrument, and the method comprises the steps: obtaining image data of a target treatment region, carrying out the preprocessing, obtaining the preprocessed image data, carrying out the analysis of the preprocessed image data, obtaining a feature data set, and carrying out the comparison of the treatment effect of the LED treatment instrument. Inputting the feature data set into a preset recognition model, and outputting corresponding treatment parameters; controlling a light source of the LED therapeutic instrument according to the therapeutic parameters, and obtaining result data after illumination is finished; and inputting the image data and the result data into a comparison model based on AI big data to obtain a treatment effect evaluation report. Quantitative evaluation of the treatment effect is achieved through an automatic process, subjectivity of traditional manual judgment is avoided, analysis accuracy and operation convenience are improved, and the problems that in the prior art, manual result observation is affected by experience and operation, and operation is tedious when multiple devices are used for image collection and monitoring are solved.

Description

Technical Field

[0001] The present application relates to the technical field of image data processing, and in particular to a treatment effect comparison method and system based on an LED therapeutic device. Background Art

[0002] In recent years, with the rapid development of photomedicine, light-based medical devices have become increasingly widely used in disease treatment. LED (Light Emitting Diode) light sources, particularly due to their controllable wavelength, stable energy, and high safety, have become a research hotspot in the field of medical phototherapy. LED therapeutic devices are widely used to treat skin diseases, chronic pain, and viral infections, stimulating tissue repair or killing pathogens through specific wavelengths of light. However, in practical applications, how to evaluate treatment efficacy in real time and optimize treatment plans remains a key challenge.

[0003] In related technologies, treatment effectiveness is typically assessed by manually observing changes in the lesion area before and after treatment. For example, in laser therapy devices, external equipment is used to observe the treatment area during treatment. The treatment effect is determined by comparing changes in color and shape using image data collected by the naked eye or external imaging equipment.

[0004] Regarding the above technical solution, although a basic assessment of tissue changes before and after treatment can be achieved through manual observation combined with external equipment detection, the results of manual observation are affected by experience and operation, and the use of multiple devices for image acquisition and monitoring has the problem of cumbersome operation. Summary of the Invention

[0005] In order to improve the existing technology, in which manual observation results are affected by experience and operation, and multiple devices are used for image acquisition and monitoring, which results in cumbersome operations, the present application provides a treatment effect comparison method and system based on an LED therapeutic device.

[0006] The present invention provides a treatment effect comparison method based on an LED therapeutic device, the comparison method comprising the following steps: acquiring image data of a target treatment area, preprocessing the image data to obtain preprocessed image data; analyzing the preprocessed image data using an image feature extraction algorithm to obtain a feature data set, inputting the feature data set into a preset recognition model, and outputting corresponding treatment parameters; controlling the LED light source of the LED therapeutic device according to the treatment parameters to illuminate the target treatment area, and obtaining result data after the illumination is completed; inputting the image data and the result data into a comparison model based on AI big data to obtain a treatment effect evaluation report.

[0007] As a preferred solution, the steps of obtaining image data of the target treatment area, preprocessing the image data, and obtaining preprocessed image data include: obtaining image data of the target treatment area through the inspection lens of the LED therapeutic device, using an adaptive median filtering method to remove random noise in the image data to obtain denoised image data; using a region segmentation algorithm to segment the target area of the denoised image data, extracting the outline of the target treatment area and generating a segmentation mask, superimposing the segmentation mask with the denoised image data to obtain a refined image; converting the refined image from RGB space to HSV space according to a color space conversion algorithm, extracting the color characteristic matrix of the target area, and combining the color characteristic matrix with the texture characteristic matrix of the target treatment area to obtain preprocessed image data.

[0008] As a preferred solution, the steps of using an image feature extraction algorithm to analyze the preprocessed image data to obtain a feature data set, inputting the feature data set into a preset recognition model, and outputting corresponding treatment parameters include: applying a local binary pattern to extract texture features from the preprocessed image data to generate a texture feature matrix; using image morphological processing to perform morphological feature analysis on the image data to generate a morphological feature matrix of the target treatment area; calculating the contrast, entropy and energy characteristics of the target treatment area through a grayscale co-occurrence matrix to generate a grayscale characteristic matrix; performing multidimensional data fusion processing on the texture feature matrix, morphological feature matrix and grayscale characteristic matrix to obtain a feature data set; inputting the feature data set into a preset recognition model, and the recognition model generates treatment parameters including LED light wavelength, light intensity, spot area and illumination time based on specific training data.

[0009] As a preferred solution, the step of applying a local binary pattern to extract texture features from the preprocessed image data and generating a texture feature matrix includes: dividing the preprocessed image data into a number of image blocks of fixed size, calculating the grayscale value difference between the central pixel and the surrounding pixels of each image block, and generating a binary code based on the grayscale value difference using a local binary pattern; performing histogram statistics on the binary code of each image block to generate a local texture histogram of each image block, normalizing the local texture histograms of all the image blocks to obtain a preliminary texture histogram of the target treatment area; applying a fast Fourier transform to perform frequency domain analysis on the preliminary texture histogram to obtain frequency domain features, combining the preliminary texture histogram with the frequency domain features to generate a texture feature matrix of the target treatment area.

[0010] As a preferred solution, the step of using image morphological processing to perform morphological feature analysis on the image data to generate a morphological feature matrix of the target treatment area includes: using a Canny edge detection algorithm to perform edge detection on the image data to generate an initial edge map of the target treatment area, applying morphology to perform an expansion operation on the initial edge map to obtain an expanded edge map, and denoising the expanded edge map through a morphological opening operation to obtain a denoised edge map; performing region segmentation on the denoised edge map through a watershed segmentation algorithm to extract the segmentation contour of the target area, and calculating the geometric characteristic parameters of the segmentation contour to generate a morphological feature matrix of the target treatment area.

[0011] As a preferred solution, the steps of controlling the LED light source of the LED therapeutic device according to the treatment parameters to illuminate the target treatment area and obtaining result data after the illumination is completed include: adjusting the light wavelength, light power density, and spot area of the LED light source of the LED therapeutic device according to the treatment parameters; using the inspection lens of the LED therapeutic device to obtain a static image after the illumination is completed, and generating result data after the illumination is completed based on the static image.

[0012] As a preferred embodiment, the step of inputting the image data and the result data into a comparison model based on AI big data to obtain a treatment effect evaluation report includes: inputting the image data and the result data into a comparison model based on AI big data, in the comparison model, using a color difference formula to calculate the hue change, brightness change and saturation change of the image data and the result data to obtain color characteristic change parameters; using an edge detection algorithm and a watershed segmentation algorithm to perform matching analysis on the morphological feature matrices of the image data and the result data to generate morphological change parameters of the target area; calculating the texture feature difference between the image data and the result data through a gray level co-occurrence matrix algorithm to generate texture change parameters of the target area; performing multi-dimensional feature fusion on the color characteristic change parameters, the morphological change parameters and the texture change parameters to generate a comprehensive comparison data set of treatment effects, and generating a treatment effect evaluation report based on the comprehensive comparison data set.

[0013] The present application also provides a treatment effect comparison system based on an LED therapeutic device, including: an acquisition unit, used to acquire image data of a target treatment area, preprocess the image data, and obtain preprocessed image data; an analysis unit, used to analyze the preprocessed image data using an image feature extraction algorithm, obtain a feature data set, input the feature data set into a preset recognition model, and output corresponding treatment parameters; a control unit, used to control the LED light source of the LED therapeutic device according to the treatment parameters to illuminate the target treatment area and obtain result data after the illumination is completed; a comparison unit, used to input the image data and the result data into a comparison model based on AI big data to obtain a treatment effect evaluation report.

[0014] Compared with the existing technology, the present application has the following beneficial effects: accurate judgment and simple operation. By obtaining image data of the target treatment area and preprocessing it, combining the image feature extraction algorithm and the recognition model to generate treatment parameters, the target treatment area is illuminated by the LED therapeutic device according to the treatment parameters, and the result data after illumination is obtained at the same time, and the image data and result data are input into the comparison model based on AI big data for analysis. The quantitative evaluation of the treatment effect is achieved through the automated process, avoiding the subjectivity of traditional manual judgment, improving the accuracy of analysis and the simplicity of operation, providing a reliable basis for the clinical application of LED therapeutic devices, and improving the existing technology. The results of manual observation are affected by experience and operation, and multiple devices are used for image acquisition and monitoring, which has cumbersome operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] The structures, proportions, sizes, etc. depicted in the drawings of this specification are only used to match the contents disclosed in the specification so as to facilitate understanding and reading by persons familiar with this technology. They are not intended to limit the conditions under which the present invention can be implemented and therefore have no substantive technical significance. Any structural modifications, changes in proportional relationships, or adjustments in size should still fall within the scope of the technical contents disclosed in the present invention without affecting the effects and objectives that can be achieved by the present invention.

[0017] Figure 1 1 is a flow chart of a method for comparing treatment effects based on an LED therapeutic apparatus provided by an embodiment of the present invention; Figure 2It is a schematic block diagram of the structure of a treatment effect comparison system based on an LED therapeutic apparatus provided in an embodiment of the present invention.

[0018] Description of reference numerals: 10. Treatment effect comparison system based on LED therapeutic device; 11. Acquisition unit; 12. Analysis unit; 13. Control unit; 14. Comparison unit. DETAILED DESCRIPTION

[0019] 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 them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0020] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, combined, or partially merged, so the actual execution order may vary depending on the actual situation.

[0021] It should also be understood that the terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit the present application. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0022] It should be further understood that the term "and / or" used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0023] The technical solution of the present invention will be further described below with reference to the accompanying drawings and through specific implementation methods. Example 1:

[0024] like Figure 1 As shown, the present application provides a treatment effect comparison method based on an LED therapeutic device, including steps S100 to S400.

[0025] Step S100: Acquire image data of the target treatment area, and preprocess the image data to obtain preprocessed image data.

[0026] In this step, image data of the target treatment area is collected through the inspection lens of the LED therapeutic device; specifically, the image data is subjected to noise reduction processing, the target area is extracted, and pre-processed image data is formed through a combination of color space conversion and feature matrix.

[0027] For example, when treating cervical lesions, image data of the lesion area is obtained, and after noise reduction and segmentation, the image of the lesion area is extracted and converted to a specific color space for subsequent analysis.

[0028] Step S200: Analyze the pre-processed image data using an image feature extraction algorithm to obtain a feature data set, input the feature data set into a preset recognition model, and output corresponding treatment parameters.

[0029] In this step, the preprocessed image data is analyzed using an image feature extraction algorithm to extract a feature data set including texture, morphology, and color characteristics. Specifically, the feature data set is input into a deep learning-based recognition model to generate treatment parameters for subsequent lighting operations.

[0030] For example, when analyzing image data of the lesion area, the surface structural characteristics of the tissue are obtained through the texture feature extraction algorithm, and the color characteristics are combined to generate personalized treatment parameters, such as light wavelength range, light power density, etc.

[0031] Step S300: Control the LED light source of the LED therapeutic device according to the treatment parameters to illuminate the target treatment area and obtain result data after the illumination is completed.

[0032] In this step, the wavelength, light power density and spot size of the LED light source are adjusted according to the treatment parameters generated by the recognition model; specifically, the inspection lens is used to record the target treatment area data after the illumination is completed to generate result data.

[0033] For example, the wavelength of the LED light source is adjusted to 405nm to 420nm according to the treatment parameters output by the recognition model, and the spot size is set to a diameter of 10mm. At the same time, a static image after illumination is collected as the result data.

[0034] Step S400: Input the image data and result data into the comparison model based on AI big data to obtain a treatment effect evaluation report.

[0035] In this step, the image data and result data are input into the comparison model, and a comprehensive comparison data set is generated through multidimensional fusion analysis of color, texture and morphological features. Specifically, a treatment effect evaluation report is generated based on the comparison results to intuitively display the changes before and after treatment.

[0036] For example, the color difference algorithm and texture analysis method are used to compare the image data before and after treatment, and a report is generated, which includes color characteristic change parameters, morphological change parameters and texture change parameters.

[0037] In this embodiment, image data of the target treatment area is obtained and preprocessed to obtain preprocessed image data; then, the preprocessed image data is analyzed using an image feature extraction algorithm to obtain a feature data set, and the feature data set is input into a preset recognition model to output corresponding treatment parameters; then, the LED light source of the LED therapy device is controlled according to the treatment parameters to illuminate the target treatment area and obtain result data after the illumination is completed; finally, the image data and the result data are input into a comparison model based on AI big data to obtain a treatment effect evaluation report. By preprocessing the image data of the target treatment area, extracting features, and analyzing the comparison model based on AI big data, a comprehensive comparison of multi-dimensional data before and after treatment is achieved, which can quantify the treatment effect and present it intuitively in the form of a report. Compared with the existing technology, the method of the present application has higher evaluation accuracy and automation level, eliminates the uncertainty of subjective judgment, simplifies the operation process, and provides strong support for the optimization of treatment plans and real-time monitoring of treatment efficacy. It improves the existing technology, where the results of manual observation are affected by experience and operation, and multiple devices are used for image acquisition and monitoring, which has cumbersome operation. Example 2:

[0038] In step S100, image data of the target treatment area is obtained through the inspection lens of the LED therapeutic device, and the random noise in the image data is removed by an adaptive median filtering method to obtain noise-reduced image data.

[0039] The LED therapy device utilizes the high-resolution inspection lens built into the device to capture multiple frames of images of the target treatment area, and combines this with an adaptive median filtering algorithm to remove random noise generated during the acquisition process. Specifically, the adaptive median filter dynamically adjusts the filter window size to enhance high-density noise areas while retaining the edge details and texture characteristics of the target treatment area, thereby improving image clarity and usability.

[0040] For example, in the image data acquisition of the cervical lesion area, a 3×3 filter window is used to replace the noise pixels. When high noise density is detected, the window is automatically expanded to 5×5 to enhance the noise reduction effect while retaining the boundaries and details of the lesion area.

[0041] The target area is segmented by the region segmentation algorithm on the denoised image data, the outline of the target treatment area is extracted and a segmentation mask is generated, and the segmentation mask is superimposed on the denoised image data to obtain a refined image.

[0042] The denoised image data is processed using a threshold-based region segmentation algorithm to automatically extract the boundary contours of the target treatment area. Specifically, the Otsu adaptive threshold segmentation method is used to determine the optimal segmentation points of the image, generate a binary mask of the target area, and superimpose this mask on the denoised image to remove irrelevant areas and enhance image details in the target area.

[0043] For example, when extracting the contour of the cervical lesion area, the Otsu algorithm is used to calculate the optimal threshold, successfully separate the lesion area and the background, and generate a segmentation mask. After superimposing it with the denoised image, a refined image with background removed and clear details is obtained.

[0044] The refined image is converted from RGB space to HSV space according to the color space conversion algorithm, the color characteristic matrix of the target area is extracted, and the color characteristic matrix is combined with the texture characteristic matrix of the target treatment area to obtain the preprocessed image data.

[0045] The refined image is converted from RGB color space to HSV color space through a color space conversion algorithm, and the hue (H), saturation (S), and brightness (V) components of the target area are extracted. Specifically, the hue histogram and saturation average of the pixels in the lesion area are calculated to form a color feature matrix, which is then combined with the texture feature matrix of the area for subsequent analysis.

[0046] For example, when analyzing cervical lesion images, after converting the RGB color space to the HSV color space, the hue variation range is extracted (such as 0° to 60° represents the lesion characteristics of red tones), and the saturation variation range is recorded at the same time (such as high saturation areas are signs of inflammation). Finally, the color feature matrix and the texture feature matrix are combined to obtain the preprocessed image data.

[0047] In step S200, local binary patterns are applied to extract texture features from the pre-processed image data to generate a texture feature matrix.

[0048] The preprocessed image data is divided into several fixed-size image blocks, and the grayscale value difference between the central pixel and the surrounding pixels of each image block is calculated; specifically, a local binary pattern code (LBP code) is generated based on the grayscale value difference, and a local texture histogram is generated by counting the frequency of LBP codes. All local texture histograms are normalized to obtain the texture feature matrix of the target treatment area.

[0049] For example, when extracting the texture features of diseased tissue, each image block is set to 5×5 in size, the grayscale difference between the central pixel and the surrounding 8 pixels is calculated, the LBP code in the form of "10101100" is generated, and the LBP code frequency in all image blocks is counted to generate a texture histogram.

[0050] Image morphological processing is used to analyze the morphological characteristics of image data and generate a morphological feature matrix of the target treatment area.

[0051] The boundary and shape characteristics of the target treatment area are analyzed by using morphological operations. Specifically, the Canny edge detection algorithm is used to generate an initial edge map, and the initial edge map is dilated to connect discontinuous boundaries. Subsequently, the morphological opening operation is applied to remove small noise areas and isolated points. Finally, the geometric characteristics of the region (such as perimeter, area, aspect ratio, etc.) are calculated to generate a morphological feature matrix of the target treatment area.

[0052] For example, when analyzing the morphological characteristics of the cervical lesion area, after extracting the edge of the lesion area using the Canny algorithm, the broken boundaries are connected using a dilation operation, and then isolated noise points are removed through an opening operation, and the perimeter and area ratio characteristics of the lesion area are calculated.

[0053] The contrast, entropy and energy characteristics of the target treatment area are calculated through the gray-level co-occurrence matrix to generate a gray-level characteristic matrix.

[0054] The statistical characteristics of the image are extracted by calculating the gray level co-occurrence matrix of the target treatment area; specifically, the gray level co-occurrence matrix calculates the contrast, entropy and energy characteristics based on the relative distribution of pixel gray values, and combines these parameters to form a gray level characteristic matrix.

[0055] For example, when analyzing cervical lesion tissue, the contrast (such as high contrast indicates clear boundaries), entropy (such as high entropy indicates high texture complexity) and energy (such as high energy indicates uniform regional grayscale distribution) of the grayscale co-occurrence matrix are calculated to generate a grayscale characteristic matrix.

[0056] The texture feature matrix, morphological feature matrix and grayscale feature matrix are subjected to multi-dimensional data fusion processing to obtain a feature data set.

[0057] By combining different feature matrices into a unified data set; specifically, a weighted fusion method is used to normalize the parameters of the texture feature matrix, morphological feature matrix and grayscale feature matrix, and combine them into a complete feature data set.

[0058] For example, when fusing feature data of a cervical lesion area, the contrast of texture features, the area ratio of morphological features, and the energy parameters of grayscale characteristics are normalized and weightedly added to generate a feature data set for identification.

[0059] The feature data set is input into a preset recognition model, which generates treatment parameters including LED light wavelength, light intensity, spot area and illumination time based on specific training data.

[0060] The fused feature data set is input into a deep learning model; specifically, the recognition model analyzes the feature data set through a neural network to generate treatment parameters for controlling the light of the LED therapeutic device.

[0061] For example, based on the characteristics of different lesion areas, the recognition model generates treatment parameters with a wavelength of 405nm~420nm, a light intensity of 50mW / cm², and a spot area of 10mm.

[0062] Among them, the steps of applying local binary patterns to extract texture features from preprocessed image data and generating a texture feature matrix include: dividing the preprocessed image data into several image blocks of fixed sizes, calculating the grayscale value difference between the central pixel and the surrounding pixels of each image block, and generating a binary code based on the grayscale value difference using the local binary pattern.

[0063] The preprocessed image data is divided into several image blocks of fixed size, and the grayscale difference between the central pixel and the surrounding 8 pixels is calculated based on the central pixel of each image block. Specifically, the grayscale difference is calculated using the differential formula: grayscale difference = central pixel grayscale value - surrounding pixel grayscale value, and the positive difference is recorded as 1 and the negative difference is recorded as 0, thereby generating an 8-bit binary code (i.e., local binary pattern coding, LBP coding).

[0064] For example, in the image processing of the cervical lesion area, the grayscale value of the central pixel of the 5×5 pixel image block is set to 120, and the grayscale values of the surrounding pixels are 115, 122, 118, 110, etc. The grayscale difference (such as +5, -2, +2, +10, etc.) is calculated and converted into binary code "10110100".

[0065] The binary code of each image block is subjected to histogram statistics to generate the local texture histogram of each image block. The local texture histograms of all image blocks are normalized to obtain the preliminary texture histogram of the target treatment area.

[0066] A local texture histogram is generated by counting the frequency of all LBP codes in each image block. Specifically, the frequency distribution of 256 possible LBP codes is calculated for each image block to generate a histogram representation. All local texture histograms are normalized, and the influence of different block sizes is eliminated by dividing each histogram value by the total frequency to obtain a preliminary texture histogram of the target treatment area.

[0067] For example, when analyzing the cervical lesion area, the LBP code in an image block may be "00000000" appearing 50 times, "11111111" appearing 30 times, and "10101010" appearing 20 times. After normalizing these coding frequencies, the generated histogram values are 0.5, 0.3, and 0.2, respectively, representing the texture characteristics of the image block.

[0068] Fast Fourier transform is applied to perform frequency domain analysis on the preliminary texture histogram to obtain frequency domain features. The preliminary texture histogram and frequency domain features are combined to generate the texture feature matrix of the target treatment area.

[0069] By applying fast Fourier transform (FFT) to the preliminary texture histogram, the texture histogram is converted from the spatial domain to the frequency domain. Specifically, the frequency domain amplitude and phase characteristics of the texture histogram are calculated to extract the high-frequency and low-frequency component features of the texture. The frequency domain features are spliced with the preliminary texture histogram in the data dimension to generate a texture feature matrix of the target treatment area.

[0070] For example, when analyzing the texture characteristics of the cervical lesion area, the histogram data was converted by FFT, and it was found that the amplitude of the high-frequency component of the histogram was higher, indicating that the surface texture of the lesion tissue was relatively rough; by combining the frequency domain amplitude characteristics with the original histogram, a texture feature matrix was generated for subsequent analysis.

[0071] Among them, the steps of using image morphological processing to analyze the morphological features of image data and generate a morphological feature matrix of the target treatment area include: using the Canny edge detection algorithm to perform edge detection on the image data to generate an initial edge map of the target treatment area, applying morphology to perform an expansion operation on the initial edge map to obtain an expanded edge map, and denoising the expanded edge map through a morphological opening operation to obtain a denoised edge map.

[0072] The boundary characteristics of the target treatment area are extracted by utilizing the Canny edge detection algorithm. Specifically, by calculating the image pixel gradient, the boundaries of the high-gradient area are detected and an initial edge map is generated. A morphological dilation operation is performed on the initial edge map to connect discontinuous boundaries, and an opening operation is applied to the dilated edge map to remove isolated noise points, thereby obtaining a complete denoised edge map.

[0073] For example, in an image of the cervical lesion area, after extracting the lesion edge using the Canny algorithm, the dilation operation connects the discontinuous boundaries (such as the boundaries interrupted by noise), and finally the morphological opening operation is used to remove isolated edge points smaller than 3 pixels to generate a clear edge map.

[0074] The denoised edge image is segmented using the watershed segmentation algorithm to extract the segmentation contour of the target area, and the geometric characteristic parameters of the segmentation contour are calculated to generate the morphological feature matrix of the target treatment area.

[0075] Region segmentation is performed by applying the watershed segmentation algorithm to the denoised edge map. Specifically, segmentation seed points are generated based on the edge gradient map, and the seed point area is expanded to cover the target area using the watershed algorithm. The contour of the segmented target area is extracted, and the geometric characteristic parameters, including perimeter, area, aspect ratio, boundary complexity, etc., are calculated to generate a morphological feature matrix of the target treatment area.

[0076] For example, when analyzing the cervical lesion area, the watershed segmentation algorithm is used to separate the lesion area from the background. The calculated area is 50 mm², the perimeter is 30 mm, the aspect ratio is 1.5, and the boundary complexity is 1.2, and finally a complete morphological feature matrix is generated.

[0077] In step S300, the light wavelength, light power density, and spot area of the LED light source of the LED therapeutic device are adjusted according to the treatment parameters.

[0078] The LED light source's parameters are controlled through a driver circuit and optical components. Specifically, the LED light source's wavelength range, optical power density, and spot size are dynamically adjusted based on the treatment parameters generated by the recognition model. The wavelength is achieved by driving the specific wavelength chip in the LED array, while the optical power density is precisely adjusted using the duty cycle of the pulse-width modulation (PWM) control signal. The spot size is adjusted through the optical lens assembly to match the target treatment area and treatment needs.

[0079] For example, when treating cervical lesions, the LED light wavelength is adjusted to 405nm-420nm based on treatment parameters to match the light absorption characteristics of the lesion tissue. The light power density is set to 50mW / cm² to avoid excessive heating of the tissue. At the same time, the light spot diameter is controlled to 10mm through an adjustable optical lens to accurately cover the lesion area.

[0080] The inspection lens of the LED therapeutic device is used to obtain a static image after the illumination is completed, and the result data after the illumination is completed is generated based on the static image.

[0081] The illumination area is monitored in real time through the inspection lens and a static image of the target treatment area is captured at the end of the illumination. Specifically, the static image records the color characteristics, texture characteristics and boundary contour changes of the target area, and is combined with the real-time recorded treatment process parameters (such as temperature distribution and light intensity curve) to generate the result data after the illumination is completed.

[0082] For example, during the treatment of cervical lesion areas, the inspection lens captures a static image of the lesion area after the illumination ends, analyzes the saturation changes and boundary smoothness of the red area in the image, and combines the temperature change curve recorded during the treatment process to generate result data for subsequent comparison and analysis.

[0083] In step S400, the image data and result data are input into a comparison model based on AI big data. In the comparison model, the color difference formula is used to calculate the hue change, brightness change, and saturation change of the image data and result data to obtain the color characteristic change parameters.

[0084] By inputting image data and result data into the comparison model, the CIEDE2000 color difference formula is used to accurately calculate the color characteristic changes of the target area; specifically, the hue difference (such as the change in hue angle), brightness difference (such as the average change in grayscale value) and saturation difference (such as the distribution change of the saturation histogram) of the area before and after treatment are calculated to form the color characteristic change parameters.

[0085] For example, when comparing the treatment effects of the lesion area, the CIEDE2000 formula is used to calculate the change in red hue (such as from 15° to 5°), brightness change (grayscale value increased by 20 units), and saturation decrease in the image before and after treatment to obtain the color characteristic change parameters.

[0086] The edge detection algorithm and watershed segmentation algorithm are used to match and analyze the morphological feature matrix of the image data and the result data to generate the morphological change parameters of the target area.

[0087] The boundaries of the target area in the images before and after treatment are extracted using an edge detection algorithm, and the edge map is segmented using a watershed segmentation algorithm to calculate geometric characteristics; specifically, the changes in the perimeter, area, and aspect ratio of the area before and after treatment are analyzed to generate the morphological change parameters of the target area.

[0088] For example, when analyzing the morphological changes of the lesion area, the Canny edge detection algorithm is used to extract the regional boundaries before and after treatment, and the watershed segmentation algorithm is combined to calculate the change in regional area (such as from 50mm² to 30mm²) and generate the morphological change parameters of the target area.

[0089] The texture feature differences between the image data and the result data are calculated using the gray level co-occurrence matrix algorithm to generate the texture change parameters of the target area.

[0090] Texture change information is extracted by calculating the gray-level co-occurrence matrix of the images before and after treatment. Specifically, the contrast, entropy and energy changes of the images before and after treatment are analyzed to generate texture change parameters of the target area.

[0091] For example, by analyzing the texture characteristics of images before and after treatment through the gray-level co-occurrence matrix, it was found that the contrast decreased by 10%, the entropy increased by 15%, and the energy increased by 20% after treatment, indicating that the texture of the lesion area tended to be smooth, and the corresponding texture change parameters were generated.

[0092] The color characteristic change parameters, morphological change parameters and texture change parameters are subjected to multi-dimensional feature fusion to generate a comprehensive comparison data set of treatment effects, and a treatment effect evaluation report is generated based on the comprehensive comparison data set.

[0093] A multidimensional feature fusion method combines color, morphology, and texture change parameters into a comprehensive comparison dataset. Specifically, a weighted feature fusion model is used to normalize the parameters according to their importance. The fusion results are then fed into an AI model to generate a treatment effect evaluation report. The report includes a detailed comparison of pre- and post-treatment images, a quantitative analysis of changes in each feature, and visual trend charts.

[0094] For example, in the analysis of treatment effects, color characteristic change parameters (such as changes in hue and brightness), morphological change parameters (such as area reduction rate and perimeter change rate), and texture change parameters (such as contrast and entropy value changes) are weighted and fused, and a report containing data tables, change curves, and treatment effect conclusions is generated to provide clinicians with intuitive treatment effect evaluation results.

[0095] In this embodiment, high-quality image data preprocessing is achieved by acquiring and preprocessing image data of the target treatment area. This includes using an adaptive median filter to remove random noise, using a region segmentation algorithm to extract the target treatment area's contours and generate a refined image, and using a color space conversion algorithm to extract color and texture feature matrices. This ensures the accuracy of subsequent analysis. Subsequently, texture features from the preprocessed image data are extracted using local binary patterns to generate a texture feature matrix. Image morphology is combined with image morphology to analyze the boundaries and morphological characteristics of the target treatment area to generate a morphological feature matrix. Furthermore, texture statistical properties such as contrast, entropy, and energy are calculated using a gray-level co-occurrence matrix to generate a gray-level feature matrix. These feature matrices are then fused through multidimensional data to generate a feature dataset, which is then fed into a recognition model. Based on the training data, the recognition model generates treatment parameters, including LED light wavelength, light intensity, spot area, and illumination duration, ensuring targeted and precise treatment. Subsequently, the light source settings of the LED therapy device are controlled based on the treatment parameters generated by the recognition model to complete the illumination treatment of the target treatment area. A static image after the illumination treatment is captured through an inspection lens to generate result data. Finally, the image data before treatment and the result data after illumination are input into the comparison model based on AI big data. Through color difference calculation, morphological change analysis, and texture characteristic change analysis, a comprehensive comparison data set is generated, and a treatment effect evaluation report is generated based on the comparison data set.

[0096] This embodiment integrates image processing, feature extraction, AI recognition and big data comparison and analysis technologies to achieve a closed loop of the entire process from treatment parameter generation to effect evaluation, thereby improving the accuracy, real-time nature and intelligence level of treatment effect evaluation, and providing strong support for the practical application of LED therapeutic devices. Example 3:

[0097] like Figure 2 As shown, the present application provides a treatment effect comparison system based on an LED therapeutic device, including an acquisition unit, an analysis unit, a control unit and a comparison unit.

[0098] The acquisition unit is mainly used to acquire image data of the target treatment area, preprocess the image data, and obtain preprocessed image data.

[0099] The analysis unit is mainly used to analyze the preprocessed image data using an image feature extraction algorithm to obtain a feature data set, input the feature data set into a preset recognition model, and output corresponding treatment parameters.

[0100] The control unit is mainly used to control the LED light source of the LED therapy device according to the treatment parameters to illuminate the target treatment area and obtain the result data after the illumination is completed.

[0101] The comparison unit is mainly used to input image data and result data into the comparison model based on AI big data to obtain a treatment effect evaluation report.

[0102] In this embodiment, a complete treatment effect comparison system is constructed by providing an acquisition unit, an analysis unit, a control unit, and a comparison unit that work together. The acquisition unit acquires and preprocesses image data of the target treatment area, including using denoising, segmentation, and feature extraction techniques to ensure clear and accurate preprocessed image data, laying the foundation for system analysis. The analysis unit applies an image feature extraction algorithm to perform multidimensional feature analysis on the preprocessed image data, including texture, morphology, and grayscale characteristics, and generates a feature dataset. Simultaneously, the feature dataset is processed using a preset recognition model to generate personalized treatment parameters, such as LED light wavelength, light intensity, spot area, and illumination duration, to achieve targeted treatment. The control unit analyzes the treatment parameters generated by the recognition model to precisely control the light source parameters of the LED therapy device, achieving personalized illumination of the target treatment area and collecting post-treatment result data in real time. The comparison unit inputs pre-treatment image data and post-treatment result data into an AI-based big data comparison model. Using multidimensional feature fusion analysis techniques such as color, morphology, and texture, it generates a comprehensive comparison dataset and ultimately outputs a detailed treatment effect evaluation report. The system in this embodiment ensures the accuracy and real-time nature of treatment effect comparison by integrating full-process functional modules of data acquisition, feature extraction, model recognition, precise control, and intelligent comparison, providing comprehensive and efficient technical support for the application of LED therapeutic devices, and can be widely used in the field of medical phototherapy.

[0103] It should be noted that technical personnel in the relevant technical field can clearly understand that for the convenience and conciseness of description, the specific working process of the system and each unit described above can refer to the corresponding process in the aforementioned embodiment of the treatment effect comparison method based on the LED therapeutic device, and will not be repeated here.

[0104] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for comparing treatment effects based on LED therapeutic devices, characterized in that: The comparison method comprises the following steps: Acquiring image data of a target treatment area, and preprocessing the image data to obtain preprocessed image data; Analyzing the preprocessed image data using an image feature extraction algorithm to obtain a feature data set, inputting the feature data set into a preset recognition model, and outputting corresponding treatment parameters; Controlling the LED light source of the LED therapeutic device according to the treatment parameters to illuminate the target treatment area and obtaining result data after the illumination is completed; The image data and the result data are input into a comparison model based on AI big data to obtain a treatment effect evaluation report.

2. The method for comparing the therapeutic effects of LED therapeutic devices according to claim 1, characterized in that: The step of acquiring image data of the target treatment area and preprocessing the image data to obtain preprocessed image data includes: Acquire image data of the target treatment area through the inspection lens of the LED therapeutic device, and use an adaptive median filtering method to remove random noise in the image data to obtain noise-reduced image data; Using a region segmentation algorithm to segment the target region of the noise-reduced image data, extracting the contour of the target treatment region and generating a segmentation mask, and superimposing the segmentation mask with the noise-reduced image data to obtain a refined image; The refined image is converted from RGB space to HSV space according to a color space conversion algorithm, the color characteristic matrix of the target area is extracted, and the color characteristic matrix is combined with the texture characteristic matrix of the target treatment area to obtain preprocessed image data.

3. The method for comparing treatment effects based on LED therapeutic apparatus according to claim 1, characterized in that: The step of analyzing the pre-processed image data using an image feature extraction algorithm to obtain a feature data set, inputting the feature data set into a preset recognition model, and outputting corresponding treatment parameters includes: Applying local binary patterns to extract texture features from the preprocessed image data to generate a texture feature matrix; Performing morphological feature analysis on the image data using image morphological processing to generate a morphological feature matrix of the target treatment area; The contrast, entropy and energy characteristics of the target treatment area are calculated through the gray-level co-occurrence matrix to generate a gray-level characteristic matrix; Performing multi-dimensional data fusion processing on the texture feature matrix, the morphological feature matrix and the grayscale feature matrix to obtain a feature data set; The feature data set is input into a preset recognition model, and the recognition model generates treatment parameters including LED light wavelength, light intensity, spot area and illumination time based on specific training data.

4. The method for comparing treatment effects based on LED therapeutic apparatus according to claim 3, characterized in that: The step of applying the local binary pattern to extract texture features from the preprocessed image data to generate a texture feature matrix includes: Dividing the preprocessed image data into a number of image blocks of fixed size, calculating the grayscale value difference between a central pixel and surrounding pixels of each image block, and generating a binary code based on the grayscale value difference using a local binary pattern; Performing histogram statistics on the binary code of each image block to generate a local texture histogram of each image block, and normalizing the local texture histograms of all the image blocks to obtain a preliminary texture histogram of the target treatment area; The preliminary texture histogram is subjected to frequency domain analysis using fast Fourier transform to obtain frequency domain features, and the preliminary texture histogram is combined with the frequency domain features to generate a texture feature matrix of the target treatment area.

5. The method for comparing treatment effects based on LED therapeutic apparatus according to claim 3, characterized in that: The step of performing morphological feature analysis on the image data using image morphological processing to generate a morphological feature matrix of the target treatment area includes: Performing edge detection on the image data using a Canny edge detection algorithm to generate an initial edge map of the target treatment area, performing a dilation operation on the initial edge map using morphology to obtain a dilated edge map, and performing denoising on the dilated edge map using a morphological opening operation to obtain a denoised edge map; The denoised edge image is segmented using a watershed segmentation algorithm to extract the segmentation contour of the target area, and the geometric characteristic parameters of the segmentation contour are calculated to generate a morphological feature matrix of the target treatment area.

6. The method for comparing treatment effects based on LED therapeutic apparatus according to claim 1, characterized in that: The step of controlling the LED light source of the LED therapeutic apparatus according to the treatment parameters to illuminate the target treatment area and obtaining result data after the illumination is completed includes: Adjust the light wavelength, light power density, and spot area of the LED light source of the LED therapeutic device according to the treatment parameters; A static image after the illumination is completed is acquired by utilizing the inspection lens of the LED therapeutic apparatus, and result data after the illumination is completed is generated based on the static image.

7. The method for comparing treatment effects based on LED therapeutic apparatus according to claim 1, characterized in that: The step of inputting the image data and the result data into a comparison model based on AI big data to obtain a treatment effect evaluation report includes: Inputting the image data and the result data into a comparison model based on AI big data, wherein the color difference formula is used in the comparison model to calculate the hue change, brightness change, and saturation change of the image data and the result data to obtain color characteristic change parameters; Performing matching analysis on the morphological feature matrix of the image data and the result data using an edge detection algorithm and a watershed segmentation algorithm to generate morphological change parameters of the target area; Calculating the texture feature difference between the image data and the result data by using a gray level co-occurrence matrix algorithm to generate a texture change parameter of the target area; The color characteristic change parameters, the morphological change parameters and the texture change parameters are subjected to multi-dimensional feature fusion to generate a comprehensive comparison data set of treatment effects, and a treatment effect evaluation report is generated based on the comprehensive comparison data set.

8. A treatment effect comparison system based on LED therapeutic device, characterized in that: include: an acquisition unit, configured to acquire image data of a target treatment area and preprocess the image data to obtain preprocessed image data; an analyzing unit, configured to analyze the preprocessed image data using an image feature extraction algorithm to obtain a feature data set, input the feature data set into a preset recognition model, and output corresponding treatment parameters; A control unit, configured to control the LED light source of the LED therapeutic apparatus according to the treatment parameters to illuminate the target treatment area and obtain result data after the illumination is completed; A comparison unit is used to input the image data and the result data into a comparison model based on AI big data to obtain a treatment effect evaluation report.

Images