Tumor Image Boundary Recognition Method and System Integrating Fluorescent Boundaries

Abstract

The present invention relates to the technical fields of medical image processing and tumor treatment, and provides a method and system for identifying tumor image boundaries by fusing fluorescence boundaries. The method includes collecting a sequence of tumor dynamic fluorescence images and respiratory waveform data through a fluorescence imaging device; performing multi-scale noise suppression processing on the image sequence to generate a preprocessed image with artifacts removed; extracting fluorescence intensity distribution features through a multi-layer feature weighted fusion network and generating an intensity distribution matrix; constructing a motion compensation model based on the respiratory waveform data to correct the boundary displacement of the intensity distribution matrix; fusing the corrected fluorescence features and tumor tissue texture features to generate an initial tumor boundary contour; performing iterative topological optimization on the initial boundary to eliminate boundary discontinuities; and mapping the optimized boundary to the space coordinate system of a laser treatment device to generate a boundary mapping model and update it in real time. The present invention can improve the accuracy and safety of laser treatment and reduce the damage to normal tissues.

Description

Technical Field

[0001] The present invention relates to the technical fields of medical image processing and tumor treatment. More specifically, the present invention relates to a method and system for identifying tumor image boundaries by fusing fluorescence boundaries. Background Art

[0002] In the fields of medical imaging technology and tumor treatment, fluorescence imaging technology has been widely used in the identification and localization of tumor boundaries due to its high sensitivity and specificity. Fluorescence imaging can label tumor tissues with fluorescent probes and can reflect the distribution of fluorescence signals in the tumor area in real time, providing a visual basis for surgery or laser treatment. However, fluorescence imaging faces many challenges in practical applications. First, the fluorescence signals in the tumor area are easily affected by respiratory motion, resulting in displacement and blurring of the boundary information in the image, which affects the accuracy of diagnosis and treatment. Second, noise and artifacts, such as high-frequency noise and fluorescence leakage, often accompany the fluorescence imaging process, further interfering with the clear identification of tumor boundaries. In addition, existing technologies lack an effective multi-dimensional feature weight allocation mechanism when fusing fluorescence signals with the structural features of tumor tissues, making it difficult to achieve accurate boundary extraction.

[0003] In the process of implementing the embodiments of the present invention, the inventors found that there are at least the following problems or defects in the prior art: it is impossible to effectively compensate for the influence of respiratory motion on the fluorescence image boundary, and it is difficult to maintain the stability of boundary recognition under dynamic conditions; the ability to suppress noise and eliminate artifacts in fluorescence images is insufficient, resulting in limited boundary extraction accuracy; there is a lack of a mechanism for dynamically optimizing tumor boundaries and designing safety buffers, and it is impossible to meet the high-precision and safety requirements of laser treatment. Summary of the Invention

[0004] The first aspect of the present invention provides a method for identifying tumor image boundaries by fusing fluorescence boundaries, including:

[0005] Continuously collecting a dynamic fluorescence image sequence of a tumor area through a fluorescence imaging device, and synchronously obtaining real-time respiratory waveform data output by a respiratory motion monitoring sensor;

[0006] Performing multi-scale noise suppression processing on the dynamic fluorescence image sequence, adaptively adjusting filtering parameters based on the image resolution, and generating a preprocessed fluorescence image with artifacts removed;

[0007] Extracting spatial fluorescence intensity distribution features from the preprocessed fluorescence image through a multi-layer feature weighted fusion network, and generating a fluorescence intensity distribution matrix that fuses multi-scale features;

[0008] Constructing a motion compensation model according to the real-time respiratory waveform data, and correcting the boundary displacement of the fluorescence intensity distribution matrix through phase angle tracking and motion vector prediction to obtain corrected fluorescence intensity distribution features;

[0009] Fuse the corrected fluorescence intensity distribution features with the texture structure features of the tumor tissue to generate an initial tumor boundary contour containing multi-dimensional feature weights;

[0010] Perform iterative topological optimization on the initial tumor boundary contour, and eliminate boundary discontinuities through dynamic threshold segmentation and morphological operations to obtain the optimized tumor boundary;

[0011] Map the optimized tumor boundary to the spatial coordinate system of the laser treatment device to generate a boundary mapping model containing a safety buffer and update it in real time.

[0012] Furthermore, the spatial fluorescence intensity distribution features extracted from the preprocessed fluorescence image through a multi-layer feature weighted fusion network include:

[0013] Use three convolutional kernels with different scales to extract local features, regional features, and global features of the fluorescence image respectively;

[0014] Correct each layer of features through a feature attenuation compensation formula, expressed as:

[0015] ;

[0016] where, represents the pixel coordinates, represents the fluorescence intensity feature extracted by the i-th layer of convolutional kernel, represents the fluorescence diffusion attenuation coefficient, represents the fluorescence attenuation compensation factor, represents the corrected fluorescence intensity feature extracted by the i-th layer of convolutional kernel;

[0017] Generate a fluorescence intensity distribution matrix that fuses multi-scale features, including: performing weighted fusion on the corrected features to generate a fluorescence intensity distribution matrix, expressed as:

[0018] ;

[0019] where, is the fluorescence intensity distribution matrix, is the weight coefficient of the i-th feature layer.

[0020] Furthermore, the corrected fluorescence intensity distribution features are obtained through the following method:

[0021] Perform a fast Fourier transform on the respiratory waveform data, extract the dominant frequency component, and calculate the respiratory phase angle;

[0022] Based on the respiratory phase angle, construct a respiratory motion compensation coefficient model to obtain the respiratory motion compensation coefficient. The respiratory motion compensation coefficient model is expressed as:

[0023] ;

[0024] Among them, represents the respiratory phase angle, represents the spatial displacement weight, represents the velocity influence factor, represents the respiratory phase change rate, and C represents the respiratory motion compensation coefficient;

[0025] According to the respiratory motion compensation coefficient, dynamic interpolation correction is performed on the fluorescence intensity distribution matrix to obtain the corrected fluorescence intensity distribution characteristics. Among them, the corrected matrix satisfies the following conditions:

[0026] ;

[0027] Among them, , are the tissue displacement reference amounts calibrated before surgery, and F is the fluorescence intensity distribution matrix function.

[0028] Furthermore, the fusion of the corrected fluorescence intensity distribution characteristics and the texture structure characteristics of the tumor tissue to generate an initial tumor boundary contour including multi-dimensional feature weights includes:

[0029] Calculate the fusion weight of the fluorescence feature and the tissue feature, expressed as follows:

[0030] ;

[0031] Among them, represents the standard deviation of the fluorescence intensity distribution matrix F, represents the standard deviation of the texture feature matrix T;

[0032] Based on the fusion weight and the corrected fluorescence intensity distribution characteristics, generate a fusion boundary feature map, expressed as follows:

[0033] ;

[0034] Among them, represents the fusion boundary feature map, represents the fusion weight, represents the texture feature matrix; represents the corrected fluorescence intensity distribution matrix;

[0035] Perform binarization processing on the fusion boundary feature map to generate an initial tumor boundary contour.

[0036] Furthermore, the performing binarization processing on the fusion boundary feature map to generate an initial tumor boundary contour includes:

[0037] Calculate the dynamic segmentation threshold based on the fusion boundary feature map, expressed as:

[0038] ;

[0039] where τ represents the dynamic segmentation threshold, N represents the total number of pixels, represents the mean of the feature map, is the sensitivity adjustment coefficient, represents the fused boundary feature map;

[0040] Compare the fused boundary feature map with the dynamic segmentation threshold. If the pixel of the fused boundary feature map is greater than the segmentation threshold, the comparison result is 1; if the pixel of the fused boundary feature map is less than the segmentation threshold, the comparison result is 0;

[0041] Generate the initial tumor boundary contour based on the comparison result;

[0042] Perform multi-scale morphological optimization, and the size of the structural element decreases geometrically according to the number of iterations to eliminate the remaining pseudo-boundaries.

[0043] Furthermore, the fluorescence decay compensation factor is determined in the following way:

[0044] Conduct a reference decay experiment using fluorescence calibration films with different depths to obtain the initial value of the fluorescence decay compensation factor;

[0045] Dynamically correct the fluorescence decay compensation factor according to the tissue optical parameters detected in real time, which is expressed as follows:

[0046] ;

[0047] where, represents the calibrated absorbance, is the tissue difference coefficient, represents the fluorescence decay compensation factor, represents the tissue optical parameter, represents the initial value.

[0048] Furthermore, the tissue displacement reference quantity is determined as follows:

[0049] Construct a respiratory phase-tissue displacement mapping relationship table based on the preoperative four-dimensional CT image, and record the tissue displacement amounts corresponding to different respiratory phases;

[0050] Query the reference displacement amount from the mapping relationship table based on the real-time respiratory phase;

[0051] Adjust the reference displacement amount according to the real-time respiratory amplitude A, and the adjustment formula is:

[0052] ;

[0053] ;

[0054] Among them, and is the reference displacement amount, is the respiration amplitude threshold.

[0055] Furthermore, the multi-scale noise suppression processing of the dynamic fluorescence image sequence includes:

[0056] Adopting non-local mean filtering to eliminate the high-frequency noise of the dynamic fluorescence image sequence, and the filtering window size is positively correlated with the image resolution;

[0057] Suppressing the fluorescence leakage artifacts through the anisotropic diffusion algorithm, and the diffusion coefficient is negatively correlated with the local gradient amplitude;

[0058] Performing soft threshold processing on the periodic noise in the wavelet transform domain to retain the low-frequency fluorescence features.

[0059] Furthermore, mapping the optimized tumor boundary to the spatial coordinate system of the laser treatment device includes the following steps:

[0060] Projecting the two-dimensional boundary segmentation result of the optimized tumor boundary into the three-dimensional spatial coordinate system to establish a voxel-level boundary mapping model;

[0061] Generating a dynamic safety buffer according to the kinematic parameters of the laser device, and the buffer width is expressed as follows:

[0062] ;

[0063] Among them, v represents the moving speed of the laser head, represents the system response delay, is the safety margin coefficient, represents the buffer width;

[0064] Generating a buffer layer with a thickness of W on the outer surface of the boundary mapping model.

[0065] In the second aspect of the present invention, a tumor image boundary recognition system integrating fluorescence boundaries is provided. It includes a data acquisition module, a preprocessing module, a matrix generation module, a feature correction module, a contour generation module, a contour optimization module, and a data mapping module;

[0066] The data acquisition module continuously acquires a dynamic fluorescence image sequence of the tumor area through a fluorescence imaging device, and synchronously acquires the real-time respiration waveform data output by the respiration movement monitoring sensor;

[0067] The preprocessing module performs multi-scale noise suppression processing on the dynamic fluorescence image sequence, adaptively adjusts the filtering parameters based on the image resolution, and generates a preprocessed fluorescence image with artifacts removed;

[0068] A matrix generation module extracts spatial fluorescence intensity distribution features from preprocessed fluorescence images through a multi-layer feature weighted fusion network, and generates a fluorescence intensity distribution matrix that fuses multi-scale features;

[0069] A feature correction module constructs a motion compensation model based on real-time respiratory waveform data, and corrects the boundary displacement of the fluorescence intensity distribution matrix through phase angle tracking and motion vector prediction to obtain corrected fluorescence intensity distribution features;

[0070] A contour generation module fuses the corrected fluorescence intensity distribution features and the texture structure features of tumor tissues to generate an initial tumor boundary contour containing multi-dimensional feature weights;

[0071] A contour optimization module performs iterative topological optimization on the initial tumor boundary contour, and eliminates boundary discontinuities through dynamic threshold segmentation and morphological operations to obtain an optimized tumor boundary;

[0072] A data mapping module maps the optimized tumor boundary to the spatial coordinate system of the laser treatment device, generates a boundary mapping model containing a safety buffer, and updates it in real time.

[0073] According to the above embodiments of the present invention, it has at least the following beneficial effects: The method of the present invention can effectively solve many problems in tumor boundary recognition in the prior art. Through a dynamic feature extraction network and a respiratory motion compensation model, combined with phase angle tracking and motion vector prediction technologies, the boundary displacement of the fluorescence intensity distribution matrix caused by respiratory motion can be accurately corrected, so as to maintain the stability and accuracy of tumor boundary recognition under dynamic conditions. At the same time, by fusing the fluorescence intensity distribution features and the texture structure features of tumor tissues, and through dynamic threshold segmentation and multi-scale morphological optimization, a high-precision tumor boundary contour can be generated, eliminating boundary discontinuities and pseudo-boundary residues, providing a reliable basis for subsequent laser treatment.

[0074] In addition, the present invention can also map the optimized tumor boundary to the spatial coordinate system of the laser treatment device in real time, generate a boundary mapping model containing a safety buffer, and dynamically adjust the buffer width according to the kinematic parameters of the laser device. At the same time, by establishing a multi-parameter safety interlock mechanism, the treatment can be paused in time when the boundary is blurred, the respiratory motion changes suddenly, or the device positioning deviation exceeds the safety range, thereby improving the safety and reliability of laser treatment and reducing the risk of accidental injury to normal tissues. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, wherein:

[0076] Figure 1 Schematic flowchart of a tumor image boundary recognition method integrating fluorescence boundaries provided by an embodiment of the present invention. Detailed implementation manners

[0077] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and then implement the present invention, rather than to limit the scope of the present invention in any way. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to be able to fully convey the scope of the present invention to those skilled in the art.

[0078] Those skilled in the art know that the embodiments of the present invention can be implemented as a system, device, equipment, method, or computer program product. Therefore, the present invention can be specifically implemented in the following forms: completely hardware, completely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0079] It should be noted that any number of elements in the drawings is for illustration rather than limitation, and any naming is only for distinction and does not have any limiting meaning.

[0080] Below, reference is made to Figure 1 , Figure 1 Schematic flowchart of a tumor image boundary recognition method integrating fluorescence boundaries provided by an embodiment of the present invention. As Figure 1 shown, a tumor image boundary recognition method 100 integrating fluorescence boundaries includes:

[0081] S1: Continuously collect a dynamic fluorescence image sequence of the tumor region through a fluorescence imaging device, and synchronously obtain real-time respiratory waveform data output by a respiratory motion monitoring sensor;

[0082] S2: Perform multi-scale noise suppression processing on the dynamic fluorescence image sequence, adaptively adjust the filtering parameters based on the image resolution, and generate a preprocessed fluorescence image with artifacts removed;

[0083] S3: Extract spatial fluorescence intensity distribution features from the preprocessed fluorescence image through a multi-layer feature weighted fusion network, and generate a fluorescence intensity distribution matrix integrating multi-scale features;

[0084] S4: Construct a motion compensation model according to the real-time respiratory waveform data, and perform boundary displacement correction on the fluorescence intensity distribution matrix through phase angle tracking and motion vector prediction;

[0085] S5: Integrate the corrected fluorescence intensity distribution features and the texture structure features of the tumor tissue to generate an initial tumor boundary contour containing multi-dimensional feature weights;

[0086] S6: Perform iterative topological optimization on the initial tumor boundary contour, and eliminate boundary discontinuities through dynamic threshold segmentation and morphological operations;

[0087] S7: Map the optimized tumor boundary to the spatial coordinate system of the laser treatment device, generate a boundary mapping model including a safety buffer zone, and update it in real time.

[0088] It should be noted that when implementing the tumor image boundary recognition method, it is first necessary to continuously collect a dynamic fluorescence image sequence of the tumor area through a fluorescence imaging device, and synchronously obtain the real-time respiratory waveform data output by the respiratory motion monitoring sensor. The fluorescence imaging device is a device capable of capturing fluorescence signals. It uses a light source with a specific wavelength to excite fluorescence probes, thereby obtaining fluorescence images of the tumor area. The dynamic fluorescence image sequence refers to multiple frames of images continuously collected within a period of time, and these images can reflect the change of fluorescence signals in the tumor area over time. The respiratory motion monitoring sensor is used to monitor the respiratory state of the patient in real time, and the respiratory waveform data output by it can reflect information such as the respiratory cycle and amplitude, which is crucial for subsequent motion compensation processing.

[0089] Specifically, the fluorescence imaging device can adopt a high-sensitivity camera and a light source with a specific wavelength, such as a near-infrared light source, to improve the capture efficiency of fluorescence signals. The acquisition frequency of the dynamic fluorescence image sequence can be adjusted according to the respiratory cycle, usually set to 5 to 10 frames per second, to ensure that the change of fluorescence signals caused by respiratory motion can be captured. The respiratory motion monitoring sensor can be a device based on pressure, capacitance or optical principles, and the respiratory waveform data output by it can be parameters such as the respiratory cycle and respiratory amplitude. These parameters can be converted into digital signals through the signal processing unit of the sensor for subsequent analysis and processing. During the image acquisition process, it is necessary to ensure the synchronization of the fluorescence imaging device and the respiratory motion monitoring sensor to ensure the temporal consistency between the dynamic fluorescence image sequence and the respiratory waveform data.

[0090] Preferably, in order to improve the accuracy and stability of image acquisition, an automatic exposure and automatic focusing function can be introduced into the fluorescence imaging device to adapt to the imaging requirements of different patients and tumor positions. At the same time, the respiratory motion monitoring sensor can select a high-precision fiber optic sensor, and the respiratory waveform data output by it can be filtered to remove noise to improve the signal quality. In addition, in order to further enhance the dynamic information of the image sequence, time-resolved technology can be introduced during the acquisition process, such as controlling the acquisition time window of fluorescence signals through time gating, so as to more clearly capture the change of fluorescence signals during the respiratory cycle.

[0091] In some embodiments, step S3 includes:

[0092] S31: Extract fluorescence intensity features of different scales using multi-layer convolutional kernels. The feature maps output by each layer are calculated and weighted and fused through the following formula:

[0093] ;

[0094] where represents the pixel coordinates, represents the fluorescence intensity feature extracted by the i-th layer convolutional kernel, represents the fluorescence diffusion attenuation coefficient, is the fluorescence attenuation compensation factor, are the weight coefficients of each feature layer and satisfy .

[0095] It should be noted that when constructing the dynamic feature extraction network, multi-layer convolutional kernels are needed to extract fluorescence intensity features of different scales. The feature maps output by each layer are used to generate the final fluorescence intensity distribution matrix through weighted fusion. In this process, the fluorescence intensity feature refers to the feature information related to the fluorescence signal extracted from the fluorescence image, which is used to reflect the fluorescence signal distribution in the tumor area. The fluorescence diffusion attenuation coefficient is a parameter used to describe the intensity attenuation of the fluorescence signal when propagating in the tissue, while the fluorescence attenuation compensation factor is used to correct this attenuation difference to make the extracted features more accurate. The weight coefficients of each feature layer are used to balance the importance of different scale features, ensuring that the finally fused feature map can comprehensively reflect the fluorescence characteristics of the tumor area.

[0096] Specifically, the design of the multi-layer convolutional kernels can include convolutional kernels of different sizes, such as 3×3, 5×5, etc., to extract fluorescence intensity features from local to global. The fluorescence intensity features extracted by each layer of convolutional kernels will be non-linearly processed through an activation function and then weighted and fused. In the weighted fusion process, the fluorescence diffusion attenuation coefficient can be obtained through experimental calibration. For example, the attenuation of the fluorescence signal is measured in tissues at different depths. The fluorescence attenuation compensation factor can be dynamically adjusted according to the optical properties of the tissue to compensate for the fluorescence signal attenuation differences caused by different depths or different tissue types. The weight coefficients of each feature layer can be assigned according to the importance of the features. Usually, these weights are automatically learned by training a neural network to ensure that their sum is 1, thus ensuring the rationality of feature fusion.

[0097] Preferably, in the feature extraction process, deep learning techniques can be introduced. For example, a convolutional neural network (CNN) can be used to automatically learn the extraction and fusion of fluorescence intensity features. By training a network model with a large amount of labeled data, the parameters and weight coefficients of the convolutional kernel can be optimized, thereby improving the accuracy and robustness of feature extraction. In addition, to further enhance the expression ability of features, an attention mechanism can be added to the feature extraction network, enabling the network to automatically focus on the key features of the tumor region. In the weighted fusion process, an adaptive weight adjustment mechanism can also be introduced to dynamically adjust the weight coefficients according to the statistical characteristics of the feature maps, so as to better adapt to the fluorescence signal distribution of different images.

[0098] In some embodiments, step S4 includes:

[0099] S41: Analyze the respiratory waveform data to calculate the real-time respiratory phase angle and the respiratory amplitude A;

[0100] S42: Construct a respiratory motion compensation coefficient C, whose expression is:

[0101] ;

[0102] where represents the spatial displacement weight, represents the speed influence factor, represents the respiratory phase change rate;

[0103] S43: Dynamically interpolate and correct the fluorescence intensity distribution matrix according to the compensation coefficient C, and the corrected matrix satisfies:

[0104] ;

[0105] where and are the tissue displacement reference amounts calibrated before surgery.

[0106] It should be noted that when constructing the motion compensation model, it is necessary to analyze the respiratory waveform data to calculate the real-time respiratory phase angle and respiratory amplitude, and then construct the respiratory motion compensation coefficient, and use this coefficient to dynamically interpolate and correct the fluorescence intensity distribution matrix. The respiratory waveform data is the signal obtained from the respiratory motion monitoring sensor, which reflects the periodicity and amplitude changes during the patient's breathing process. The real-time respiratory phase angle refers to the phase position in the current respiratory cycle, and the respiratory amplitude represents the movement range of the chest and abdomen during breathing. The respiratory motion compensation coefficient is a parameter used to correct the displacement of the fluorescence image caused by respiratory motion. Through this coefficient, the fluorescence intensity distribution matrix can be corrected, thereby eliminating the interference of respiratory motion on boundary recognition.

[0107] Specifically, when analyzing respiratory waveform data, methods such as Fourier transform or wavelet transform can be used to extract the period and amplitude information of the respiratory signal, and then the real-time respiratory phase angle can be calculated. The respiratory amplitude can be determined by measuring the peak-valley difference of the respiratory waveform. The construction of the respiratory motion compensation coefficient involves parameters such as the spatial displacement weight, the speed influence factor, and the respiratory phase change rate. The spatial displacement weight is used to adjust the influence degree of the compensation coefficient on the displacement, the speed influence factor takes into account the influence of the respiratory motion speed on the image displacement, and the respiratory phase change rate reflects the dynamic characteristics of the respiratory motion. During the dynamic interpolation correction process, the preoperatively calibrated tissue displacement reference quantity is used to determine the specific influence degree of the respiratory motion on the fluorescence intensity distribution matrix, so as to achieve precise correction of the boundary displacement.

[0108] Preferably, when analyzing respiratory waveform data, a high-precision signal processing algorithm, such as a Kalman filter, can be adopted to improve the calculation accuracy of the respiratory phase angle and the respiratory amplitude. For the construction of the respiratory motion compensation coefficient, machine learning methods, such as support vector machines or neural networks, can be used to dynamically adjust the spatial displacement weight and the speed influence factor according to the individual respiratory characteristics of the patient, so as to improve the adaptability and accuracy of the compensation coefficient. In addition, the determination of the tissue displacement reference quantity can be achieved through multimodal image fusion technology, combined with preoperative CT or MRI data, to more accurately calibrate the influence of respiratory motion on tissue displacement and provide a more reliable reference for dynamic interpolation correction.

[0109] In some embodiments, step S5 includes:

[0110] S51: Extract the vascular density distribution map and the cell heterogeneity map of the current case from the tumor tissue database;

[0111] S52: Calculate the fusion weight of the fluorescence feature and the tissue feature :

[0112] ;

[0113] where, represents the standard deviation of the fluorescence intensity distribution matrix F, represents the standard deviation of the texture feature matrix T;

[0114] S53: Generate a fused boundary feature map , which satisfies:

[0115] ;

[0116] It should be noted that when fusing the corrected fluorescence intensity distribution characteristics with the texture structure characteristics of tumor tissues, it is necessary to generate an initial tumor boundary contour containing multi-dimensional feature weights. The fluorescence intensity distribution characteristics refer to the distribution of fluorescence signals after motion compensation correction, reflecting the fluorescence characteristics of the tumor region; while the texture structure characteristics of tumor tissues describe the spatial distribution, shape, and texture information of tumor tissues in the image. The fusion weight is used to balance the importance of these two characteristics to generate a more accurate initial tumor boundary contour. In this way, the fluorescence signal and tissue structure information can be comprehensively considered to improve the accuracy and robustness of boundary recognition.

[0117] Specifically, the standard deviation of the fluorescence intensity distribution matrix ( ) is used to measure the intensity change range of the fluorescence signal, while the standard deviation of the texture feature matrix ( ) reflects the complexity of the texture information. The calculation of the fusion weight is based on the ratio of these two standard deviations, which is used to dynamically adjust the contribution ratio of the fluorescence feature and the texture feature in boundary generation. When the fluorescence signal changes greatly, the weight of the fluorescence feature will increase; while when the texture information is more abundant, the weight of the texture feature will dominate. When generating the fused boundary feature map, by weighted fusing the corrected fluorescence intensity distribution characteristics and texture characteristics, a boundary contour considering multi-dimensional features can be obtained, providing a basis for subsequent boundary optimization.

[0118] Preferably, in order to further optimize the fusion process, an adaptive mechanism can be introduced when calculating the fusion weight to dynamically adjust the parameters in the weight calculation formula according to different tumor types or image qualities. For example, for tumors with weak fluorescence signals, the weight of the texture feature can be appropriately increased to improve the reliability of boundary recognition. In addition, various methods can be used to extract texture features, such as the gray-level co-occurrence matrix (GLCM) or local binary pattern (LBP), to extract richer texture information. In the process of generating the fused boundary feature map, a regularization term can also be introduced to reduce the influence of noise on the boundary contour. For example, methods such as Gaussian smoothing or wavelet denoising can be used to preprocess the fused boundary feature map, thereby further improving the accuracy of boundary recognition.

[0119] In some embodiments, step S6 includes:

[0120] S61: Calculate the dynamic segmentation threshold τ based on the fused boundary feature map, and its expression is:

[0121] ;

[0122] Where N represents the total number of pixels, represents the mean value of the feature map, is the sensitivity adjustment coefficient;

[0123] S62: Perform multi-scale morphological optimization, where the size of the structural element decreases geometrically according to the number of iterations to eliminate the remaining pseudo-boundaries.

[0124] It should be noted that when performing iterative topological optimization on the initial tumor boundary contour, it is necessary to eliminate boundary discontinuities through dynamic threshold segmentation and morphological operations. Dynamic threshold segmentation is a method that dynamically adjusts the segmentation threshold according to image features and is used to separate the tumor boundary from the background. Morphological operations optimize the boundary through mathematical morphology methods to remove pseudo-boundaries and fill boundary gaps, thereby generating a smoother and more continuous tumor boundary. This process aims to improve the accuracy and robustness of boundary recognition and provide a more reliable basis for subsequent treatment.

[0125] Specifically, the threshold calculation of dynamic threshold segmentation is based on the statistical characteristics of the fused boundary feature map, such as the mean and standard deviation of the feature map. By introducing a sensitivity adjustment coefficient, the threshold can be dynamically adjusted according to the complexity of the image to adapt to the segmentation requirements of different tumor boundary features. Morphological operations include erosion, dilation, opening, and closing operations, etc. These operations process the boundary through the structural element. The size of the structural element can be adjusted according to the complexity of the boundary features and is usually optimized in a geometrically decreasing manner during the iteration process to gradually refine the boundary. In each iteration, these operations remove the noise points and pseudo-boundaries on the boundary and fill the small gaps on the boundary at the same time, thereby achieving the elimination of boundary discontinuities.

[0126] Preferably, during the dynamic threshold segmentation process, a local adaptive threshold segmentation algorithm can be introduced, such as the threshold segmentation algorithm based on the Otsu method or local regions, to better adapt to the local changes of the boundary features. For morphological operations, a multi-scale structural element can be adopted and optimized by combining structural elements in different directions to more comprehensively process the discontinuities on the boundary. In addition, deep learning methods, such as convolutional neural networks (CNNs), can be introduced to further refine and correct the optimized boundary to improve the accuracy and robustness of boundary recognition. For example, a CNN can be used to classify the optimized boundary to distinguish the real boundary from the pseudo-boundary, thereby further improving the accuracy of boundary recognition.

[0127] In some embodiments, the fluorescence decay compensation factor is determined in the following manner:

[0128] S61: Conduct a reference decay experiment using fluorescence calibration films with different depths to obtain the initial value of ;

[0129] S62: According to the tissue optical parameters detected in real time for Perform dynamic correction, and the correction formula is:

[0130] ;

[0131] Wherein, represents the calibrated absorbance, is the tissue difference coefficient.

[0132] It should be noted that when determining the fluorescence attenuation compensation factor, first, a reference attenuation experiment needs to be carried out with fluorescence calibration films at different depths to obtain the initial value, and then the initial value is dynamically corrected according to the tissue optical parameters detected in real time. The fluorescence attenuation compensation factor is used to correct the attenuation difference of the fluorescence signal caused by depth and tissue characteristics during its propagation in the tissue, thereby improving the accuracy of the fluorescence intensity distribution characteristics. The fluorescence calibration film is a standard sample with known fluorescence characteristics, used to simulate the fluorescence signal attenuation at different depths, while the tissue optical parameters reflect the absorption and scattering characteristics of the tissue to light. The real-time detection of these parameters can provide a basis for the dynamic correction of the fluorescence attenuation compensation factor.

[0133] Specifically, in the reference attenuation experiment, by placing the fluorescence calibration film at different depth positions and measuring the change in its fluorescence signal intensity, the initial attenuation compensation factor can be obtained. This initial value reflects the attenuation law of the fluorescence signal under standard conditions. The tissue optical parameters detected in real time include the absorbance and scattering coefficient of the tissue, etc., and these parameters can be obtained through optical sensors or imaging devices. The absorbance reflects the degree of light absorption by the tissue, while the scattering coefficient describes the scattering situation of light in the tissue. Dynamically correcting the initial attenuation compensation factor according to these parameters can more accurately compensate for the attenuation difference of the fluorescence signal in the actual tissue, thereby improving the accuracy of the fluorescence intensity distribution characteristics.

[0134] Preferably, when performing the reference attenuation experiment, fluorescence calibration films of multiple depths can be used to cover different depth ranges where tumors may exist. For example, calibration films with depths of 0.5 cm, 1.0 cm, and 1.5 cm, etc., can be selected for the experiment to obtain more comprehensive initial attenuation compensation factor data. For the detection of tissue optical parameters, multispectral imaging technology or photoacoustic imaging technology can be used, and these technologies can provide more accurate tissue optical property information. In the dynamic correction process, machine learning algorithms, such as linear regression or neural networks, can be introduced to automatically adjust the correction formula of the attenuation compensation factor according to the tissue optical parameters detected in real time, thereby further improving the accuracy and adaptability of the compensation.

[0135] In some embodiments, the determination of the tissue displacement reference quantity , includes:

[0136] S71: Construct a respiration phase-tissue displacement mapping table based on preoperative four-dimensional CT images;

[0137] S72: Adjust the reference quantity according to the real-time respiration amplitude A, and the adjustment formula is:

[0138] ;

[0139] ;

[0140] where and are the static calibration reference quantities, and is the respiration amplitude threshold.

[0141] It should be noted that when determining the tissue displacement reference quantity, it is necessary to construct a respiration phase-tissue displacement mapping table based on preoperative four-dimensional CT images and adjust the reference quantity according to the real-time respiration amplitude. Preoperative four-dimensional CT images are an imaging technique that can reflect tissue movement during the respiratory cycle, which records the position changes of tissues at different respiration phases. The respiration phase-tissue displacement mapping table is established by analyzing four-dimensional CT image data to establish a quantitative relationship between the respiration phase and tissue displacement, which is used to predict the influence of respiratory movement on tissue position. The real-time respiration amplitude refers to the actual amplitude of the patient's respiratory movement during the operation, which is obtained through monitoring equipment and is used to adjust the tissue displacement reference quantity to adapt to the patient's current respiratory state.

[0142] Specifically, when constructing the respiration phase-tissue displacement mapping table, it is necessary to extract the displacement data of tissues at each respiration phase from preoperative four-dimensional CT images. These data can be obtained through image registration technology, that is, aligning CT images at different time points and calculating the displacement vectors of tissues during the respiratory cycle. The static calibration reference quantity is a tissue displacement reference value pre-calculated based on preoperative image data, usually including the amplitude and direction of displacement. The respiration amplitude threshold is a preset reference value used to determine whether the real-time respiration amplitude is within the normal range. When adjusting the reference quantity according to the real-time respiration amplitude, a non-linear function can be used to describe the relationship between the respiration amplitude and tissue displacement, such as a logarithmic function or an exponential function, to more accurately reflect the influence of respiratory movement on tissue displacement.

[0143] Preferably, when constructing the respiration phase-tissue displacement mapping relationship table, a more refined image registration algorithm can be adopted, such as feature point-based registration or voxel-based registration, to improve the accuracy of displacement data. For the real-time monitoring of respiration amplitude, a high-precision respiration sensor can be used, such as a fiber optic sensor or a capacitance sensor, to obtain a more stable and accurate respiration signal. When adjusting the reference quantity, an adaptive adjustment mechanism can be introduced. For example, through a machine learning algorithm, the parameters of the non-linear function can be dynamically adjusted according to the patient's respiration pattern, so as to better adapt to the respiration characteristics of different patients. In addition, the physiological characteristics of the patient, such as age, weight, and lung function, can be combined to optimize the adjustment formula of the reference quantity, so as to further improve the accuracy of tissue displacement prediction.

[0144] In some embodiments, the multi-scale noise suppression in step S2 includes:

[0145] S81: Use non-local means filtering to eliminate high-frequency noise, and the size of the filtering window is positively correlated with the image resolution;

[0146] S82: Suppress fluorescence leakage artifacts through an anisotropic diffusion algorithm, and the diffusion coefficient is negatively correlated with the local gradient magnitude;

[0147] S83: Perform soft threshold processing on periodic noise in the wavelet transform domain to retain low-frequency fluorescence features.

[0148] It should be noted that the multi-scale noise suppression processing is a preprocessing method for dynamic fluorescence image sequences, aiming to eliminate noise and artifacts in the images through various technical means, so as to improve the image quality and provide a clearer input for subsequent tumor boundary recognition. Non-local means filtering is a noise reduction method based on image similarity, which removes high-frequency noise by calculating the weighted average of similar pixels in the image while retaining the detailed information of the image. The anisotropic diffusion algorithm is an image processing technology based on partial differential equations, which suppresses fluorescence leakage artifacts by controlling the direction and intensity of diffusion while retaining the edge information of the image. The soft threshold processing in the wavelet transform domain decomposes the image signal into different frequency levels and performs threshold processing on the high-frequency noise components to remove periodic noise and retain low-frequency fluorescence features.

[0149] Specifically, the filtering window size of non-local means filtering is positively correlated with the image resolution. That is, for high-resolution images, a larger window needs to be used to cover more similar pixels, so as to achieve better noise reduction effects. The diffusion coefficient of the anisotropic diffusion algorithm is negatively correlated with the local gradient magnitude, which means that diffusion is reduced in regions with large image gradients (such as edges), while diffusion is increased in regions with small gradients (such as smooth regions), so as to retain edge information while suppressing artifacts. Soft threshold processing in the wavelet transform domain removes the noise components corresponding to small coefficients by performing threshold processing on wavelet coefficients, while retaining the low-frequency fluorescence features corresponding to large coefficients. The parameter settings of these methods need to be optimized according to specific application scenarios and image characteristics, such as the filtering window size, the initial value and attenuation rate of the diffusion coefficient, etc.

[0150] Preferably, in non-local means filtering, an adaptive window size method can be adopted to dynamically adjust the window size according to the complexity of the local image, so as to achieve a better balance between noise reduction and detail retention. For the anisotropic diffusion algorithm, an adaptive adjustment mechanism for the time step can be introduced to accelerate the convergence speed of the algorithm and improve the processing efficiency. In the soft threshold processing in the wavelet transform domain, the threshold can be dynamically selected in combination with the statistical characteristics of the image, such as using a histogram-based threshold selection method, so as to better remove noise and retain useful information. In addition, the advantages of multiple noise suppression methods can be combined to design a hybrid noise suppression algorithm to further improve the effect of image preprocessing. For example, first remove high-frequency noise through non-local means filtering, then use the anisotropic diffusion algorithm to suppress artifacts, and finally enhance low-frequency features through soft threshold processing in the wavelet transform domain, so as to provide higher-quality image input for subsequent tumor boundary recognition.

[0151] In some embodiments, step S7 includes:

[0152] S91: Project the two-dimensional boundary segmentation result onto a three-dimensional space coordinate system to establish a voxel-level boundary mapping model;

[0153] S92: Generate a dynamic safety buffer according to the kinematic parameters of the laser device, and the buffer width W satisfies:

[0154] ;

[0155] where v represents the moving speed of the laser head, represents the system response delay, is the safety margin coefficient.

[0156] It should be noted that when mapping the two-dimensional boundary segmentation result to the spatial coordinate system of the laser treatment device, a boundary mapping model including a safety buffer will be generated. This process aims to ensure that the optimized tumor boundary can accurately fit the operation space of the laser treatment device. At the same time, through the design of the safety buffer, accidental injury to surrounding normal tissues can be avoided. Among them, the voxel-level boundary mapping model is a three-dimensional grid-based model that can accurately represent the tumor boundary and the surrounding safety area, while the kinematic parameters of the laser device involve characteristics such as the moving speed and acceleration of the laser head. These parameters are used to calculate the width of the safety buffer to ensure the safety and effectiveness of the treatment.

[0157] Specifically, when projecting the two-dimensional boundary segmentation result onto the three-dimensional spatial coordinate system, it is necessary to convert the pixel coordinates in the two-dimensional image into three-dimensional coordinates in the operation space of the laser treatment device. This process can be achieved through geometric transformation or interpolation algorithms. The width of the safety buffer is jointly determined by the moving speed of the laser head, the system response delay, and the safety margin coefficient. The moving speed of the laser head refers to the speed at which the laser moves along a predetermined path during treatment. The system response delay refers to the time difference between detecting the boundary information and the laser actually performing an action. The safety margin coefficient is a parameter determined based on experience or experiments, which is used to increase an additional safety distance to ensure the safety of the treatment.

[0158] Preferably, when establishing the voxel-level boundary mapping model, a high-resolution three-dimensional grid can be used to more accurately represent the tumor boundary and the safety buffer. For example, a grid with a voxel size of 1 mm³ can be used to build the model, thereby improving the accuracy of the treatment. For the calculation of the safety buffer, it can be optimized according to the specific parameters of the laser device. For example, for a high-precision laser device, the safety margin coefficient can be appropriately reduced to improve the treatment efficiency; while for a device with lower precision, the safety margin coefficient needs to be increased to ensure the safety of the treatment. In addition, a real-time feedback mechanism can be introduced to dynamically adjust the width of the safety buffer according to the real-time data during the laser treatment process to further improve the safety and adaptability of the treatment.

[0159] In some embodiments, it further includes:

[0160] Establish a multi-parameter safety interlock mechanism, and trigger treatment suspension when the following conditions are simultaneously met:

[0161] (1) The boundary blur index > threshold, where is the boundary gradient change amount, is the maximum boundary intensity value;

[0162] (2) The respiratory motion mutation rate > preset warning value, A is the respiratory amplitude;

[0163] (3)Device positioning deviation >Precision tolerance, which is the Euclidean distance between the actual coordinates and the theoretical coordinates.

[0164] It should be noted that the multi-parameter safety interlock mechanism established in the present invention is used to monitor key parameters in real time and trigger treatment suspension during tumor image boundary recognition and laser treatment to ensure the safety and reliability of treatment. The boundary blurriness index is a parameter that measures the boundary clarity by calculating the ratio of the boundary gradient change amount to the maximum boundary intensity value. The respiratory motion mutation rate refers to the change rate of the respiratory amplitude per unit time and is used to evaluate the stability of respiratory motion. The device positioning deviation refers to the Euclidean distance between the actual coordinates and the theoretical coordinates of the laser device and is used to monitor the positioning accuracy of the device. When these parameters exceed the preset thresholds, treatment suspension is triggered to avoid treatment risks caused by inaccurate boundary recognition, abnormal respiratory motion, or device positioning deviation.

[0165] Specifically, the calculation of the boundary blurriness index is based on the boundary gradient change amount and the maximum boundary intensity value. The boundary gradient change amount reflects the degree of change in pixel intensity in the boundary region, while the maximum boundary intensity value is the maximum value of pixel intensity in the boundary region. The calculation of the respiratory motion mutation rate requires real-time monitoring of the change in respiratory amplitude and calculating its change rate through the difference method. The calculation of the device positioning deviation requires real-time acquisition of the actual coordinates and theoretical coordinates of the laser device and calculating the deviation value through the Euclidean distance formula. The preset warning value and precision tolerance are thresholds determined based on experimental data and clinical experience and are used to judge whether to trigger treatment suspension. For example, the threshold of the boundary blurriness index can be set to 0.1, the warning value of the respiratory motion mutation rate can be set to 0.5, and the precision tolerance of the device positioning deviation can be set to 1 mm.

[0166] Preferably, in the multi-parameter safety interlock mechanism, more refined monitoring methods can be introduced. For example, for the boundary blurriness index, dynamic evaluation can be combined with local regions of the image instead of a single global calculation to more accurately reflect the change in boundary clarity. For the respiratory motion mutation rate, the sliding window method can be used to calculate its short-term change trend to improve the monitoring sensitivity. For the device positioning deviation, predictive deviation evaluation can be combined with the kinematic model of the laser device to detect potential positioning problems in advance. In addition, a redundant monitoring mechanism can also be introduced, such as simultaneously monitoring respiratory motion or device positioning through multiple sensors to improve the reliability and accuracy of monitoring results.

[0167] The above-mentioned embodiments of the present invention have the following beneficial effects: By integrating the fluorescence boundary with the tumor tissue characteristics and combining technical means such as multi-scale noise suppression, dynamic feature extraction, respiratory motion compensation, boundary optimization, and safety buffer design, the present invention can improve the accuracy and stability of tumor image boundary recognition. In the image preprocessing stage, methods such as non-local mean filtering, anisotropic diffusion, and wavelet transform are used to eliminate high-frequency noise, suppress fluorescence leakage artifacts, and retain low-frequency fluorescence characteristics, thereby improving the image quality and providing a clearer input for subsequent boundary extraction; in the feature extraction and fusion process, a dynamic feature extraction network is constructed and a fluorescence decay compensation factor is introduced to extract the fluorescence intensity distribution characteristics and compensate for the fluorescence decay difference. At the same time, combined with the texture structure characteristics of the tumor tissue, by dynamically calculating the fusion weight, an initial tumor boundary contour containing multi-dimensional feature weights can be generated, further improving the accuracy of boundary recognition; through the respiratory motion compensation model to dynamically interpolate and correct the fluorescence intensity distribution matrix, the boundary displacement problem caused by respiratory motion can be solved, ensuring the stability of boundary recognition; in the boundary optimization stage, dynamic threshold segmentation and multi-scale morphological optimization are used to eliminate boundary discontinuity and pseudo-boundary residues, generating a more accurate tumor boundary contour; mapping the optimized boundary to the spatial coordinate system of the laser treatment device and generating a boundary mapping model can provide an accurate positioning basis for laser treatment. At the same time, combined with the dynamic safety buffer design, the buffer width can be adjusted in real time according to the kinematic parameters of the laser device, further improving the safety of treatment.

[0168] In addition, the present invention also establishes a multi-parameter safety interlock mechanism, which can trigger the treatment pause in time when the boundary blur exceeds the threshold, the respiratory motion mutation rate exceeds the warning value, or the device positioning deviation exceeds the accuracy tolerance, thereby effectively avoiding the treatment risks caused by inaccurate boundary recognition or device operation errors and improving the safety and reliability of laser treatment; by pre-operative calibration and real-time correction of the tissue displacement reference quantity, the influence of respiratory motion on the boundary can be compensated more accurately, further improving the dynamic adaptability of boundary recognition; overall, the present invention can provide more efficient and safer technical support for the precise diagnosis and laser treatment of tumors.

[0169] Further, the storage medium of the embodiments of the present application stores program instructions capable of implementing all of the above methods. Among them, the program instructions can be stored in the above storage medium in the form of a software product, including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the methods described in the various embodiments of the present application. The foregoing storage medium includes: various media that can store program codes such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs, or terminal devices such as computers, servers, mobile phones, and tablets.

[0170] The above description is only some preferred embodiments of the present invention and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solutions formed by the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the technical solutions formed by mutually replacing the above features with the technical features (but not limited to) having similar functions disclosed in the embodiments of the present invention.

Claims

1. A method for identifying the boundary of a tumor image that integrates a fluorescence boundary, characterized in that, Including the following steps: Continuously acquire a dynamic fluorescence image sequence of the tumor region through a fluorescence imaging device, and synchronously obtain real-time respiratory waveform data output by a respiratory motion monitoring sensor; Perform multi-scale noise suppression processing on the dynamic fluorescence image sequence, adaptively adjust the filtering parameters based on the image resolution, and generate a preprocessed fluorescence image with artifacts removed; Extract the spatial fluorescence intensity distribution features from the preprocessed fluorescence image through a multi-layer feature weighted fusion network, and generate a fluorescence intensity distribution matrix integrating multi-scale features; Construct a motion compensation model according to the real-time respiratory waveform data, and perform boundary displacement correction on the fluorescence intensity distribution matrix through phase angle tracking and motion vector prediction to obtain the corrected fluorescence intensity distribution features. Specifically, it includes the following steps: perform fast Fourier transform on the respiratory waveform data, extract the dominant frequency component to calculate the respiratory phase angle; based on the respiratory phase angle, construct a respiratory motion compensation coefficient model to obtain the respiratory motion compensation coefficient; perform dynamic interpolation correction on the fluorescence intensity distribution matrix according to the respiratory motion compensation coefficient to obtain the corrected fluorescence intensity distribution features; Fuse the corrected fluorescence intensity distribution features with the texture structure features of the tumor tissue to generate an initial tumor boundary contour containing multi-dimensional feature weights; Perform iterative topological optimization on the initial tumor boundary contour, and eliminate boundary discontinuities through dynamic threshold segmentation and morphological operations to obtain the optimized tumor boundary; Map the optimized tumor boundary to the spatial coordinate system of the laser treatment device, generate a boundary mapping model containing a safety buffer, and update it in real time.

2. The method according to claim 1, wherein Extracting the spatial fluorescence intensity distribution features from the preprocessed fluorescence image through a multi-layer feature weighted fusion network includes: Using three-layer convolution kernels with different scales to extract local features, regional features, and global features of the fluorescence image respectively; Correct each layer of features through a feature attenuation compensation formula, expressed as: ； Among them, represents the pixel coordinates, represents the fluorescence intensity feature extracted by the i-th layer of convolution kernels, represents the fluorescence diffusion attenuation coefficient, represents the fluorescence attenuation compensation factor, represents the fluorescence intensity feature extracted by the corrected i-th layer of convolution kernels; Generating a fluorescence intensity distribution matrix integrating multi-scale features includes: performing weighted fusion on the corrected features to generate a fluorescence intensity distribution matrix, expressed as: ； Among them, is the fluorescence intensity distribution matrix, is the weight coefficient of the i-th characteristic layer.

3. The method according to claim 2, wherein The respiratory motion compensation coefficient model, expressed as: ； Among them, represents the respiratory phase angle, represents the spatial displacement weight, represents the speed influence factor, represents the respiratory phase change rate, and C represents the respiratory motion compensation coefficient; The fluorescence intensity distribution matrix after the dynamic interpolation correction satisfies the following conditions: ； Among them, , are the tissue displacement reference amounts calibrated before the operation, and F is the fluorescence intensity distribution matrix function.

4. The method according to claim 3, characterized in that, Fusing the corrected fluorescence intensity distribution features with the texture structure features of the tumor tissue to generate an initial tumor boundary contour containing multi-dimensional feature weights includes: Calculate the fusion weight of the fluorescence feature and the tissue feature, expressed as follows: ； Among them, represents the standard deviation of the fluorescence intensity distribution matrix F, represents the standard deviation of the texture feature matrix T; Based on the fusion weight and the corrected fluorescence intensity distribution features, generate a fused boundary feature map, expressed as follows: ； Among them, represents the fused boundary feature map, represents the fusion weight, represents the texture feature matrix; represents the corrected fluorescence intensity distribution matrix; Perform binary processing on the fused boundary feature map to generate an initial tumor boundary contour.

5. The method according to claim 4, characterized in that Performing binary processing on the fused boundary feature map to generate an initial tumor boundary contour includes: Calculate the dynamic segmentation threshold based on the fused boundary feature map, expressed as: ； where τ represents the dynamic segmentation threshold, N represents the total number of pixels, represents the mean of the feature map, is the sensitivity adjustment coefficient, represents the fused boundary feature map; Compare the fused boundary feature map with the dynamic segmentation threshold. If the pixel of the fused boundary feature map is greater than the segmentation threshold, the comparison result is 1; if the pixel of the fused boundary feature map is less than the segmentation threshold, the comparison result is 0; Generate an initial tumor boundary contour based on the comparison result; Perform multi-scale morphological optimization, and the size of the structural element decreases geometrically according to the number of iterations to eliminate the remaining pseudo-boundaries.

6. The method according to claim 2, wherein The fluorescence decay compensation factor is determined in the following manner: Perform a reference decay experiment using fluorescence calibration films at different depths to obtain the initial value of the fluorescence decay compensation factor; Dynamically correct the fluorescence decay compensation factor based on the optically detected tissue parameters in real time, which is expressed as follows: ； Among them, represents the calibrated absorbance, is the tissue difference coefficient, represents the fluorescence decay compensation factor, represents the tissue optical parameter, represents the initial value.

7. The method according to claim 3, wherein The tissue displacement reference quantity is determined as follows: Construct a respiratory phase-tissue displacement mapping relationship table based on the preoperative four-dimensional CT images, and record the tissue displacement amounts corresponding to different respiratory phases; Query the reference displacement amount from the mapping relationship table based on the real-time respiratory phase; Adjust the reference displacement amount according to the real-time respiratory amplitude A, and the adjustment formula is: ； ； Among them, , is the reference displacement amount, is the breathing amplitude threshold.

8. The method according to claim 1, wherein The multi-scale noise suppression process for the dynamic fluorescence image sequence includes: Use non-local means filtering to eliminate the high-frequency noise in the dynamic fluorescence image sequence, and the filtering window size is positively correlated with the image resolution; Suppress the fluorescence leakage artifacts through the anisotropic diffusion algorithm, and the diffusion coefficient is negatively correlated with the local gradient amplitude; Perform soft threshold processing on the periodic noise in the wavelet transform domain to retain the low-frequency fluorescence features.

9. The method according to claim 1, characterized in that Mapping the optimized tumor boundary to the spatial coordinate system of the laser treatment device includes the following steps: Project the two-dimensional boundary segmentation result of the optimized tumor boundary into the three-dimensional spatial coordinate system to establish a voxel-level boundary mapping model; Generate a dynamic safety buffer according to the kinematic parameters of the laser device, and the buffer width is expressed as follows: ； Among them, v represents the moving speed of the laser head, represents the system response delay, is the safety margin coefficient, represents the buffer width; Generate a buffer layer with a thickness of W on the outer surface of the boundary mapping model.

10. A tumor image boundary recognition system integrating fluorescence boundaries, characterized in that, Including a data acquisition module, a preprocessing module, a matrix generation module, a feature correction module, a contour generation module, a contour optimization module, and a data mapping module; The data acquisition module continuously acquires a dynamic fluorescence image sequence of the tumor area through a fluorescence imaging device, and simultaneously acquires the real-time respiratory waveform data output by the respiratory motion monitoring sensor; The preprocessing module performs multi-scale noise suppression processing on the dynamic fluorescence image sequence, adaptively adjusts the filtering parameters based on the image resolution, and generates a preprocessed fluorescence image with artifacts removed; The matrix generation module extracts the spatial fluorescence intensity distribution features from the preprocessed fluorescence image through a multi-layer feature weighted fusion network, and generates a fluorescence intensity distribution matrix that fuses multi-scale features; The feature correction module constructs a motion compensation model based on the real-time respiratory waveform data, and corrects the boundary displacement of the fluorescence intensity distribution matrix through phase angle tracking and motion vector prediction to obtain the corrected fluorescence intensity distribution features; The contour generation module fuses the corrected fluorescence intensity distribution features with the texture structure features of the tumor tissue to generate an initial tumor boundary contour containing multi-dimensional feature weights; The contour optimization module performs iterative topological optimization on the initial tumor boundary contour, and eliminates the boundary discontinuity through dynamic threshold segmentation and morphological operations to obtain the optimized tumor boundary; The data mapping module maps the optimized tumor boundary to the spatial coordinate system of the laser treatment device, generates a boundary mapping model including a safety buffer, and updates it in real time.

Images