Helium-neon laser therapy method and application based on adaptive beam focusing

Abstract

Disclosed in the present invention are a helium-neon laser therapy method and application based on adaptive beam focusing. The technical method belongs to the technical field of medical lasers, and realizes adaptive beam focusing optimization for helium-neon laser therapy. The process comprises: patient assessment, which involves acquiring an image and data of a treatment region; image processing and analysis, which involves performing feature extraction and lesion segmentation by using a convolutional neural network (CNN) and a U-Net algorithm; laser parameter setting based on wavefront optimization, which involves adjusting laser parameters by means of a feedback control system; therapy execution, which involves tracking and adjusting a beam in real time by means of wavefront optimization and an optical flow algorithm; result assessment and adjustment, which involves assessing a treatment effect by using the CNN and the U-Net algorithm again, and adjusting the laser parameters for repeated treatments if necessary; and finally, assessing a treatment result for the patient after the treatment is completed. The method improves the accuracy, effect and safety of treatment, and increases the degree of satisfaction of a patient.

Description

A method and application of He-Ne laser treatment based on adaptive beam focusing Technical Field

[0001] The present invention relates to the field of medical laser technology, and in particular to a method and application of He-Ne laser treatment based on adaptive beam focusing. Background Art

[0002] Helium-neon lasers are commonly used to treat epidermal lesions, alleviate inflammation, and in cosmetic dermatology, demonstrating particular advantages in stimulating collagen production and improving skin quality. Adaptive optics systems dynamically adjust the beam to compensate for tissue irregularities and optical distortion, ensuring precise laser delivery to the target area, significantly improving the precision, safety, and personalization of treatment. The integration of these technologies not only enhances treatment effectiveness but also offers broad application prospects for modern medicine, particularly in providing precise and personalized treatment plans. Existing treatment technologies are unable to adapt their capabilities to treat specific conditions based on lesion type. Technical Solutions

[0003] To solve these problems, this patent proposes a method and application of He-Ne laser treatment based on adaptive beam focusing, which solves the related problems through the setting and arrangement of the circuit:

[0004] 1. Lack of automated laser parameter adjustment based on lesion image information to treat different conditions: Poor treatment outcomes: Laser parameters cannot be precisely adjusted based on lesion characteristics (such as size, depth, or type), potentially resulting in suboptimal treatment outcomes. For example, too low an energy level may result in incomplete lesion healing, while too high an energy level may damage surrounding healthy tissue. Increased treatment risks: Inappropriate laser settings may cause unnecessary tissue damage, such as burns, hyperpigmentation, and scarring, due to laser parameters not being optimized for specific lesion characteristics. Prolonged treatment times: Laser parameters cannot automatically adjust to the specific conditions of the lesion, requiring physicians to manually test and modify settings, which increases the duration of individual treatments and may require more treatments to achieve a cure. Patient dissatisfaction and decreased trust: Frequent adjustments and treatment failures may lead to patient dissatisfaction, undermining their trust in the treatment and overall healthcare delivery. Financial burden: Ineffective or suboptimal treatment outcomes may require more follow-up visits and additional treatments, increasing the patient's medical expenses. Waste of medical resources: More time and resources are needed to deal with the complexity and uncertainty of treatment, which may lead to waste of medical resources, including time, equipment usage and labor of medical staff. Therefore, there is an urgent need for a method and application of helium-neon laser treatment that can change laser parameters according to different symptoms.

[0005] 2. During treatment, the laser focus may drift away from the lesion, causing the treatment position to shift. Suboptimal treatment efficacy: Failure to accurately target the lesion will result in inadequate treatment, potentially leaving the lesion untreated and requiring more treatments to achieve the desired outcome. Damage to healthy tissue: Inadvertent laser exposure to surrounding healthy tissue can cause unnecessary damage, such as burns, hyperpigmentation, or scarring. This damage not only affects aesthetics but can also lead to long-term functional impairment or paresthesia. Increased treatment risks: Lack of treatment precision increases treatment risks, such as pain, infection, and a more complex recovery process. Increased treatment time and cost: Poor initial treatment efficacy may require multiple follow-up visits to correct the problem, prolonging the treatment cycle and increasing the overall cost. Reduced patient satisfaction: Inaccurate treatment can directly impact patient satisfaction and trust in the treatment provider, potentially leading to resistance to future treatment or seeking alternative treatment options. Therefore, a method and application for HeNe laser treatment that can adapt the laser focus position to the patient's movement is urgently needed.

[0006] 3. Inability to Automatically Extract Features: Inaccurate Treatment Results: The lack of automated feature extraction means that laser parameters cannot be adjusted based on the specific characteristics of the lesion (such as size, shape, and depth) during treatment, potentially resulting in less targeted treatment and impacting treatment effectiveness. Increased Operational Complexity: Physicians or technicians must manually analyze images and data to determine lesion characteristics, which not only increases operational complexity and time, but also increases reliance on operator skill. Increased Risk of Human Error: Manual processing and interpretation of image data is more prone to errors, especially in busy or stressful medical environments. Misinterpretation of data can lead to inappropriate treatment parameter settings. Prolonged Treatment Time: Without automated feature extraction, additional time is required for data analysis and preparation before each treatment, extending the overall treatment cycle. Reduced Patient Throughput: Increased processing time for each patient results in a decrease in the number of patients that can be seen per day. This can mean decreased efficiency and revenue for medical institutions, necessitating an urgent need for a method and application for automated feature extraction for He-Ne laser treatment. Beneficial effects

[0007] The main purpose of the present invention is to provide a method and application of He-Ne laser treatment based on adaptive beam focusing, which can effectively solve the problems in the background technology.

[0008] To achieve the above object, the technical solution adopted by the present invention is:

[0009] A method and application of He-Ne laser treatment based on adaptive beam focusing, including wavefront optimization, a feedback control system, a convolutional neural network (CNN), an image segmentation algorithm (U-Net), and an optical flow algorithm. The method steps are as follows:

[0010] Step 1: Patient Assessment: Input: Images and data of the patient's treatment area. Output: Images and patient data are prepared.

[0011] Step 2: Image Processing and Analysis: Use a convolutional neural network (CNN) to extract image features and identify lesions. Use a U-Net to accurately segment the treatment area. Output: Precise location of the treatment area and confirmation of lesion type.

[0012] Step 3: Wavefront Optimization Laser Parameter Setting: Based on the image analysis results, the laser parameters (wavelength, energy density, pulse width, etc.) are adjusted through the feedback control system. Output: The laser equipment preset parameters are completed.

[0013] Step 4: Treatment Execution: Wavefront optimization technology is used to optimize the beam's focus. Optical flow algorithms are used to track the movement of the affected area in real time and adjust the beam's position. Output: The treatment process is dynamically adjusted and executed.

[0014] Step 5: Results Evaluation and Subsequent Adjustments: Based on the post-treatment images and data feedback, CNN and U-Net are used again to evaluate the results. If adjustments are required, return to Step 3, reset the laser parameters, and perform the treatment.

[0015] Step 6: Treatment completion: Patient evaluation is completed and the treatment results are satisfactory.

[0016] Furthermore, the convolutional neural network (CNN) uses optimization algorithms such as ReLU function, softmax function, SGD (stochastic gradient descent), and Adam. The process of extracting image information is as follows: Image preprocessing: Before input into CNN, the image needs to undergo a series of preprocessing steps to improve the learning efficiency and effect of the network: Normalization: Scale the image pixel values ​​to a certain range, such as 0 to 1 or -1 to 1, to reduce numerical instability during model training. Resize: Modify the image size to meet the requirements of the network input layer. Enhancement: Including rotation, scaling, cropping, etc., to increase the generalization ability of the model. The convolution layer is responsible for extracting features from the image, using a set of learned filters (or called convolution kernels) to perform convolution operations with the image to extract local features of the image. Each filter captures a different feature of the image, such as edges, corners, or more complex textures. ReLU (Rectified Linear Unit) is used as an activation function to increase the network's nonlinear capabilities and help capture complex patterns. Pooling layers follow convolutional layers to reduce the spatial size of features, reducing the number of parameters and computational effort, while also making the features more spatially invariant. The maximum value within a specific region is selected as the representative of that region, and the average value within that region is calculated. Fully connected layers, after multiple convolutional and pooling layers, are used to classify or regress the extracted features. The feature maps output by the convolutional and pooling layers are flattened into one-dimensional vectors. The fully connected layers learn how to make accurate predictions based on these feature vectors, classifying the lesion type. The final output layer is designed based on the task (classification or regression). In classification tasks, the output layer uses the softmax function, which converts the output into a probability distribution. Training and optimization: The loss function defines how to calculate the error between the predicted value and the true value (cross-entropy loss is used for classification). Optimization algorithms such as SGD (stochastic gradient descent) and Adam are used to adjust network weights to minimize the loss function.

[0017] Furthermore, the image segmentation algorithm consists of multiple convolutional layers and pooling layers, which gradually reduce the spatial dimension and increase the depth of features. These layers extract increasingly complex features by learning local features in the image, restoring the spatial dimension of the image and reducing the feature depth. Through upsampling and convolution, the expansion path maps high-level features back to higher-resolution outputs for precise positioning. Utilizing the skip connection of U-Net, it directly connects the feature map of the contraction path to the corresponding layer of the expansion path. These connections help retain the position information lost by the downsampling process during the upsampling process, allowing the network to better make pixel-level accurate predictions. In the expansion path, the upsampling layer gradually increases the resolution of the feature map, while refining these features through the convolution layer. Transposed convolution (sometimes called deconvolution) or simple interpolation methods are used to increase the size of the feature map. Convolutional layers are used to merge features from skip connections, refine features, and restore details. The last layer is a convolutional layer that converts the final feature map into a segmentation map. A 1x1 convolution kernel is used to convert the multi-channel feature map into a single-channel output. Each pixel in the output layer predicts the category of its corresponding input pixel. The segmentation task uses the Dice loss function for cases where the data categories are unbalanced. Backpropagation and optimization algorithms (such as Adam) adjust network parameters, minimize the loss function, and provide data reference during segmentation.

[0018] Furthermore, in a He-Ne laser treatment system, the feedback control system utilizes sensors to collect real-time data on the treatment area, including light absorption and scattering rates. Through data analysis and processing, it identifies laser parameters that require adjustment. The system employs a real-time monitoring algorithm to continuously assess treatment effectiveness and, through a parameter adjustment algorithm, to modify the laser power and pulse width in real time to adapt to the skin's response. Furthermore, a result feedback and iterative optimization algorithm adjusts subsequent treatment strategies based on treatment effectiveness, while a safety protection algorithm ensures patient safety throughout the treatment process.

[0019] Furthermore, in adaptive beam-focusing He-Ne laser therapy systems, wavefront optimization technology works in conjunction with a feedback control system to ensure the accuracy and effectiveness of laser treatment. Wavefront optimization uses sensors to collect real-time distortion data of reflected light waves, analyzes this data, calculates correction parameters, and uses a deformable mirror or optical modulator to adjust the beam in real time to optimize focus quality. Simultaneously, the feedback control system monitors the optical properties of the treatment area and the skin's response, adjusting the laser's power and pulse width based on this real-time data to ensure that the laser parameters match the treatment needs and skin condition.

[0020] Furthermore, in adaptive beam-focusing He-Ne laser therapy systems, the optical flow algorithm is used to track the dynamic position and movement of the treatment area in real time, ensuring that the laser is precisely adjusted to the correct treatment location. This algorithm analyzes pixel changes between consecutive image frames before and after treatment to estimate the movement of the treatment area, calculate the speed and direction of movement, and automatically adjust the laser's position, angle, focus depth, and intensity.

[0021] Compared with the prior art, the present invention has the following beneficial effects:

[0022] 1. This invention uses a convolutional neural network (CNN) and an image segmentation algorithm (U-Net) to analyze and capture disease images and generate relevant data. A feedback control system then modifies laser data to treat different conditions: Enhanced treatment precision: Analyzing disease images using CNN and U-Net can precisely identify and locate lesions, ensuring that laser light only targets the target lesion and minimizing the impact on surrounding healthy tissue. Personalized treatment plans: Through image analysis, the system can adjust laser parameters, such as power, pulse width, and frequency, based on the specific characteristics of the disease (e.g., size, shape, depth), providing a personalized treatment plan. Automated operation: The feedback control system automatically adjusts laser parameters based on real-time data, reducing manual intervention by the physician and the risk of operational errors. Improved treatment efficiency and safety: Automated and precise laser control enables treatment to be completed in a shorter time, while reducing the risk of laser overexposure and enhancing treatment safety. Optimized resource utilization: More efficient treatment planning and rapid response times mean higher patient throughput and better resource allocation. Improved patient satisfaction: Accurate treatment results and shorter recovery times can enhance patient satisfaction and trust in treatment providers. Reduced need for follow-up treatment: Accurate first-time treatment reduces the likelihood of re-treatment, reducing long-term treatment costs and inconvenience for patients.

[0023] 2. This invention aims to track the treatment area through dynamic tracking using an optical flow algorithm, ensuring that laser focus remains accurate even with slight patient movement. Continuous Precision Treatment: Even with slight patient movement during treatment, the optical flow algorithm adjusts the laser focus in real time, ensuring that the laser remains precisely focused on the intended lesion area, thereby improving treatment effectiveness and accuracy. Reduced Treatment Risk: By ensuring the precision of the laser focus, potential damage to surrounding healthy tissue is minimized, avoiding unwanted side effects such as burns or scarring caused by focusing errors. Improved Patient Comfort: Patients do not need to remain completely still during treatment and can breathe naturally, reducing discomfort and anxiety. This enhances patient comfort and satisfaction during treatment. Accelerated Treatment: Dynamic tracking allows for continuous, uninterrupted treatment without frequent equipment adjustments or patient repositioning, thereby shortening the total duration of a single treatment session. Improved Automation of Treatment Equipment: The highly automated tracking system reduces the burden on medical staff during operation, allowing them to focus more on monitoring and evaluating treatment effectiveness rather than manually adjusting the equipment. Optimize treatment results: Continuous and precise laser irradiation ensures consistent and repeatable treatment effects, helping to achieve better treatment results, especially when treating complex or extended lesions.

[0024] 3. This invention aims to reduce the need for manual identification through automated feature extraction using a convolutional neural network (CNN). Improving diagnostic accuracy and consistency: CNNs can accurately identify and extract key features from complex skin images, reducing errors caused by subjectivity and experience differences in manual diagnosis, and improving treatment accuracy and consistency. Accelerating processing speed: Automated feature extraction significantly increases data processing speed, enabling rapid analysis and processing of large numbers of images, thereby accelerating diagnostic and treatment decision-making. Reducing the burden on medical staff: The workload of doctors and technicians on image recognition and data analysis is reduced, allowing them to devote more time and energy to optimizing treatment strategies and patient care. Increasing treatment reliability: Automated feature extraction increases the reliability and predictability of treatment processes by providing standardized and reproducible analysis results. Optimizing resource allocation: Automated processing allows for more efficient allocation of medical resources, improving the overall efficiency of healthcare services. Improving the patient experience: Waiting times and treatment preparation time are reduced, improving the overall patient experience while also reducing the anxiety and dissatisfaction that can arise from long wait times. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] FIG1 is a schematic diagram of the operation of the overall process of the present invention;

[0026] FIG2 is a schematic diagram of the operation flow of the convolutional neural network (CNN) of the present invention;

[0027] FIG3 is a schematic diagram of the operation flow of the image segmentation algorithm (U-Net) of the present invention;

[0028] FIG4 is a schematic diagram of the process of dynamic tracking using the optical flow algorithm of the present invention. Best Mode for Carrying Out the Invention

[0029] In order to make the technical means, creative features, objectives and effects achieved by the present invention easier to understand, the present invention is further described below in conjunction with specific implementation methods.

[0030] In the description of the present invention, it should be noted that the terms "upper," "lower," "inner," "outer," "front end," "rear end," "two ends," "one end," "the other end," and the like, indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limiting the present invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0031] In the description of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "installed," "provided with," "connected," etc., should be understood in a broad sense. For example, "connected" may refer to a fixed connection, a detachable connection, or an integral connection; it may refer to a mechanical connection or an electrical connection; it may refer to a direct connection or an indirect connection through an intermediate medium; it may refer to internal communication between two components. Those skilled in the art will be able to understand the specific meanings of the above terms in the present invention in specific circumstances.

[0032] Example

[0033] Please refer to Figures 1-4, the present invention provides a technical solution:

[0034] A method and application of He-Ne laser treatment based on adaptive beam focusing, including wavefront optimization, a feedback control system, a convolutional neural network (CNN), an image segmentation algorithm (U-Net), and an optical flow algorithm. The method steps are as follows:

[0035] Step 1: Patient Assessment: Input: Images and data of the patient's treatment area. Output: Images and patient data are prepared.

[0036] Step 2: Image Processing and Analysis: Use a convolutional neural network (CNN) to extract image features and identify lesions. Use a U-Net to accurately segment the treatment area. Output: Precise location of the treatment area and confirmation of lesion type.

[0037] Step 3: Wavefront Optimization Laser Parameter Setting: Based on the image analysis results, the laser parameters (wavelength, energy density, pulse width, etc.) are adjusted through the feedback control system. Output: The laser equipment preset parameters are completed.

[0038] Step 4: Treatment Execution: Wavefront optimization technology is used to optimize the beam's focus. Optical flow algorithms are used to track the movement of the affected area in real time and adjust the beam's position. Output: The treatment process is dynamically adjusted and executed.

[0039] Step 5: Results Evaluation and Subsequent Adjustments: Based on the post-treatment images and data feedback, CNN and U-Net are used again to evaluate the results. If adjustments are required, return to Step 3, reset the laser parameters, and perform the treatment.

[0040] Step 6: Treatment completion: Patient evaluation is completed and the treatment results are satisfactory.

[0041] It should be noted that a method and application of He-Ne laser treatment based on adaptive beam focusing has the following specific implementation steps:

[0042] Step 1: Patient Assessment: This initial step involves collecting images of the treatment area and relevant health data from the patient, which will be used for subsequent image processing and treatment parameter setting. After this step, the system organizes and formats the images and data for detailed analysis.

[0043] Step 2: Image Processing and Analysis: A convolutional neural network (CNN) performs deep learning analysis on the captured images, automatically extracting key features such as lesion shape, size, and texture. Next, a U-Net algorithm is used to precisely segment the treatment area from healthy skin, ensuring that the laser only affects the target area. The output of this step is the precise location and confirmation of the lesion type, providing a basis for adjusting laser parameters.

[0044] Step 3: Wavefront Optimization Laser Parameter Settings: Based on the image analysis results, a feedback control system adjusts key laser parameters such as wavelength, energy density, and pulse width to suit the specific lesion characteristics and treatment needs. This step ensures that the laser device operates at the most optimized settings during treatment, enhancing the therapeutic effect while minimizing potential damage to surrounding healthy tissue.

[0045] Step 4: Treatment Execution: Wavefront optimization technology is used to adjust the beam's focus, while optical flow algorithms track minute movements of the affected area in real time, ensuring precise laser focus on the dynamically changing treatment area. During this process, the laser's position and parameters are continuously adjusted based on real-time data to accommodate any changes in the patient's position during treatment.

[0046] Step 5: Results Evaluation and Subsequent Adjustments: After treatment is complete, the CNN and U-Net algorithms are used again to evaluate the treatment's effectiveness. Based on subsequent image and data feedback, further parameter adjustments or additional treatments are determined. This step ensures continuity and adaptability of treatment, improving the success rate.

[0047] Step 6: Treatment Completion: The final step is to confirm whether the patient is satisfied with the treatment results and complete the patient's evaluation. This step marks the end of the entire treatment process. The patient can then proceed with follow-up health management or further treatment based on the treatment results and the doctor's recommendations.

[0048] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. [Corrected 16.05.2024 in accordance with Rule 26] A method and application of He-Ne laser therapy based on adaptive beam focusing, including wavefront optimization, feedback control system, convolutional neural network (CNN), image segmentation algorithm (U-Net), optical flow algorithm, characterized in that: The method steps are as follows: Step 1: Patient Assessment: Input: images and data of the patient’s treatment area, Output: images and patient data are ready; Step 2: Image processing and analysis: Use convolutional neural network (CNN) to extract features and identify lesions from images, and use U-Net to accurately segment the treatment area. Output: accurate positioning of the treatment area and confirmation of lesion type; Step 3: Wavefront optimization laser parameter setting: According to the image analysis results, the laser parameters (wavelength, energy density, pulse width, etc.) are adjusted through the feedback control system. Output: The laser equipment preset parameters are completed; Step 4: Treatment execution: Use wavefront optimization technology to optimize the focusing effect of the beam, use optical flow algorithm to track the movement of the affected part in real time, adjust the beam position, output: dynamic adjustment and execution of the treatment process; Step 5: Result evaluation and subsequent adjustments: Based on the images and data feedback after treatment, use CNN and U-Net to evaluate the effect again. If adjustments are needed, return to step 3, reset the laser parameters and perform the treatment; Step 6: Treatment completed: Patient evaluation is completed and treatment results are satisfactory.

2. A method and application of He-Ne laser treatment based on adaptive beam focusing according to claim 1, characterized in that: The convolutional neural network (CNN) uses optimization algorithms such as ReLU function, softmax function, SGD (stochastic gradient descent), Adam, etc. The process of extracting image information is as follows: Image preprocessing: Before input into CNN, the image needs to undergo a series of preprocessing steps to improve the learning efficiency and effect of the network: Normalization: Scale the image pixel values ​​to a certain range, such as 0 to 1 or -1 to 1, to reduce numerical instability during model training, Resize: Modify the image size to meet the requirements of the network input layer, Enhancement: Including rotation, scaling, cropping, etc., to increase the generalization ability of the model, the convolution layer is responsible for extracting features from the image, using a set of learned filters (or called convolution kernel) performs convolution operation with the image to extract local features of the image. Each filter captures different features of the image, such as edges, corners or more complex patterns. ReLU (rectifier linear unit) is used as the activation function to increase the nonlinear ability of the network and help capture complex patterns. The pooling layer is after the convolution layer to reduce the spatial size of the features, reduce the number of parameters and the amount of calculation, and make the features have better spatial invariance. The maximum value in a specific area is selected as the representative of the area, and the average value in the specific area is calculated. The fully connected layer is after multiple convolution and pooling layers. The fully connected layer is used to classify or regress the extracted features. The feature map (feature map) is flattened into a one-dimensional vector, and the fully connected layer learns how to make correct predictions based on these feature vectors, classify the lesion type, and the final output layer is designed according to the task (classification or regression). In the classification task, the output layer uses the softmax function, which converts the output into a probability distribution, training and optimization: Loss function: defines how to calculate the error between the predicted value and the true value (cross entropy loss is used for classification), and uses optimization algorithms such as SGD (stochastic gradient descent) and Adam to adjust the network weights and minimize the loss function.

3. A method and application of He-Ne laser treatment based on adaptive beam focusing according to claim 1, characterized in that: The image segmentation algorithm consists of multiple convolutional layers and pooling layers, which gradually reduce the spatial dimension and increase the depth of features. These layers extract more and more complex features by learning local features in the image, restore the spatial dimension of the image, and reduce the feature depth. Through upsampling and convolution, the expansion path maps high-level features back to higher-resolution outputs for precise positioning. Using the jump connection of U-Net, it directly connects the feature map of the contraction path to the corresponding layer of the expansion path. These connections help retain the position information lost by the downsampling process during the upsampling process, allowing the network to better perform pixel-level accurate predictions. In the expansion path, the upsampling layer gradually converts the feature map of the contraction path into a higher-resolution output. The resolution is increased while refining these features through convolutional layers. Transposed convolution (sometimes called deconvolution) or simple interpolation methods are used to increase the size of the feature map. Convolutional layers are used to merge features from jump connections, refine features and restore details. The last layer is a convolutional layer that converts the final feature map into a segmentation map. A 1x1 convolution kernel is used to convert multi-channel feature maps into single-channel outputs. Each pixel in the output layer predicts the category of its corresponding input pixel. The segmentation task uses the Dice loss function for cases where the data categories are unbalanced. Backpropagation and optimization algorithms (such as Adam) adjust network parameters, minimize the loss function, and provide data references during segmentation.

4. A method and application of He-Ne laser treatment based on adaptive beam focusing according to claim 1, characterized in that: In the helium-neon laser treatment system, the feedback control system uses sensors to collect data on the treatment area, light absorption rate and scattering rate in real time, and identifies the laser parameters that need to be adjusted through data analysis and processing. The system uses a real-time monitoring algorithm to continuously evaluate the treatment effect, and uses a parameter adjustment algorithm to modify the laser power and pulse width in real time to adapt to the skin's reaction. In addition, the result feedback and iterative optimization algorithm adjusts the subsequent treatment strategy according to the treatment effect, and the safety protection algorithm ensures the safety of the patient during the entire treatment process.

5. A method and application of He-Ne laser treatment based on adaptive beam focusing according to claim 1, characterized in that: In the adaptive beam focusing He-Ne laser treatment system, the wavefront optimization technology works together with the feedback control system to ensure the accuracy and effectiveness of laser treatment. The wavefront optimization uses sensors to collect distortion data of reflected light waves in real time, calculates correction parameters after analyzing these data, and uses deformable mirrors or optical modulators to adjust the light beam in real time to optimize the focusing quality. At the same time, the feedback control system monitors the optical properties and skin response of the treatment area, and adjusts the parameters of laser power and pulse width according to the real-time data to match the laser parameters with the treatment needs and skin conditions.

6. A method and application of He-Ne laser treatment based on adaptive beam focusing according to claim 1, characterized in that: In the adaptive beam focusing He-Ne laser treatment system, the optical flow algorithm is used to track the dynamic position and movement of the treatment area in real time to ensure that the laser can be accurately adjusted to the correct treatment position. This algorithm analyzes the pixel changes between consecutive image frames before and after treatment to estimate the movement of the treatment area, calculate the movement speed and direction, and automatically adjust the position, angle, focus depth and intensity of the laser.

Images