Multi-spectral hair-fixing, hair-growing, hair-nourishing, hair-repairing and hair-regenerating therapeutic apparatus

Abstract

The invention provides a multispectral hair-fixing, hair-growing, hair-nourishing, hair-repairing and hair-regenerating therapeutic apparatus, and relates to the technical field of hair-growing therapeutic apparatuses, a multispectral illumination scheme for aid decision making is generated through a deep learning algorithm on the basis of user data, and a light source module is controlled to work according to the multispectral illumination scheme for aid decision making. A user can be assisted to simply and quickly use the therapeutic apparatus, a multispectral illumination scheme suitable for the user can be found, and the use efficiency of the therapeutic apparatus can be effectively improved.

Description

Technical Field

[0001] The present application relates to the technical field of hair growth treatment devices, and in particular to a multi-spectrum hair growth, hair repair and hair regeneration treatment device. Background Art

[0002] With the accelerated pace of life and increased work pressure, more and more people are facing the problem of hair loss and thinning hair. At present, although there are some technologies for hair growth and hair regeneration, such as drug treatment, hair transplant surgery, etc., these methods often have certain risks and side effects. Therefore, the existing technology often uses a hair growth instrument for hair growth and hair regeneration. A hair growth instrument is a device used to promote hair growth and improve hair quality. It stimulates hair follicles, promotes hair growth and prevents hair loss by using a specific light source and optical technology, thereby improving the overall hair quality. Hair growth instruments are usually used to treat various hair loss symptoms, such as male pattern baldness, female pattern baldness, circular alopecia, etc. The working principle of the hair growth instrument mainly uses light of different wavelengths, such as red, near infrared, far infrared, blue, etc., to achieve different therapeutic effects. These lights can penetrate the surface of the scalp, stimulate the activity of hair follicle cells, promote blood circulation, and thus promote hair growth.

[0003] In the prior art, hair growth devices are generally used by users themselves, which results in users being unable to accurately find a multi-spectral usage strategy that is suitable for them; or users consult professional physicians for use, but the professional physicians may not reply to messages in a timely manner, resulting in complicated and inefficient use. Summary of the invention

[0004] The embodiment of the present application provides a multi-spectrum hair growth, hair repair and hair regeneration treatment device to solve the technical problems existing in the prior art.

[0005] This application adopts the following technical solutions:

[0006] A multi-spectrum hair growth, hair repair and hair regeneration treatment device, comprising: a redundant power supply module, a data processing module, a cap module and a light source module;

[0007] The redundant power supply module is used to supply power to the data processing module and the light source module by using a mobile power supply mode and a wired power supply mode;

[0008] The data processing module is used to collect user data through wireless communication, and based on the user data, generate a multi-spectral lighting scheme for auxiliary decision-making through a deep learning algorithm, and control the light source module to work according to the multi-spectral lighting scheme for auxiliary decision-making;

[0009] The cap body module is used to form a hollow hemispherical cap body structure, and the light source module is installed on the cap body structure;

[0010] The light source module is used to accept the control of the data processing module and generate light in a multi-spectral lighting manner to achieve hair growth, hair repair and hair regeneration.

[0011] Furthermore, the data processing module includes a communication submodule, a deep learning submodule, an auxiliary decision submodule and a control submodule;

[0012] The communication submodule is used to connect to the user's device through wireless communication and receive user data input by the user through human-computer interaction;

[0013] The deep learning submodule is used to pre-deploy the deep learning algorithm and receive the scheduling of the auxiliary decision submodule;

[0014] The auxiliary decision submodule is used to generate an auxiliary decision-making multi-spectral lighting plan through a deep learning algorithm based on the user data;

[0015] The control submodule is used to control the light source module to work according to the auxiliary decision-making multi-spectral illumination scheme.

[0016] Furthermore, the user data includes gender, age, weight, average daily sleep amount, and the number of hair follicles per unit area of ​​hair loss;

[0017] The number of hair follicles per unit is data obtained by the user through professional equipment or detection by a professional physician, and the number of hair follicles per unit is used to represent the number of normal hair follicles in a unit area.

[0018] Furthermore, deep learning algorithms are pre-deployed, including:

[0019] Initialize the deep learning algorithm using a BP neural network algorithm or a CNN algorithm, and determine a plurality of different algorithm parameter particles about the deep learning algorithm; wherein each algorithm parameter particle includes all algorithm parameters to be optimized of the deep learning algorithm;

[0020] Acquire pre-stored training data; wherein the training data includes historical user data and manual labels corresponding to the historical user data; wherein the manual labels are used to characterize type labels of multi-spectral illumination schemes manually marked by professional physicians;

[0021] Based on the training data, an improved particle swarm algorithm is used to optimize multiple algorithm parameter particles to obtain the optimal algorithm parameter particles;

[0022] The parameters included in the optimal algorithm parameter particle are used as the final parameters of the deep learning algorithm, and the deep learning algorithm is deployed to complete the pre-deployment.

[0023] Furthermore, based on the training data, an improved particle swarm algorithm is used to optimize multiple algorithm parameter particles to obtain the optimal algorithm parameter particles, including:

[0024] For each algorithm parameter particle, the parameters contained in the algorithm parameter particle are applied to the deep learning algorithm, and the historical user data in the training data is used as the input of the deep learning algorithm, and the manual labels corresponding to the historical user data are used as the expected output of the deep learning algorithm to obtain the fitness corresponding to each algorithm parameter particle;

[0025] According to the fitness corresponding to each algorithm parameter particle, determine the optimal algorithm parameter particle;

[0026] Based on the optimal algorithm parameter particle, the algorithm parameter particle is nonlinearly locally optimized by using adaptive memory and mutation optimization strategies to obtain the algorithm parameter particle after nonlinear local optimization;

[0027] For the algorithm parameter particles after nonlinear local optimization, the local influence ability of each algorithm parameter particle is determined, and based on the local influence ability, a mutation strategy is used to perform global mutation optimization on the algorithm parameter particles to obtain the algorithm parameter particles after global mutation optimization;

[0028] For the algorithm parameter particles after global mutation optimization, the population information fusion strategy is used to perform local fine optimization on the algorithm parameter particles to obtain the algorithm parameter particles after local fine optimization;

[0029] Determine whether the current optimization times have reached the maximum optimization times. If so, redetermine the optimal algorithm parameter particles according to the algorithm parameter particles after local fine optimization, and output the redetermined optimal algorithm parameter particles. Otherwise, redetermine the optimal algorithm parameter particles according to the algorithm parameter particles after local fine optimization, and return to the nonlinear local optimization step.

[0030] Furthermore, based on the optimal algorithm parameter particle, the algorithm parameter particle is nonlinearly locally optimized by using adaptive memory and mutation optimization strategies to obtain the algorithm parameter particle after nonlinear local optimization, including:

[0031] Based on the current optimization times, the adaptive mutation factor is determined as: ξ=(ξ min +(ξ max -ξ min )t / T)*(1+γ*(r1-1)); where ξ represents the adaptive mutation factor, max represents the maximum value corresponding to the adaptive mutation factor, ξ minrepresents the minimum value corresponding to the adaptive mutation factor, t represents the current optimization times, T represents the maximum optimization times, γ represents the mutation probability, and r1 represents the first random number between [0,1];

[0032] Based on the adaptive mutation factor, the speed of the algorithm parameter particle is adaptively memorized, and the adaptive memory result is: V i t-1 *exp(-R*t); where V i t-1 represents the speed of the i-th algorithm parameter particle in the t-1 optimization process, i = 1, 2, ..., I, I represents the total number of algorithm parameter particles;

[0033] According to the optimal algorithm parameter particle, the guide item corresponding to the algorithm parameter particle is determined as: in, represents the optimal algorithm parameter particle, represents the i-th algorithm parameter particle in the t-1 optimization process;

[0034] According to the adaptive memory result and the guide item corresponding to the algorithm parameter particle, the speed of the i-th algorithm parameter particle in the t-th optimization process is determined as:

[0035] According to the speed of the ith algorithm parameter particle in the tth optimization process, a nonlinear local optimization is performed on the ith algorithm parameter particle, and the algorithm parameter particle after the nonlinear local optimization is obtained as follows: in, Represents the algorithm parameter particle after the i-th nonlinear local optimization.

[0036] Furthermore, for the algorithm parameter particles after nonlinear local optimization, the local influence ability of each algorithm parameter particle is determined, and based on the local influence ability, a mutation strategy is used to perform global mutation optimization on the algorithm parameter particles to obtain the algorithm parameter particles after global mutation optimization, including:

[0037] For any algorithm parameter particle after nonlinear local optimization, the distance influencing parameter between the algorithm parameter particle and other different algorithm parameter particles is determined as: in, It represents the distance influence parameter between the algorithm parameter particle after the mth nonlinear local optimization and the algorithm parameter particle after the jth nonlinear local optimization, It represents the historical optimal value of the algorithm parameter particle after the mth nonlinear local optimization. It represents the historical optimal value of the algorithm parameter particle after the jth nonlinear local optimization, and m is not equal to j; The best historical The d-th dimension parameter in , Represents the historical best value The d-th dimension parameter in, D represents the total dimension of the parameter, m = 1, 2, ..., I, I represents the total number of algorithm parameter particles;

[0038] Based on the distance influence parameter between the algorithm parameter particle and other particles with different algorithm parameters, the local influence ability of each algorithm parameter particle is determined as: and in, It represents the local influence ability of the algorithm parameter particle after the mth nonlinear local optimization. represents the neighborhood particle determination factor between the algorithm parameter particle after the mth nonlinear local optimization and the algorithm parameter particle after the jth nonlinear local optimization, and R represents the neighborhood radius;

[0039] According to the local influence capability corresponding to each algorithm parameter particle, the algorithm parameter particles after nonlinear local optimization are arranged in descending order of local influence capability, and the historical optimal values ​​corresponding to the first half of the algorithm parameter particles are copied to obtain the copied particles;

[0040] For any replica particle, determine Is it true? If so, the replica particle is mutated to obtain the replica particle after the mutation process. Otherwise, the replica particle is not mutated.

[0041] Among them, exp represents the exponential function with the natural constant e as the base, Indicates the fitness of the particle corresponding to the optimal algorithm parameter, represents the fitness of the replicating particle; η represents the probability of mutation decision;

[0042] The replica particles are mutated and the replica particles after mutation are: Where r2 represents the second random number between [0,1], Represents the replica particles after mutation processing;

[0043] According to the replicated particles after mutation processing, the algorithm parameter particles in the second half of the arrangement are updated to obtain the algorithm parameter particles after global mutation optimization.

[0044] Furthermore, the mutation decision probability is: in, represents the fitness of the fzth replica, and K represents the total number of replicas.

[0045] Furthermore, for the algorithm parameter particles after global mutation optimization, the population information fusion strategy is used to perform local fine optimization on the algorithm parameter particles, and the algorithm parameter particles after local fine optimization are obtained, including:

[0046] Based on the current number of optimizations, the adaptive quantity factor is determined as: N t =INT(βN t-1 ), where N t Represents the adaptive quantity factor during the t-th training process, N t-1 represents the adaptive quantity factor in the t-1th training process, β represents the adaptive coefficient, and β=αcos(πt / T), INT represents the upward rounding function, α represents the attenuation factor between (0,1), cos represents the cosine function, π represents pi, and T represents the maximum number of optimization times;

[0047] For the algorithm parameter particles after global mutation optimization, re-obtain the fitness corresponding to the algorithm parameter particles, and according to the re-obtained fitness, arrange the algorithm parameter particles after global mutation optimization in order from large to small fitness, and take out the top N t algorithm parameter particles, and obtain the target algorithm parameter particles;

[0048] The population information fusion strategy is used to perform local fine optimization on the target algorithm parameter particles. The target algorithm parameter particles after local fine optimization are: in, represents the nth target algorithm parameter particle, r3 represents the third random number between [0,1], and r4 represents the fourth random number between [0,1]. Represents the target algorithm parameter particle after local fine optimization Represents the population center corresponding to all algorithm parameter particles, Represents the sub-population center corresponding to all target algorithm parameter particles.

[0049] Further, controlling the light source module to work according to the auxiliary decision-making multi-spectral illumination scheme includes:

[0050] The multi-spectral lighting scheme for assisting decision-making is transmitted to the user's device to receive a scheme execution instruction fed back by the user; wherein the scheme execution instruction includes executing the multi-spectral lighting scheme for assisting decision-making or not executing the multi-spectral lighting scheme for assisting decision-making;

[0051] When the scheme execution instruction is to execute the multi-spectral lighting scheme for auxiliary decision-making, the multi-spectral lighting scheme for auxiliary decision-making is executed to control the light source module to work.

[0052] The present application provides a multi-spectral hair growth, hair maintenance and hair regeneration therapeutic device, which generates a decision-making auxiliary multi-spectral lighting plan based on the user data through a deep learning algorithm, and controls the light source module to work according to the decision-making auxiliary multi-spectral lighting plan. This can assist the user to use the therapeutic device simply and quickly, find a multi-spectral lighting plan suitable for themselves, and effectively improve the use efficiency of the therapeutic device. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a structural block diagram of a multi-spectrum hair growth, hair repair and hair regeneration therapeutic device provided in an embodiment of the present application.

[0054] Figure 2 This is a first structural example diagram of a multi-spectrum hair growth, hair repair and hair regeneration therapeutic device provided in an embodiment of the present application.

[0055] Figure 3 This is a second structural example diagram of a multi-spectrum hair growth, hair repair and hair regeneration therapeutic device provided in an embodiment of the present application.

[0056] Among them, 1 is a cap body, 2 is a data processing device, 3 is a power button, 4 is a first power cord, 5 is a mobile power supply, 6 is a second power cord, 7 is a third power cord, and 8 is a multi-spectrum light strip. DETAILED DESCRIPTION

[0057] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0058] like Figure 1 As shown, the present application provides a multi-spectrum hair growth, hair repair and hair regeneration treatment device, including: a redundant power supply module, a data processing module, a cap module and a light source module;

[0059] The redundant power supply module is used to supply power to the data processing module and the light source module by using a mobile power supply mode and a wired power supply mode;

[0060] By setting up a redundant power supply module, the embodiment of the present application allows users to choose whether to use mobile power supply or wired fixed power supply, which can greatly improve the convenience of using the therapeutic device.

[0061] The data processing module is used to collect user data through wireless communication, and based on the user data, generate a multi-spectral lighting scheme for auxiliary decision-making through a deep learning algorithm, and control the light source module to work according to the multi-spectral lighting scheme for auxiliary decision-making;

[0062] For example, the data processing module can be connected to the user's device through wireless communication to adopt the user data input by the user through the device. Based on the user data, a deep learning algorithm is used to perform data analysis and recommendation, so that a multi-spectral lighting solution for auxiliary decision-making can be obtained. After the multi-spectral lighting solution for auxiliary decision-making is generated, it can also be transmitted back to the user's device so that the user can determine whether to adopt the multi-spectral lighting solution for auxiliary decision-making.

[0063] Before using the deep learning algorithm, historical user data can also be collected, and professional physicians can determine the type of multispectral illumination scheme that should be used for the historical user data, and store the type of multispectral illumination scheme and the multispectral illumination scheme in association. Then, after learning the historical user data and the corresponding type of multispectral illumination scheme through the deep learning algorithm, a recommendation task for the multispectral illumination scheme can be formed. The deep learning algorithm can be used for recommendation in the future to replace the work of professional physicians, improve the efficiency of users' use of the multispectral hair growth, hair repair and hair regeneration treatment device, and use the multispectral illumination scheme that suits them.

[0064] The cap body module is used to form a hollow hemispherical cap body structure, and the light source module is installed on the cap body structure;

[0065] By providing the cap body module, it is convenient to install the light source module and to use the multi-spectrum hair growth, hair repair and hair regeneration treatment device.

[0066] The light source module is used to accept the control of the data processing module and generate light in a multi-spectral lighting manner to achieve hair growth, hair repair and hair regeneration.

[0067] The light source module can be composed of lamp beads with different wavelengths to generate light of different wavelengths such as red, near infrared, far infrared, blue, etc., to achieve hair growth, hair repair and hair regeneration.

[0068] The present application provides a multi-spectral hair growth, hair maintenance and hair regeneration therapeutic device, which generates a decision-making auxiliary multi-spectral lighting plan based on the user data through a deep learning algorithm, and controls the light source module to work according to the decision-making auxiliary multi-spectral lighting plan. This can assist the user to use the therapeutic device simply and quickly, find a multi-spectral lighting plan suitable for themselves, and effectively improve the use efficiency of the therapeutic device.

[0069] like Figure 2 - Figure 3 As shown, in order for those skilled in the art to better understand the embodiments of the present application, a structural example is given below, which may include a cap body 1, a data processing device 2, a power button 3, a first power cord 4, a mobile power supply 5, a second power cord 6, a third power cord 7 and a multi-spectrum light strip 8;

[0070] The multi-spectrum light strip 8 may be composed of lamp beads with different wavelengths, and the multi-spectrum light strip 8 may be evenly arranged inside the cap body 1 for user convenience. The mobile power supply 5 is connected to the power button 3 and the data processing device 2 through the first power line 4, and the power button 3 controls the switch state of the data processing device 2, and then connected to the multi-spectrum light strip 8 through the third power line 7 to control the multi-spectrum light strip 8 through the data processing device 2. Similarly, the second power line 6 can also power the data processing device 2 and the multi-spectrum light strip 8 in the same way to form a redundant power supply.

[0071] For example, the first power cord 4 and the second power cord 6 can be first connected to the power button 3. When the power button 3 is pressed, they are connected to the data processing device 2 and the third power cord 7, so that the data processing device 2 starts working. Then, the third power cord 7 and the multi-spectrum light strip 8 can be turned on and off by control methods such as optocouplers, transistors or relays, and the lamp beads of each wavelength are controlled separately, thereby forming multi-spectrum illumination.

[0072] Optionally, the first power cord 4, the mobile power supply 5 and the second power cord 6 can be used together as a redundant power supply module, the data processing device 2 can be used as a data processing module, the cap body 1 can be used as a cap body module, and the multi-spectrum light strip 8 can be used as a light source module.

[0073] In a possible implementation, the data processing module includes a communication submodule, a deep learning submodule, an auxiliary decision submodule and a control submodule;

[0074] The communication submodule is used to connect to the user's device through wireless communication and receive user data input by the user through human-computer interaction;

[0075] The deep learning submodule is used to pre-deploy the deep learning algorithm and receive the scheduling of the auxiliary decision submodule;

[0076] The auxiliary decision submodule is used to generate an auxiliary decision-making multi-spectral lighting plan through a deep learning algorithm based on the user data;

[0077] The control submodule is used to control the light source module to work according to the auxiliary decision-making multi-spectral illumination scheme.

[0078] Based on the above examples, the communication submodule, deep learning submodule, auxiliary decision submodule and control submodule may be software deployed in the data processing device 2. The data processing device 2 may be a device with data processing and wireless data communication (such as Bluetooth).

[0079] In a possible implementation, the user data includes gender, age, weight, average daily sleep volume, and the number of hair follicles per unit area of ​​hair loss;

[0080] The number of hair follicles per unit is the data obtained by the user through professional equipment or professional doctors, and the number of hair follicles per unit is used to represent the number of normal hair follicles in a unit area. The average daily sleep volume refers to the average daily sleep volume in the past 15 days, 30 days or other time periods.

[0081] It is worth mentioning that in addition to the above data, other user data can also be collected, such as heart rate and blood oxygen, etc.

[0082] In a possible implementation, pre-deploying a deep learning algorithm includes:

[0083] Initialize the deep learning algorithm using a BP (Back Propagation) neural network algorithm or a CNN (Convolutional Neural Network) algorithm, and determine a plurality of different algorithm parameter particles about the deep learning algorithm; wherein each algorithm parameter particle includes all algorithm parameters to be optimized of the deep learning algorithm;

[0084] The algorithm parameters to be optimized can be part of the parameters or all of the parameters of the deep learning algorithm. In the embodiment of the present application, the BP neural network algorithm is preferably used for data learning and analysis, so the algorithm parameters to be optimized can be all or part of the weights of the BP neural network algorithm. However, it is worth noting that the above two deep learning algorithms are only used as an example, and other deep learning algorithms can also be used for data learning and analysis.

[0085] Acquire pre-stored training data; wherein the training data includes historical user data and manual labels corresponding to the historical user data; wherein the manual labels are used to characterize type labels of multi-spectral illumination schemes manually marked by professional physicians;

[0086] Professional doctors can label historical user data and then learn the data relationship to form a recommendation task. After that, users only need to input their data to get a multi-spectral lighting plan suitable for them. This not only improves the efficiency of use, but also reduces the workload of doctors.

[0087] A multi-spectral lighting regimen may include a sequence of different wavelengths of light exposure and the duration of each wavelength of light exposure.

[0088] Based on the training data, an improved particle swarm algorithm is used to optimize multiple algorithm parameter particles to obtain the optimal algorithm parameter particles;

[0089] Particle Swarm Optimization (PSO) algorithm is a computational method that simulates the social behavior of bird flocks, fish schools and human society to find the optimal solution in multidimensional space. In PSO, each potential solution is regarded as a particle in the search space. All particles have a fitness value determined by the objective function, and a speed determines the direction and distance they move in the search space. Particles adjust their speed and position based on their own experience and the experience of neighboring particles. Although the particle swarm algorithm has the characteristics of fast convergence speed, the particle swarm algorithm is extremely easy to fall into the local optimum, which ultimately leads to inaccurate data analysis and makes it impossible for users to use the therapeutic device correctly. Therefore, the embodiment of the present application provides an improved particle swarm algorithm to solve the defects existing in the existing particle swarm algorithm.

[0090] The parameters included in the optimal algorithm parameter particle are used as the final parameters of the deep learning algorithm, and the deep learning algorithm is deployed to complete the pre-deployment.

[0091] In a possible implementation, based on the training data, an improved particle swarm algorithm is used to optimize multiple algorithm parameter particles to obtain the optimal algorithm parameter particle, including:

[0092] For each algorithm parameter particle, the parameters contained in the algorithm parameter particle are applied to the deep learning algorithm, and the historical user data in the training data is used as the input of the deep learning algorithm, and the manual labels corresponding to the historical user data are used as the expected output of the deep learning algorithm to obtain the fitness corresponding to each algorithm parameter particle;

[0093] For example, the root mean square error function or the cross entropy error function can be used to obtain the error value corresponding to each algorithm parameter particle, and then the error value is added to a very small constant (such as 0.0001), and then the inverse is taken to obtain the fitness corresponding to the algorithm parameter particle.

[0094] According to the fitness corresponding to each algorithm parameter particle, the optimal algorithm parameter particle is determined, that is, the algorithm parameter particle with the largest fitness is determined as the optimal algorithm parameter particle.

[0095] Based on the optimal algorithm parameter particle, the algorithm parameter particle is nonlinearly locally optimized by using adaptive memory and mutation optimization strategies to obtain the algorithm parameter particle after nonlinear local optimization;

[0096] For the algorithm parameter particles after nonlinear local optimization, the local influence ability of each algorithm parameter particle is determined, and based on the local influence ability, a mutation strategy is used to perform global mutation optimization on the algorithm parameter particles to obtain the algorithm parameter particles after global mutation optimization;

[0097] For the algorithm parameter particles after global mutation optimization, the population information fusion strategy is used to perform local fine optimization on the algorithm parameter particles to obtain the algorithm parameter particles after local fine optimization;

[0098] Determine whether the current optimization times have reached the maximum optimization times. If so, redetermine the optimal algorithm parameter particles according to the algorithm parameter particles after local fine optimization, and output the redetermined optimal algorithm parameter particles. Otherwise, redetermine the optimal algorithm parameter particles according to the algorithm parameter particles after local fine optimization, and return to the nonlinear local optimization step.

[0099] In a possible implementation, based on the optimal algorithm parameter particle, an adaptive memory and mutation optimization strategy is used to perform nonlinear local optimization on the algorithm parameter particle to obtain the algorithm parameter particle after nonlinear local optimization, including:

[0100] Based on the current optimization times, the adaptive mutation factor is determined as: ξ=(ξ min +(ξ max -ξ min )t / T)*(1+γ*(r1-1)); where ξ represents the adaptive mutation factor, max represents the maximum value corresponding to the adaptive mutation factor, ξ min represents the minimum value corresponding to the adaptive mutation factor, t represents the current optimization times, T represents the maximum optimization times, γ represents the mutation probability, and r1 represents the first random number between [0,1];

[0101] Based on the adaptive mutation factor, the speed of the algorithm parameter particle is adaptively memorized, and the adaptive memory result is: V i t-1 *exp(-R*t); where V i t-1 represents the speed of the i-th algorithm parameter particle in the t-1 optimization process, i = 1, 2, ..., I, I represents the total number of algorithm parameter particles;

[0102] According to the optimal algorithm parameter particle, the guide item corresponding to the algorithm parameter particle is determined as: in, represents the optimal algorithm parameter particle, represents the i-th algorithm parameter particle in the t-1 optimization process;

[0103] According to the adaptive memory result and the guide item corresponding to the algorithm parameter particle, the speed of the i-th algorithm parameter particle in the t-th optimization process is determined as:

[0104] According to the speed of the ith algorithm parameter particle in the tth optimization process, a nonlinear local optimization is performed on the ith algorithm parameter particle, and the algorithm parameter particle after the nonlinear local optimization is obtained as follows: in, Represents the algorithm parameter particle after the i-th nonlinear local optimization.

[0105] The adaptive memory and mutation optimization strategy provided in the embodiment of the present application can adaptively change and mutate the speed while moving in the optimal direction, which is more conducive to exploring unknown areas and improving the possibility of exploring the global optimal solution.

[0106] In a possible implementation, for the algorithm parameter particles after nonlinear local optimization, the local influence capability of each algorithm parameter particle is determined, and based on the local influence capability, a mutation strategy is used to perform global mutation optimization on the algorithm parameter particles to obtain the algorithm parameter particles after global mutation optimization, including:

[0107] For any algorithm parameter particle after nonlinear local optimization, the distance influencing parameter between the algorithm parameter particle and other different algorithm parameter particles is determined as: in, It represents the distance influence parameter between the algorithm parameter particle after the mth nonlinear local optimization and the algorithm parameter particle after the jth nonlinear local optimization, It represents the historical optimal value of the algorithm parameter particle after the mth nonlinear local optimization. It represents the historical optimal value of the algorithm parameter particle after the jth nonlinear local optimization, and m is not equal to j; The best historical The d-th dimension parameter in , Represents the historical best value The d-th dimension parameter in, D represents the total dimension of the parameter, m = 1, 2, ..., I, I represents the total number of algorithm parameter particles;

[0108] Based on the distance influence parameter between the algorithm parameter particle and other particles with different algorithm parameters, the local influence ability of each algorithm parameter particle is determined as: and in, It represents the local influence ability of the algorithm parameter particle after the mth nonlinear local optimization. represents the neighborhood particle determination factor between the algorithm parameter particle after the mth nonlinear local optimization and the algorithm parameter particle after the jth nonlinear local optimization, and R represents the neighborhood radius;

[0109] According to the local influence capability corresponding to each algorithm parameter particle, the algorithm parameter particles after nonlinear local optimization are arranged in descending order of local influence capability, and the historical optimal values ​​corresponding to the first half of the algorithm parameter particles are copied to obtain the copied particles;

[0110] For any replica particle, determine Is it true? If so, the replica particle is mutated to obtain the replica particle after the mutation process. Otherwise, the replica particle is not mutated.

[0111] Among them, exp represents the exponential function with the natural constant e as the base, Indicates the fitness of the particle corresponding to the optimal algorithm parameter, represents the fitness of the replicating particle; η represents the probability of mutation decision;

[0112] The replica particles are mutated and the replica particles after mutation are: Where r2 represents the second random number between [0,1], Represents the replica particles after mutation processing;

[0113] According to the replicated particles after mutation processing, the algorithm parameter particles in the second half of the arrangement are updated to obtain the algorithm parameter particles after global mutation optimization.

[0114] The mutation strategy provided in the embodiment of the present application can search in a non-optimal direction for some particles with poor fitness, so that the algorithm always maintains the ability of global search to avoid falling into the local optimum.

[0115] In a possible implementation, the variation decision probability is: in, represents the fitness of the fzth replica, and K represents the total number of replicas.

[0116] In a possible implementation, for the algorithm parameter particles after global variation optimization, a population information fusion strategy is used to perform local fine optimization on the algorithm parameter particles to obtain the algorithm parameter particles after local fine optimization, including:

[0117] Based on the current number of optimizations, the adaptive quantity factor is determined as: Nt =INT(βN t-1 ), where N t Represents the adaptive quantity factor during the t-th training process, N t-1 represents the adaptive quantity factor in the t-1th training process, β represents the adaptive coefficient, and β=αcos(πt / T), INT represents the upward rounding function, α represents the attenuation factor between (0,1), cos represents the cosine function, π represents pi, and T represents the maximum number of optimization times;

[0118] For the algorithm parameter particles after global mutation optimization, re-obtain the fitness corresponding to the algorithm parameter particles, and according to the re-obtained fitness, arrange the algorithm parameter particles after global mutation optimization in order from large to small fitness, and take out the top N t algorithm parameter particles, and obtain the target algorithm parameter particles;

[0119] The population information fusion strategy is used to perform local fine optimization on the target algorithm parameter particles. The target algorithm parameter particles after local fine optimization are: in, represents the nth target algorithm parameter particle, r3 represents the third random number between [0,1], and r4 represents the fourth random number between [0,1]. Represents the target algorithm parameter particle after local fine optimization Represents the population center corresponding to all algorithm parameter particles, Represents the sub-population center corresponding to all target algorithm parameter particles.

[0120] The population information fusion strategy provided in the embodiment of the present application can enable the algorithm to fuse population information, explore more unknown local areas in the early stage of the algorithm, and improve the convergence accuracy in the later stage of the algorithm. It can effectively improve the convergence speed and convergence effect of the algorithm, and always search towards the better area. At the same time, some particles with poor positions are not searched, which improves the speed and the global search capability at the same time.

[0121] In a possible implementation, controlling the light source module to work according to the decision-assisted multi-spectral illumination scheme includes:

[0122] The multi-spectral lighting scheme for assisting decision-making is transmitted to the user's device to receive a scheme execution instruction fed back by the user; wherein the scheme execution instruction includes executing the multi-spectral lighting scheme for assisting decision-making or not executing the multi-spectral lighting scheme for assisting decision-making;

[0123] When the scheme execution instruction is to execute the multi-spectral lighting scheme for auxiliary decision-making, the multi-spectral lighting scheme for auxiliary decision-making is executed to control the light source module to work.

[0124] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or other arbitrary combinations. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When a computer instruction or computer program is loaded or executed on a computer, a process or function according to an embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable systems. The computer program or instruction can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer program or instruction can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center containing one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD) or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0125] In this application, "at least one" means one or more, and "more" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can be represented by: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or multiple.

[0126] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0127] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0128] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0129] In the several embodiments provided in the present application, it should be understood that the disclosed systems, systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the system or unit can be electrical, mechanical or other forms.

[0130] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0131] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0132] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0133] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A multi-spectrum hair growth, hair repair and hair regeneration treatment device, characterized in that: include: Redundant power supply module, data processing module, cap module and light source module; The redundant power supply module is used to power the data processing module and the light source module using a mobile power supply mode and a wired power supply mode; The data processing module is used to collect user data through wireless communication, and based on the user data, generate a multi-spectral lighting scheme for auxiliary decision-making through a deep learning algorithm, and control the light source module to work according to the multi-spectral lighting scheme for auxiliary decision-making; The cap body module is used to form a hollow hemispherical cap body structure, and the light source module is installed on the cap body structure; The light source module is used to accept the control of the data processing module and generate light in a multi-spectral light mode to achieve hair growth, hair repair and hair regeneration; The data processing module includes a communication submodule, a deep learning submodule, an auxiliary decision submodule and a control submodule; The communication submodule is used to connect to the user's device through wireless communication and receive user data input by the user through human-computer interaction; The deep learning submodule is used to pre-deploy the deep learning algorithm and receive the scheduling of the auxiliary decision submodule; The auxiliary decision submodule is used to generate an auxiliary decision-making multi-spectral lighting plan through a deep learning algorithm based on the user data; The control submodule is used to control the light source module to work according to the auxiliary decision-making multi-spectral illumination scheme.

2. The multi-spectrum hair growth, hair repair and hair regeneration treatment device according to claim 1, characterized in that: The user data includes gender, age, weight, average daily sleep amount, and number of hair follicles per unit area of ​​hair loss; The number of hair follicles per unit is data obtained by the user through professional equipment or detection by a professional physician, and the number of hair follicles per unit is used to represent the number of normal hair follicles in a unit area.

3. The multi-spectrum hair growth, hair repair and hair regeneration treatment device according to claim 1, characterized in that: Pre-deployed deep learning algorithms, including: Initialize the deep learning algorithm using a BP neural network algorithm or a CNN algorithm, and determine a plurality of different algorithm parameter particles about the deep learning algorithm; wherein each algorithm parameter particle includes all algorithm parameters to be optimized of the deep learning algorithm; Acquire pre-stored training data; wherein the training data includes historical user data and manual labels corresponding to the historical user data; wherein the manual labels are used to characterize type labels of multi-spectral illumination schemes manually marked by professional physicians; Based on the training data, an improved particle swarm algorithm is used to optimize multiple algorithm parameter particles to obtain the optimal algorithm parameter particles; The parameters included in the optimal algorithm parameter particle are used as the final parameters of the deep learning algorithm, and the deep learning algorithm is deployed to complete the pre-deployment.

4. The multi-spectrum hair growth, hair repair and hair regeneration treatment device according to claim 3, characterized in that: Based on the training data, an improved particle swarm algorithm is used to optimize multiple algorithm parameter particles to obtain the optimal algorithm parameter particles, including: For each algorithm parameter particle, the parameters contained in the algorithm parameter particle are applied to the deep learning algorithm, and the historical user data in the training data is used as the input of the deep learning algorithm, and the manual labels corresponding to the historical user data are used as the expected output of the deep learning algorithm to obtain the fitness corresponding to each algorithm parameter particle; According to the fitness corresponding to each algorithm parameter particle, determine the optimal algorithm parameter particle; Based on the optimal algorithm parameter particle, the algorithm parameter particle is nonlinearly locally optimized by using adaptive memory and mutation optimization strategies to obtain the algorithm parameter particle after nonlinear local optimization; For the algorithm parameter particles after nonlinear local optimization, the local influence ability of each algorithm parameter particle is determined, and based on the local influence ability, a mutation strategy is used to perform global mutation optimization on the algorithm parameter particles to obtain the algorithm parameter particles after global mutation optimization; For the algorithm parameter particles after global mutation optimization, the population information fusion strategy is used to perform local fine optimization on the algorithm parameter particles to obtain the algorithm parameter particles after local fine optimization; Determine whether the current optimization times have reached the maximum optimization times. If so, redetermine the optimal algorithm parameter particles according to the algorithm parameter particles after local fine optimization, and output the redetermined optimal algorithm parameter particles. Otherwise, redetermine the optimal algorithm parameter particles according to the algorithm parameter particles after local fine optimization, and return to the nonlinear local optimization step.

5. The multi-spectral hair growth, hair repair and hair regeneration treatment device according to claim 4, characterized in that: Based on the optimal algorithm parameter particle, the algorithm parameter particle is nonlinearly locally optimized by using adaptive memory and mutation optimization strategies to obtain the algorithm parameter particle after nonlinear local optimization, including: Based on the current optimization times, the adaptive mutation factor is determined as: ξ=(ξ min +(ξ max -ξ min )t / T)*(1+γ*(r1-1)); where ξ represents the adaptive mutation factor, max represents the maximum value corresponding to the adaptive mutation factor, ξ min represents the minimum value corresponding to the adaptive mutation factor, t represents the current optimization times, T represents the maximum optimization times, γ represents the mutation probability, and r1 represents the first random number between [0,1]; Based on the adaptive mutation factor, the speed of the algorithm parameter particle is adaptively memorized, and the adaptive memory result is: V i t-1 *exp(-R*t); where V i t-1 represents the speed of the i-th algorithm parameter particle in the t-1 optimization process, i = 1, 2, ..., I, I represents the total number of algorithm parameter particles; According to the optimal algorithm parameter particle, the guide item corresponding to the algorithm parameter particle is determined as: in, represents the optimal algorithm parameter particle, represents the i-th algorithm parameter particle in the t-1 optimization process; According to the adaptive memory result and the guide item corresponding to the algorithm parameter particle, the speed of the i-th algorithm parameter particle in the t-th optimization process is determined as: According to the speed of the ith algorithm parameter particle in the tth optimization process, a nonlinear local optimization is performed on the ith algorithm parameter particle, and the algorithm parameter particle after the nonlinear local optimization is obtained as follows: in, Represents the algorithm parameter particle after the i-th nonlinear local optimization.

6. The multi-spectrum hair growth, hair repair and hair regeneration treatment device according to claim 5, characterized in that: For the algorithm parameter particles after nonlinear local optimization, the local influence ability of each algorithm parameter particle is determined, and based on the local influence ability, a mutation strategy is used to perform global mutation optimization on the algorithm parameter particles to obtain the algorithm parameter particles after global mutation optimization, including: For any algorithm parameter particle after nonlinear local optimization, the distance influencing parameter between the algorithm parameter particle and other different algorithm parameter particles is determined as: in, It represents the distance influence parameter between the algorithm parameter particle after the mth nonlinear local optimization and the algorithm parameter particle after the jth nonlinear local optimization, It represents the historical optimal value of the algorithm parameter particle after the mth nonlinear local optimization. It represents the historical optimal value of the algorithm parameter particle after the jth nonlinear local optimization, and m is not equal to j; The best historical The d-th dimension parameter in , Represents the historical best value The d-th dimension parameter in, D represents the total dimension of the parameter, m = 1, 2, ..., I, I represents the total number of algorithm parameter particles; Based on the distance influence parameter between the algorithm parameter particle and other different algorithm parameter particles, the local influence capability of each algorithm parameter particle is determined as: and in, It represents the local influence ability of the algorithm parameter particle after the mth nonlinear local optimization. represents the neighborhood particle determination factor between the algorithm parameter particle after the mth nonlinear local optimization and the algorithm parameter particle after the jth nonlinear local optimization, and R represents the neighborhood radius; According to the local influence capability corresponding to each algorithm parameter particle, the algorithm parameter particles after nonlinear local optimization are arranged in descending order of local influence capability, and the historical optimal values ​​corresponding to the first half of the algorithm parameter particles are copied to obtain the copied particles; For any replica particle, determine Is it true? If so, the replica particle is mutated to obtain the replica particle after the mutation process. Otherwise, the replica particle is not mutated. Among them, exp represents the exponential function with the natural constant e as the base, Indicates the fitness of the particle corresponding to the optimal algorithm parameter, represents the fitness of the replicating particle; η represents the probability of mutation decision; The replica particles are mutated and the replica particles after mutation are: Where r2 represents the second random number between [0,1], Represents the replica particles after mutation processing; According to the replicated particles after mutation processing, the algorithm parameter particles in the second half of the arrangement are updated to obtain the algorithm parameter particles after global mutation optimization.

7. The multi-spectral hair growth, hair repair and hair regeneration treatment device according to claim 6, characterized in that: The mutation decision probability is: in, represents the fitness of the fzth replica, and K represents the total number of replicas.

8. The multi-spectrum hair growth, hair repair and hair regeneration treatment device according to claim 7, characterized in that: For the algorithm parameter particles after global mutation optimization, the population information fusion strategy is used to perform local fine optimization on the algorithm parameter particles, and the algorithm parameter particles after local fine optimization are obtained, including: Based on the current number of optimizations, the adaptive quantity factor is determined as: N t =INT(βN t-1 ), where N t Represents the adaptive quantity factor during the t-th training process, N t-1 represents the adaptive quantity factor in the t-1th training process, β represents the adaptive coefficient, and β=αcos(πt / T), INT represents the upward rounding function, α represents the attenuation factor between (0,1), cos represents the cosine function, π represents pi, and T represents the maximum number of optimization times; For the algorithm parameter particles after global mutation optimization, re-obtain the fitness corresponding to the algorithm parameter particles, and according to the re-obtained fitness, arrange the algorithm parameter particles after global mutation optimization in order from large to small fitness, and take out the top N t algorithm parameter particles, and obtain the target algorithm parameter particles; The population information fusion strategy is used to perform local fine optimization on the target algorithm parameter particles. The target algorithm parameter particles after local fine optimization are: in, represents the nth target algorithm parameter particle, r3 represents the third random number between [0,1], and r4 represents the fourth random number between [0,1]. Represents the target algorithm parameter particle after local fine optimization Represents the population center corresponding to all algorithm parameter particles, Indicates the sub-population center corresponding to all target algorithm parameter particles.

9. The multi-spectrum hair growth, hair repair and hair regeneration treatment device according to claim 8, characterized in that: Controlling the light source module to work according to the decision-assisted multi-spectral illumination scheme includes: The multi-spectral lighting scheme for assisting decision-making is transmitted to the user's device to receive a scheme execution instruction fed back by the user; wherein the scheme execution instruction includes executing the multi-spectral lighting scheme for assisting decision-making or not executing the multi-spectral lighting scheme for assisting decision-making; When the scheme execution instruction is to execute the multi-spectral lighting scheme for auxiliary decision-making, the multi-spectral lighting scheme for auxiliary decision-making is executed to control the light source module to work.

Images