Intelligent recommended herbal extraction method and equipment

Abstract

The invention provides an intelligent recommendation herbal extraction method and device, and belongs to the technical field of intelligent recommendations, and the method comprises the steps: collecting a face image and a scalp image of a user, and carrying out the image processing and recognition, and obtaining the face representation information and scalp representation information of the user; receiving questionnaire survey information related to health input by a user on the interactive interface of the extraction device, and performing information analysis to obtain current basic information; preliminarily estimating the current health condition of the user, and extracting a plurality of health strengthening indexes from the current health condition; and determining the conditioning direction of the user depending on all health strengthening indexes, intelligently matching the conditioning direction with the conditioning effect of the traditional Chinese medicine herbs in the extraction device and multiple extraction modes of the traditional Chinese medicine herbs, and outputting a recommendation mode for the user to select. The medicine needing to be extracted can conveniently meet the actual pathological requirements of users, so that the high efficiency of medicine extraction and the subsequent effective use of the medicine are ensured to the maximum extent.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent recommendation, and particularly to an intelligent recommendation method and device for herbal extraction. Background Art

[0002] Traditional Chinese medicine preparations are drugs made from traditional Chinese medicine as raw materials under the guidance of traditional Chinese medicine theory, processed into certain specifications for direct use in preventing and treating diseases. Since the traditional pharmaceutical method is manual extraction, there will be waste of traditional Chinese medicine herbs and reduction of drug efficacy due to unsatisfactory drug extraction during this process. Or during the extraction process, general intelligent extraction devices provide several known drug juices for users to choose, without determining the actual needs of users from the users themselves, further reducing the ineffective use of drugs.

[0003] Therefore, the present invention proposes an intelligent recommendation method and device for herbal extraction. Summary of the Invention

[0004] The present invention provides an intelligent recommendation method and device for herbal extraction, which is used to determine the user's health strengthening indicators through comprehensive analysis of three aspects: face, scalp and questionnaire survey, so that the required extracted drugs can meet the actual pathological needs of users, and then determine the conditioning direction and extraction method through the health strengthening indicators to ensure the highest efficiency of drug extraction and the effective use of drugs subsequently.

[0005] The present invention provides an intelligent recommendation method for herbal extraction, including:

[0006] Step 1: When the user triggers the extraction device, collect the facial image and scalp image of the user, perform image processing and recognition, and obtain the facial characterization information and scalp characterization information of the user;

[0007] Step 2: Receive the questionnaire survey information related to health input by the user on the interaction interface of the extraction device, and perform information analysis to obtain the current basic information. Among them, the questionnaire survey information includes: the actual pulse rate measured by the user on site using a pulse sensor and several related health index information;

[0008] Step 3: Perform comprehensive analysis and processing on the facial characterization information, scalp characterization information and current basic information, preliminarily estimate the current health condition of the user, and extract several health strengthening indicators from the current health condition;

[0009] Step 4: Determine the conditioning direction of the user depending on all health strengthening indicators, and perform intelligent matching with the conditioning effects of traditional Chinese medicine herbs and various extraction methods of traditional Chinese medicine herbs existing in the extraction device, and output a recommended method for the user to choose.

[0010] Preferably, obtaining the facial representation information of the user includes:

[0011] Extract the three-channel components of each pixel point in the facial image, compare the three-channel components with the component combination - skin color database respectively, and match skin color labels to the corresponding pixel points;

[0012] Perform clustering processing on all skin color labels to obtain several skin color blocks, determine the facial color characteristics of each skin color block respectively, and construct an initial sequence in combination with the facial position where the corresponding skin color block is located;

[0013] Perform low-saturation adjustment and high-saturation adjustment on the image blocks corresponding to each skin color block to obtain low-saturation image blocks and high-saturation image blocks respectively;

[0014] Extract the pixel values of the same pixel point from the low-saturation image and the high-saturation image respectively, and combine the first skin color feature and the second skin color feature of the corresponding skin color block under the low-saturation image and the high-saturation image to obtain a reference pair for the corresponding pixel point;

[0015] Perform a peripheral search on each pixel point respectively, and lock the tomographic boundary points based on the pixel value of the corresponding pixel point as a reference basis to obtain the tomographic locking quantity of each pixel point;

[0016] Expand the initial sequence according to the tomographic locking quantity and in combination with the reference pair of the corresponding pixel point to obtain a facial color sequence;

[0017] Input the facial color sequence into the skin color sequence analysis model to obtain the skin color information of the corresponding skin color block;

[0018] Based on all skin color information and the facial occupation size of each skin color information corresponding skin color block, obtain the facial representation information.

[0019] Preferably, expanding the initial sequence to obtain a facial color sequence includes:

[0020] Perform a first mean processing on the two pixel values in the reference pair of the pixel point to obtain a first value;

[0021] Determine the absolute value of the difference between the pixel value of each tomographic boundary point locked each time in the tomographic locking quantity and the pixel value of the pixel point corresponding to the reference basis, and perform a second mean processing to obtain a second value;

[0022] Determine the auxiliary skin color information of the corresponding pixel point according to the first value, the second value, the tomographic locking quantity, and the straight-line distance between each tomographic boundary point locked and the pixel point corresponding to the reference basis;

[0023] Add the auxiliary skin color information to the corresponding facial positions in the initial sequence to obtain a complexion sequence.

[0024] Preferably, perform information parsing to obtain the current basic information, including:

[0025] Retrieve the buried points of several responses of the user to each health problem from the questionnaire information respectively, and adjust the initial weights of the buried points in combination with the modal particles of each response buried point to obtain a buried point vector;

[0026] Input the buried point vector into an answer parsing model matching the health problem to obtain a health disorder;

[0027] Among them, all health disorders are used as the current basic information.

[0028] Preferably, comprehensively analyze and process the facial feature information, scalp feature information, and current basic information, including:

[0029] Perform a first extraction on the facial feature information according to the facial index set;

[0030] Perform a second extraction on the scalp feature information according to the scalp index set;

[0031] Perform a third extraction on the current basic information according to the Q&A index set;

[0032] Input the first extraction result, the second extraction result, and the third extraction result into the result analysis table in sequence.

[0033] Preferably, extract several health strengthening indicators from the current health condition, including:

[0034] Perform a full-number combination analysis on the input results in the result analysis table according to the analysis criteria to obtain the combination analysis results of the corresponding times;

[0035] Extract non-health factors from each combination analysis result respectively and use them as initial indicators;

[0036] Perform a union operation on all initial indicators to obtain several health strengthening indicators.

[0037] Preferably, determine the conditioning direction of the user depending on all health strengthening indicators, including:

[0038] Determine the number of occurrences of each health strengthening indicator from the combination analysis results of all times;

[0039] Determine the interaction relationship between the health strengthening indicator and each of the remaining indicators in the corresponding combination analysis result, and determine the preliminary radiation range based on the corresponding combination analysis result;

[0040] Determine all interaction relationships involved in each health enhancement indicator in the corresponding combined analysis results, sort them in order of relationship strength, and calculate the first strength differences between adjacent interaction relationships respectively;

[0041] Lock the jump strength difference from the first strength differences, determine the optimization strength based on the first strength differences before the position where the jump strength difference is located, and adjust the interaction relationships before the position according to the optimization strength to obtain the enhancement factor of the corresponding health enhancement indicator in the corresponding combined analysis result;

[0042] Obtain the extended radiation range based on the preliminary radiation range and the enhancement factor, and construct an association expression;

[0043] Obtain the health attributes of each association expression based on the expression-attribute comparison table, and combine the number of occurrences to obtain the attribute set of the corresponding health enhancement indicator;

[0044] Obtain the sub-directions and index weights that match the corresponding attribute set from the set-direction comparison table;

[0045] Determine the conditioning direction based on all sub-directions and index weights.

[0046] Preferably, it is intelligently matched with the conditioning effect of traditional Chinese medicine herbs existing in the extraction device and various extraction methods of traditional Chinese medicine herbs, including:

[0047] Eliminate the conditioning conflict of the sub-directions with index weights greater than the preset weight according to the traditional Chinese medicine conditioning theory mechanism, and perform functional effect matching between the conditioning content after conflict elimination and the traditional Chinese medicine herbs existing in the extraction device, screen out the first herbs with conditioning effects, and determine the herb functions and herb dosages of each first herb;

[0048] Input the sub-directions and index weights of all health enhancement indicators into the user drug recognition model to obtain the drug absorption effect of the user;

[0049] Determine the herb states of each first herb in the extraction device respectively, and combine the drug absorption effect of the user to determine several recommended extraction methods for the user to choose.

[0050] Preferably, the extraction method is related to overpressure purification filtration technology, slow release homogenization technology, micro-nano collision technology, and intelligent automatic temperature and pressure technology.

[0051] The present invention provides an intelligent recommended herb extraction device, including:

[0052] An image acquisition module, configured to collect the facial image and scalp image of the user when the extraction device is triggered by the user, perform image processing and recognition, and obtain the facial characterization information and scalp characterization information of the user;

[0053] An information parsing module, configured to receive the health-related questionnaire information input by the user on the interaction interface of the extraction device, and perform information parsing to obtain the current basic information;

[0054] A health prediction module, configured to comprehensively analyze and process the facial characterization information, scalp characterization information, and current basic information, preliminarily predict the current health condition of the user, and extract several health enhancement indicators from the current health condition;

[0055] An extraction recommendation module, configured to determine the conditioning direction of the user based on all health enhancement indicators, and perform intelligent matching with the conditioning effects of traditional Chinese medicine herbs and various extraction methods of traditional Chinese medicine herbs existing in the extraction device, and output recommended methods for the user to choose.

[0056] Compared with the prior art, the beneficial effects of the present application are as follows:

[0057] By comprehensively analyzing from three aspects of the face, scalp, and questionnaire survey to determine the health enhancement indicators of the user, it is convenient for the required extracted medicine to meet the actual pathological needs of the user, and then the conditioning direction and extraction method are determined through the health enhancement indicators to ensure the highest efficiency of medicine extraction and the effective use of the medicine subsequently to the greatest extent.

[0058] Other features and advantages of the present invention will be described in the following specification, and, in part, will be obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be realized and obtained by the structure specifically pointed out in the written specification and the drawings.

[0059] The technical solutions of the present invention will be further described in detail below through the drawings and embodiments. Description of the Drawings

[0060] The drawings are used to provide further understanding of the present invention, and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, and do not constitute a limitation to the present invention. In the drawings:

[0061] Figure 1 is a flowchart of an intelligent recommended herb extraction method in an embodiment of the present invention;

[0062] Figure 2 is a structural diagram of an intelligent recommended herb extraction device in an embodiment of the present invention;

[0063] Figure 3This is the structural diagram of the extraction device in the embodiments of the present invention. Detailed implementation manners

[0064] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only for the purpose of illustrating and explaining the present invention, and are not intended to limit the present invention.

[0065] The present invention provides an intelligent recommendation method for herbal extraction, as Figure 1 shown, including:

[0066] Step 1: When the user triggers the extraction device, collect the facial image and scalp image of the user, perform image processing and recognition, and obtain the facial characterization information and scalp characterization information of the user;

[0067] Step 2: Receive the health-related questionnaire information input by the user on the interaction interface of the extraction device, and perform information parsing to obtain the current basic information. Among them, the questionnaire information includes: the actual pulse rate measured by the user on-site using a pulse sensor and several related health index information;

[0068] Step 3: Perform comprehensive analysis and processing on the facial characterization information, scalp characterization information, and current basic information, preliminarily estimate the current health condition of the user, and extract several health enhancement indicators from the current health condition;

[0069] Step 4: Determine the conditioning direction of the user depending on all health enhancement indicators, and perform intelligent matching with the conditioning effects of traditional Chinese medicine herbs and various extraction methods of traditional Chinese medicine herbs existing in the extraction device, and output a recommended method for the user to choose.

[0070] In this embodiment, the extraction device, as Figure 3 shown, includes a housing part, a crushing part, an extraction part, a transmission part, a measurement part, a cooling part, a power part, an interaction part, a collection part, and a data processing part. Obtain the information triggered by the user through the interaction part, and then realize facial collection and scalp collection based on the collection part. Further, display the questionnaire content through the interaction part for the user to input to obtain the questionnaire information. Then, perform comprehensive analysis on the results of the three aspects through the data processing part to determine the health enhancement indicators, and then determine the conditioning direction to obtain the intelligent matching result (extraction method). After the user selects the recommended extraction method on the interaction part, control the crushing part, extraction part, transmission part, measurement part, and cooling part, and the power part starts to work to realize the extraction operation.

[0071] In this embodiment, the acquisition of facial images is for determining the user's skin color, whether there are acne on the face, whether the face is oily, etc. The acquisition of scalp images is for determining whether the user's scalp is oily, the amount of hair, the thickness and texture of the hair, etc.

[0072] In this embodiment, the questionnaire survey information includes: pulse beating information, drug allergy information, discomfort information existing in the current body, the information of the most recent food, etc.

[0073] In this embodiment, the comprehensive analysis and processing is for determining the current health condition of the user. For example, the user is generally healthy, but has excessive dampness and strong liver fire. At this time, the health strengthening indicators are: dampness elimination indicator, liver fire weakening indicator, etc.

[0074] In this embodiment, the conditioning direction is determined comprehensively based on the individual direction of the corresponding indicators and the indicator weights to ensure the order of priority of conditioning.

[0075] In this embodiment, for example, the traditional Chinese medicine herbs used to reduce liver fire are: honeysuckle, astragalus, gardenia, rehmannia glutinosa, rhubarb, moutan bark, alisma orientale, anemarrhena asphodeloides. For example, the poria cocos used to condition dampness. At this time, the weight ratio of the corresponding indicator of dampness accounts for 60%, and the weight ratio of the corresponding indicator of liver fire accounts for 40%. At this time, according to the conditioning effect of the formula, the grams and the addition of other herbs are supplemented, and the extraction method is recommended.

[0076] Among them, as long as there is a difference in one parameter in the aspects of ultra-pressure purification and filtration technology, slow release and homogenization technology, micro-nano collision technology, and intelligent automatic temperature and pressure technology for the extraction method, it is regarded as a corresponding method.

[0077] The beneficial effects of the above technical solution are: by comprehensively analyzing from three aspects of the face, scalp and questionnaire survey to determine the health strengthening indicators of the user, it is convenient for the required extracted drugs to meet the actual pathological needs of the user, and then the conditioning direction and extraction method are determined through the health strengthening indicators to ensure the high efficiency of drug extraction to the greatest extent and the effective use of the drugs subsequently.

[0078] The present invention provides an intelligent recommendation herbal extraction method, obtaining the facial characterization information of the user, including:

[0079] Extracting the three-channel components of each pixel point in the facial image, and comparing the three-channel components with the component combination - skin color database respectively, and matching skin color labels to the corresponding pixel points;

[0080] Performing clustering processing on all skin color labels to obtain several skin color blocks, respectively determining the facial color characteristics of each skin color block, and combining the facial positions where the corresponding skin color blocks are located to construct an initial sequence;

[0081] Perform low-saturation adjustment and high-saturation adjustment on the image blocks corresponding to each skin color block to obtain low-saturation image blocks and high-saturation image blocks respectively;

[0082] Extract the pixel values of the same pixel point from the low-saturation image and the high-saturation image respectively, and combine the first skin color feature and the second skin color feature of the corresponding skin color block in the low-saturation image and the high-saturation image to obtain a reference pair for the corresponding pixel point;

[0083] Perform a peripheral search on each pixel point respectively, and lock the fault boundary points based on the pixel value of the corresponding pixel point as a reference basis to obtain the fault locking quantity of each pixel point;

[0084] Expand the initial sequence according to the fault locking quantity and combine the reference pair of the corresponding pixel point to obtain a complexion sequence;

[0085] Input the complexion sequence into the skin color sequence analysis model to obtain the skin color information of the corresponding skin color block;

[0086] Based on all skin color information and the facial occupation size of each skin color information corresponding skin color block, obtain facial representation information.

[0087] In this embodiment, the three-channel components are: the red-channel component, the blue-channel component, and the green-channel component. For example, the three-channel components of pixel point A1 are; r1, b1, g1. At this time, the component combination - skin color database contains the three-channel components and the corresponding skin color labels, which are all pre-stored. For example: oily skin color, acne skin color, red temperature skin color, uneven skin color, etc., and then the skin color label of each pixel point can be obtained.

[0088] In this embodiment, the clustering analysis is implemented by using the k-means algorithm, which belongs to the prior art.

[0089] In this embodiment, after clustering, the pixel points involved in the same clustering result and the manifestation position of this pixel point on the face are regarded as a skin color block, that is, there are as many skin color blocks as there are clustering results.

[0090] In this embodiment, the complexion feature refers to the result obtained by inputting all the skin color labels involved in the corresponding skin color block into the facial analysis model, and this facial analysis model is trained for the neural network model based on different combinations of skin color conditions and the diagnosis results of experts on the skin color conditions of this combination. Therefore, the facial features of different skin color blocks can be directly obtained.

[0091] In this embodiment, the facial position is the manifestation position of the corresponding facial feature on the face, that is, it is related to the position of the pixel point corresponding to the label.

[0092] In this embodiment, the initial sequence = {facial position - complexion feature}.

[0093] In this embodiment, the low-saturation adjustment and high-saturation adjustment are for more obvious analysis of facial features. The set value of the low-saturation adjustment can be -100, and the set value of the high-saturation adjustment can be +100. After directly adjusting the settings, a saturated image block can be obtained.

[0094] In this embodiment, the information of the reference pair includes: the pixel values of the same pixel point in the low-saturation image and the high-saturation image, and the skin color features corresponding to the low-saturation image and the high-saturation image.

[0095] In this embodiment, the principles of obtaining the first skin color feature, the second skin color feature and the complexion feature are similar and will not be elaborated here.

[0096] In this embodiment, the peripheral search refers to the layer-by-layer search of neighboring pixel points until the fault boundary point is determined. That is, the difference between the pixel value of the fault boundary point and the pixel value of the corresponding pixel point is suddenly too large. For example, the pixel value of the original pixel point is a1, and the pixel value of the determined fault boundary point is b1. At this time, the absolute value of the difference between b1 and a1 is greater than the set value. It should be noted that during the layer-by-layer search process, the pixel point where the absolute value of the difference first satisfies being greater than the set value is regarded as the fault boundary point.

[0097] In this embodiment, since the peripheral search is performed on each pixel point, each pixel point may be used as a fault boundary point. Even if it is not used as the fault boundary point of pixel point B1, it may be used as the fault boundary point of pixel point B2. Therefore, it is necessary to perform a quantity statistics to determine the number of pixel points used as the fault boundary point, that is, the fault lock-in quantity.

[0098] In this embodiment, the complexion sequence = the initial sequence + the extended sequence.

[0099] In this embodiment, the skin color sequence analysis model is trained for the neural network model based on different skin color sequences and the judgment results of experts on different skin color sequences as samples. Therefore, skin color information can be obtained, further ensuring the accuracy of the determined skin color information.

[0100] In this embodiment, the facial representation information is: the skin color information of each skin color block, the position of the skin color block on the face, and the occupied size on the face.

[0101] The beneficial effects of the above technical solution are as follows: directly set skin color tags for the captured images to achieve preliminary clustering analysis and sequence construction, and then further enrich the facial information by adjusting the low saturation and high saturation of image blocks, and combine tomographic locking to expand the sequence, ensuring the effective and reasonable acquisition of facial representation information and providing a basis for subsequent analysis.

[0102] The present invention provides an intelligent recommendation herbal extraction method, which expands the initial sequence to obtain a complexion sequence, including:

[0103] Perform a first mean process on two pixel values in the reference pair of the pixel points to obtain a first value;

[0104] Determine the absolute value of the difference between the pixel value of each locked tomographic boundary point in the tomographic locking quantity and the pixel value of the pixel point corresponding to the reference basis, and perform a second mean process to obtain a second value;

[0105] According to the first value, the second value, the tomographic locking quantity, and the straight-line distance between each locked tomographic boundary point and the pixel point corresponding to the reference basis, determine the auxiliary skin color information of the corresponding pixel point;

[0106] Add the auxiliary skin color information to the corresponding facial position in the initial sequence to obtain a complexion sequence.

[0107] In this embodiment, the first value = (the sum of two pixel values) / 2.

[0108] In this embodiment, the second value = (the sum of all absolute values of differences involved in the pixel points of the same reference basis) / the tomographic locking quantity.

[0109] In this embodiment, the corresponding Among them, L i1 represents the straight-line distance between the i1-th locked tomographic boundary point and the pixel point corresponding to the reference basis; L max the maximum value among the distances between all any two cluster centers in all clustering results; ln represents the logarithmic function symbol; n1 represents the tomographic locking quantity, and a1 + a2 represents the sum of the characteristic differences between the first skin color feature and the second skin color feature of the skin color block where the corresponding pixel point is located and the corresponding complexion feature.

[0110] It should be noted that the auxiliary skin color information is the calculated pixel value.

[0111] In this embodiment, the complexion sequence = {the auxiliary skin color information of the pixel points corresponding to the initial sequence}.

[0112] The beneficial effects of the above technical solution are as follows: by performing a mean process on the pixel values in the reference pair, and

[0113] The present invention provides an intelligent recommendation herbal extraction method, which performs information parsing to obtain current basic information, including:

[0114] Retrieve several response buried points of the user for each health problem from the questionnaire survey information respectively, and adjust the initial weight of the buried point in combination with the modal particles of each response buried point to obtain a buried point vector;

[0115] Input the buried point vector into an answer parsing model matching the health problem to obtain a health disorder;

[0116] Among them, all health disorders are used as the current basic information.

[0117] In this embodiment, the response buried point refers to the information that needs to be captured after the user answers the corresponding health problem. For example, for the pulse frequency, the user's answer is: currently 120 times / min. At this time, the buried point is to only capture 120.

[0118] In this embodiment, the modal particles can be: possibly, probably, often, occasionally, etc. For example, the questionnaire question is: Do you have insomnia? At this time, the user's reply is: Occasionally have insomnia. For example, originally the weight set for insomnia is c1. At this time, the modal particle is occasionally, that is, not serious, and c1 can be subtracted by r01, where r01 is the reduction weight for the corresponding modal particle.

[0119] In this embodiment, the buried point vector = {the extraction result of the response buried points involved in the same question and the adjusted weight of the buried point}.

[0120] In this embodiment, the answer parsing model includes the same questionnaire question and the response result for this question, and the expert's judgment on the solution result is used as a sample for training the neural network model. Therefore, the health disorder can be directly obtained.

[0121] The beneficial effect of the above technical solution is that by performing buried point retrieval and weight adjustment, it is ensured that the obtained vector is more in line with the user's own disorder, thereby providing a data basis for subsequent analysis.

[0122] The present invention provides an intelligent recommendation herbal extraction method, which comprehensively analyzes and processes the facial feature information, scalp feature information and current basic information, including:

[0123] Perform a first extraction on the facial feature information according to the facial index set;

[0124] Perform a second extraction on the scalp feature information according to the scalp index set;

[0125] Perform a third extraction on the current basic information according to the question-and-answer index set;

[0126] Input the first extraction result, the second extraction result, and the third extraction result into the result analysis table in sequence.

[0127] In this embodiment, the facial index set includes: facial greasiness index, facial sallowness index, facial redness and temperature index, facial acne index, etc.

[0128] In this embodiment, the scalp index set includes: scalp greasiness index, hair quality index, etc.

[0129] In this embodiment, the Q&A index set includes: pulse frequency index, allergy index, last food index, etc.

[0130] The purpose of extraction is to analyze the relevant results.

[0131] In this embodiment, the result analysis table contains the index descriptions of various different indexes and the blank placement positions corresponding to the results of the corresponding indexes. Only need to place the extracted results in the corresponding blank positions.

[0132] The beneficial effects of the above technical solution are: By using facial indexes, scalp indexes, and Q&A indexes to extract from the corresponding information in sequence, it effectively ensures the reliability of result acquisition.

[0133] The present invention provides an intelligent recommendation herbal extraction method, which extracts several health strengthening indexes from the current health condition, including:

[0134] Perform full-number combination analysis on the input results in the result analysis table according to the analysis standard to obtain the combination analysis results corresponding to the corresponding times;

[0135] Respectively extract non-health factors from each combination analysis result and use them as initial indexes;

[0136] Perform union processing on all initial indexes to obtain several health strengthening indexes.

[0137] In this embodiment, the full-number combination analysis refers to randomly shuffling and combining the results of the three aspects, and it is required that all indexes have been combined. For example, there are results corresponding to index 1, index 2, and index 3 respectively. At this time, combine index 1 and index 2, index 2 and index 3, index 1, index 2, and index 3 respectively, that is, obtain 3 combination analysis results.

[0138] In this embodiment, non-health factors refer to factors that affect physical health, such as heavy dampness, excessive liver fire, etc.

[0139] For example, in the combined analysis result 01, there are initial indicators u1 and u2, and in the combined analysis result 02, there are initial indicators u3 and u1. At this time, the health enhancement indicators obtained after the union processing are: initial indicator u1, initial indicator u2, and initial indicator u3.

[0140] The beneficial effect of the above technical solution is that all possibilities can be included by performing the combined analysis for all times, ensuring the comprehensiveness of obtaining health enhancement indicators and providing a reliable basis for determining the conditioning direction subsequently.

[0141] The present invention provides an intelligent recommendation herbal extraction method, which determines the conditioning direction of the user depending on all health enhancement indicators, including:

[0142] Determine the occurrence times of each health enhancement indicator from the combined analysis results of all times;

[0143] Determine the interaction relationship between the health enhancement indicator and each of the remaining indicators in the corresponding combined analysis result, and determine the preliminary radiation range based on the corresponding combined analysis result;

[0144] Sort in sequence according to the relationship strength all the interaction relationships involved by each health enhancement indicator in the corresponding combined analysis result, and calculate the first strength difference between adjacent interaction relationships respectively;

[0145] Lock the jump strength difference from the first strength differences, determine the optimization strength based on the first strength differences before the position where the jump strength difference is located, and adjust the interaction relationships before the position according to the optimization strength to obtain the enhancement factor of the corresponding health enhancement indicator in the corresponding combined analysis result;

[0146] Obtain the extended radiation range based on the preliminary radiation range and the enhancement factor, and construct an association expression;

[0147] Obtain the health attributes of each association expression based on the expression-attribute comparison table, and obtain the attribute set of the corresponding health enhancement indicator in combination with the occurrence times;

[0148] Obtain the sub-direction and the index weight that match the corresponding attribute set from the set-direction comparison table;

[0149] Determine the conditioning direction based on all sub-directions and index weights.

[0150] In this embodiment, for example, the health enhancement indicator u1 appears in both the combined analysis result 01 and the combined analysis result 02. At this time, the occurrence times are regarded as 2.

[0151] In this embodiment, the interaction relationship is obtained by inputting all the health enhancement indicators in the combined analysis result into the mutual analysis model. The mutual analysis model is trained with neural network models using samples based on the interaction among health enhancement indicators. Therefore, it can be directly obtained.

[0152] In this embodiment, for example, there is no relationship between health enhancement indicator u1 and health enhancement indicator u2, and there is a strength relationship of 0.2 with health enhancement indicator u3. At this time, the result corresponding to health enhancement indicator u2 is not associated with health enhancement indicator u1, and 20% of the results associated in the result of health enhancement indicator u3 are associated with health enhancement indicator u1. At this time, the preliminary radiation range for this health enhancement indicator u1 can be obtained, that is, the content related to other health enhancement indicators.

[0153] In this embodiment, the interaction relationship and the relationship strength are realized in mutual contrast. For example, when the interaction relationship is non - existent, the relationship strength at this time is 0; when the interaction relationship is a close relationship, the relationship strength at this time is 1. The two are realized based on a relationship - strength comparison table, which includes different interaction relationships and the corresponding relationship strengths.

[0154] In this embodiment, the first strength difference is the strength difference between adjacent interaction relationships after sorting.

[0155] In this embodiment, the beating strength difference refers to the strength in the first strength difference where the value is greater than r01 (a set value).

[0156] In this embodiment, the positions locked by the beating strength difference correspond to adjacent interaction relationships, and there are two positions. The earlier of the two positions is regarded as the position where it is located.

[0157] In this embodiment, where floor(m1 / 2) is the floor value of m1 / 2, and m1 is the number of all the first strength differences before the position where it is located;

[0158] In this embodiment, the adjustment results of the relationship strength corresponding to each interaction relationship before the position where it is located:

[0159] where wx represents the position serial number of the relationship strength corresponding to the corresponding interaction relationship before the position where it is located, and the value range of the position serial number is 1, 2, 3,..., m1.

[0161] In this embodiment, the enhancement factor = (the sum of all adjustment results - the sum of the strength before adjustment) / (m1 + 1).

[0162] In this embodiment, the enhancement factor is regarded as an extension of the associated results corresponding to the remaining strengthening indicators whose relationship strength is not 1. If the value of the original strength + the enhancement factor is greater than 1, full association extension is performed. If it is still less than 1, relevant association extension can be carried out according to the newly obtained value.

[0163] In this embodiment, the association expression is: the combination of the representative symbols of the associated results with each of the remaining health indicators after actual extension, and each association is unique, and thus the representative symbol is unique.

[0164] In this embodiment, the expression-attribute comparison table contains different symbol combinations and the health attributes matching the symbol combinations, and the health attributes are related to dampness, phlegm, cold air, liver fire, lung fire, etc.

[0165] In this embodiment, the attribute set = {the health attributes of the health strengthening indicators under different occurrence times}.

[0166] In this embodiment, the set-direction comparison table is related to different combinations of attribute sets and the conditioning sub-directions and index weights matching the set, and thus the conditioning direction can be obtained.

[0167] The beneficial effects of the above technical solution are: determining the preliminary radiation range through the interaction relationship, and then obtaining the enhancement factor through the sorting of intensities, the determination of intensity differences, and the adjustment of relationships, which can ensure the comprehensiveness of associations to realize the reasonable acquisition of expressions, and provide convenience for the subsequent acquisition of sub-directions and weights of different health strengthening indicators.

[0168] The present invention provides an intelligent recommendation herbal extraction method, which intelligently matches the conditioning effects of traditional Chinese medicine herbs and various extraction methods of traditional Chinese medicine herbs existing in the extraction device, including:

[0169] Eliminating the conditioning conflicts of sub-directions with index weights greater than the preset weight according to the traditional Chinese medicine conditioning theory mechanism, and performing functional effect matching between the conditioning content after conflict elimination and the traditional Chinese medicine herbs existing in the extraction device, screening out the first herbs with conditioning effects and determining the herbal functions and herbal dosages of each first herb;

[0170] Inputting the sub-directions and index weights of all health strengthening indicators into the user drug recognition model to obtain the drug absorption effect of the user;

[0171] Respectively determining the herbal states of each first herb in the extraction device, and combining the drug absorption effect of the user to determine several recommended extraction methods for the user to choose.

[0172] Among them, after the user makes a selection, all the required first herbs are stirred and medicinally extracted according to the extraction method, and the user is reminded to receive the traditional Chinese medicine based on the extraction device. The reminder can be an interface reminder based on the interaction part or a broadcast reminder, etc.

[0173] Preferably, the extraction method is related to the overpressure purification and filtration technology, the slow release and homogenization technology, the micro-nano collision technology, and the intelligent automatic temperature and pressure technology.

[0174] In this embodiment, the theoretical mechanism of traditional Chinese medicine conditioning is pre-established, which is based on the basic theories of traditional Chinese medicine such as observation, auscultation and olfaction, interrogation, and palpation, providing the association between the most basic and comprehensive traditional Chinese medicine herb information and the medication situation.

[0175] In this embodiment, the value of the preset weight is 0.1.

[0176] In this embodiment, the conditioning conflict is to avoid the situation that although herbs e1 and e2 can respectively condition the health problems in the corresponding directions y1 and y2, they cannot be used simultaneously.

[0177] In this embodiment, the functional effect is the effect of each herb, which is known.

[0178] In this embodiment, the function and dosage can be directly output after matching.

[0179] In this embodiment, the drug recognition model is trained on a neural network model with different directions, weight combinations, and the estimated absorption effect results based on this combination as samples. Therefore, the drug absorption effect can be directly obtained.

[0180] In this embodiment, the herb state refers to solid block, powder, etc.

[0181] In this embodiment, the extraction method is obtained by matching from a state-effect-method comparison table, which includes different herb states, drug absorption effects, and matching extraction methods.

[0182] The beneficial effects of the above technical solutions are: by eliminating drug conflicts and matching functional effects, reasonable herbs can be screened, and then combined with the drug absorption effect, the extraction method is recommended to achieve intelligent extraction.

[0183] The present invention provides an intelligent recommended herb extraction device, as Figure 2 shown, including:

[0184] An image acquisition module, configured to collect the facial image and scalp image of the user when the user triggers the extraction device, perform image processing and recognition, and obtain the facial characterization information and scalp characterization information of the user;

[0185] An information analysis module, configured to receive the health-related questionnaire information input by the user on the interaction interface of the extraction device, and perform information analysis to obtain the current basic information;

[0186] A health prediction module, configured to comprehensively analyze and process the facial feature information, scalp feature information, and the current basic information, preliminarily predict the current health condition of the user, and extract several health enhancement indicators from the current health condition;

[0187] An extraction recommendation module, configured to determine the conditioning direction of the user depending on all the health enhancement indicators, and perform intelligent matching with the conditioning effects of the traditional Chinese medicine herbs existing in the extraction device and various extraction methods of the traditional Chinese medicine herbs, and output a recommended method for the user to select.

[0188] The beneficial effects of the above technical solution are: By comprehensively analyzing from three aspects of the face, scalp, and questionnaire survey to determine the health enhancement indicators of the user, it is convenient for the required extracted medicine to meet the actual pathological needs of the user, and then the conditioning direction and extraction method are determined through the health enhancement indicators to ensure the high efficiency of the medicine extraction and the effective use of the medicine subsequently to the greatest extent.

[0189] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. An intelligent recommendation herbal extraction method, characterized in that: include: Step 1: When the user triggers the extraction device, the facial image and scalp image of the user are collected and image processing and recognition are performed to obtain the facial representation information and scalp representation information of the user; Step 2: receiving the health-related questionnaire information input by the user in the interactive interface of the extraction device, and performing information analysis to obtain current basic information, wherein the questionnaire information includes: the actual pulse rate measured by the user using a pulse sensor on site and several related health index information; Step 3: Comprehensively analyzing and processing the facial representation information, scalp representation information and current basic information, preliminarily estimating the current health condition of the user, and extracting a number of health enhancement indicators from the current health condition; Step 4: Determine the conditioning direction of the user based on all health enhancement indicators, and intelligently match the conditioning effects of the Chinese medicinal herbs in the extraction device and the various extraction methods of the Chinese medicinal herbs, and output a recommended method for the user to choose.

2. The intelligent recommendation herbal extraction method according to claim 1, characterized in that: Obtaining facial representation information of the user, including: Extracting three-channel components of each pixel in the facial image, and comparing the three-channel components with the component combination-skin color database, and matching skin color labels to corresponding pixels; Clustering all skin color labels to obtain several skin color blocks, and determining the facial color features of each skin color block respectively, and combining the facial position of the corresponding skin color block to construct an initial sequence; Performing low saturation adjustment and high saturation adjustment on the image block corresponding to each skin color block, to obtain a low saturation image block and a high saturation image block respectively; Extracting pixel values ​​of the same pixel from the low-saturation image and the high-saturation image respectively, and combining the first skin color feature and the second skin color feature of the corresponding skin color block in the low-saturation image and the high-saturation image to obtain a reference pair of the corresponding pixel; Perform a perimeter search on each pixel point respectively, and lock the fault boundary point based on the pixel value of the corresponding pixel point as a reference to obtain the fault locking number of each pixel point; According to the number of locked slices and in combination with reference pairs of corresponding pixel points, the initial sequence is expanded to obtain a facial color sequence; Inputting the facial color sequence into a skin color sequence analysis model to obtain skin color information of a corresponding skin color block; Based on all skin color information and the facial occupancy size of the skin color block corresponding to each skin color information, the facial representation information is obtained.

3. The intelligent recommendation herbal extraction method according to claim 2, characterized in that: The initial sequence is expanded to obtain a facial color sequence, including: Performing a first mean processing on two pixel values ​​in a reference pair of the pixel points to obtain a first value; Determine the absolute value of the difference between the pixel value of each locked fault boundary point in the fault locking quantity and the pixel value of the corresponding pixel point as the reference basis, and perform a second mean value processing to obtain a second value; Determine the auxiliary skin color information of the corresponding pixel point according to the first value, the second value and the number of fault locks, based on the straight-line distance between each locked fault boundary point and the corresponding pixel point as a reference basis; The auxiliary skin color information is added to the corresponding facial position in the initial sequence to obtain a facial color sequence.

4. The intelligent recommendation herbal extraction method according to claim 1, characterized in that: Analyze the information and obtain the current basic information, including: Retrieve several answer embedding points of the user to each health question from the questionnaire information, and adjust the initial weight of the embedding point based on the modal particle of each answer embedding point to obtain the embedding point vector; Inputting the buried point vector into an answer parsing model matching the health question to obtain health symptoms; Among them, all health conditions are taken as current basic information.

5. The intelligent recommendation herbal extraction method according to claim 1, characterized in that: The facial representation information, scalp representation information and current basic information are comprehensively analyzed and processed, including: Performing a first extraction on the facial representation information according to a facial indicator set; Performing a second extraction on the scalp representation information according to the scalp indicator set; Performing a third extraction of the current basic information according to the question-answer indicator set; The first extraction result, the second extraction result, and the third extraction result are sequentially input into the result analysis table.

6. The intelligent recommendation herbal extraction method according to claim 5, characterized in that: Several health enhancement indicators are extracted from the current health status, including: Perform a combination analysis of all the input results in the result analysis table according to the analysis standard to obtain a combination analysis result of a corresponding number of times; Extract the non-healthy factors from each combined analysis result and use them as initial indicators; All initial indicators are combined to obtain several health enhancement indicators.

7. The intelligent recommendation herbal extraction method according to claim 6, characterized in that: All health enhancement indicators are relied upon to determine the conditioning direction of the user, including: The number of occurrences of each health enhancement indicator was determined from the combined analysis results of all times; Determine the interaction relationship between the health enhancement index and each of the other indicators in the corresponding combined analysis results, and determine the preliminary radiation range based on the corresponding combined analysis results; Determine all the interaction relationships involved in the corresponding combination analysis results of each health enhancement indicator, sort them in order according to the relationship strength, and calculate the first intensity difference between adjacent interaction relationships respectively; The beating intensity difference is locked from the first intensity difference, and the optimization intensity is determined based on the first intensity difference before the position where the beating intensity difference is located, and the interaction relationship before the position is adjusted according to the optimization intensity to obtain the enhancement factor of the corresponding health enhancement index in the corresponding combination analysis result; Based on the preliminary radiation range and the enhancement factor, an extended radiation range is obtained, and a correlation expression is constructed; Based on the expression-attribute comparison table, the health attribute of each associated expression is obtained, and the attribute set of the corresponding health enhancement indicator is obtained in combination with the number of occurrences; Obtain the sub-direction and indicator weight matching the corresponding attribute set from the set-direction comparison table; Based on all sub-directions and indicator weights, the adjustment direction is determined.

8. The intelligent recommendation herbal extraction method according to claim 7, characterized in that: Intelligent matching with the conditioning effects of Chinese medicinal herbs in the extraction device and the various extraction methods of Chinese medicinal herbs, including: Eliminate the conditioning conflicts of the sub-directions whose index weights are greater than the preset weights according to the TCM conditioning theory mechanism, and match the functional effects of the conditioning contents after the conflicts are eliminated with the TCM herbs in the extraction device, screen out the first herbs with conditioning effects, and determine the herbal functions and herbal dosage of each first herb; Inputting the sub-directions and indicator weights of all health enhancement indicators into the user drug identification model to obtain the drug absorption effect of the user; The herbal state of each first herb in the extraction device is determined respectively, and in combination with the drug absorption effect of the user, several recommended extraction methods are determined for the user to choose.

9. The intelligent recommendation herbal extraction method according to claim 8, characterized in that: The extraction method is related to ultra-pressure purification and filtration technology, release and equalization technology, micro-nano collision technology and intelligent automatic temperature and pressure technology.

10. An intelligent recommendation herbal extraction device, characterized in that: include: An image acquisition module, used for acquiring a facial image and a scalp image of the user and performing image processing and recognition to obtain facial representation information and scalp representation information of the user when the user triggers the extraction device; An information analysis module, used to receive the health-related questionnaire information input by the user in the interactive interface of the extraction device, and perform information analysis to obtain current basic information; A health prediction module is used to comprehensively analyze and process the facial representation information, scalp representation information and current basic information, preliminarily predict the current health status of the user, and extract a number of health enhancement indicators from the current health status; The extraction recommendation module is used to determine the conditioning direction of the user based on all health enhancement indicators, and intelligently match the conditioning effects of the Chinese medicinal herbs in the extraction device and the various extraction methods of the Chinese medicinal herbs, and output the recommended method for the user to choose.

Images