CN115587829A - Dish recommendation method and system - Google Patents
Dish recommendation method and system Download PDFInfo
- Publication number
- CN115587829A CN115587829A CN202110691095.6A CN202110691095A CN115587829A CN 115587829 A CN115587829 A CN 115587829A CN 202110691095 A CN202110691095 A CN 202110691095A CN 115587829 A CN115587829 A CN 115587829A
- Authority
- CN
- China
- Prior art keywords
- target object
- data
- face
- ordering
- dish
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 75
- 230000008921 facial expression Effects 0.000 claims abstract description 62
- 230000001815 facial effect Effects 0.000 claims abstract description 36
- 238000012545 processing Methods 0.000 claims description 22
- 238000001514 detection method Methods 0.000 claims description 16
- 238000000605 extraction Methods 0.000 claims description 8
- 238000013528 artificial neural network Methods 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 4
- 230000000694 effects Effects 0.000 abstract description 12
- 230000008569 process Effects 0.000 description 33
- 230000008859 change Effects 0.000 description 10
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 235000019640 taste Nutrition 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000008451 emotion Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 229910044991 metal oxide Inorganic materials 0.000 description 1
- 150000004706 metal oxides Chemical class 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0281—Customer communication at a business location, e.g. providing product or service information, consulting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/12—Hotels or restaurants
Landscapes
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Finance (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a dish recommending method and a dish recommending system, wherein the dish recommending method comprises the following steps: acquiring human face video data comprising at least one frame of image; performing face attribute recognition on at least partial image including a target object in the face video data to obtain facial expression data and facial posture data of the target object; generating a current click list preference of the target object according to the facial expression data and the facial pose data; when determining that the historical point list data of the target object exist, generating historical point list preference of the target object according to the historical point list data; and according to a preset rule, obtaining a dish to be recommended according to the historical ordering preference and the current ordering preference, and displaying the dish to be recommended to the target object. Is used for improving dish recommending effect.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a dish recommending method and system.
Background
Because the traditional manual ordering efficiency is low, the prior face recognition ordering system is provided, the ordering preference of a user is recorded by generating a historical database, when the user orders again, the user identity is recognized by using a face recognition technology, and products which are frequently ordered before the user is automatically retrieved and presented to the user. The whole process does not consider the real-time feeling of the user and the change of taste, and the dish recommending effect is poor. In addition, similar eating tastes can be recommended to users with similar facial features on the basis of face recognition, however, the eating tastes with similar facial features can be different, and the dish recommendation practicability is not strong. Therefore, it becomes additionally important how to improve the dish recommending effect.
Disclosure of Invention
The invention provides a dish recommending method and system, which are used for improving dish recommending effect.
In a first aspect, an embodiment of the present invention provides a dish recommendation method, including:
acquiring human face video data comprising at least one frame of image;
carrying out face attribute recognition on at least partial image of the face video data, which comprises a target object, to obtain face expression data and face posture data of the target object;
generating a current click list preference of the target object according to the facial expression data and the facial pose data;
when determining that the historical point list data of the target object exist, generating historical point list preference of the target object according to the historical point list data;
and according to a preset rule, obtaining a dish to be recommended according to the historical ordering preference and the current ordering preference, and displaying the dish to be recommended to the target object.
In one possible implementation, the generating a current click list preference of the target object according to the facial expression data and the facial pose data includes:
determining an expression code for representing a target expression of the target object according to the facial expression data;
determining a target moment corresponding to the expression code;
determining face orientation data corresponding to the target time from the face posture data;
determining a heat power area of interest corresponding to the face orientation data and a first dish set corresponding to the heat power area of interest;
and taking the first dish set as the current ordering preference of the target object.
In one possible implementation manner, the generating, when it is determined that there is history point list data of the target object, a history point list preference of the target object according to the history point list data includes:
determining a historical ordering record of the target object according to the historical ordering data;
determining a second dish set and the number of times of ordering each dish from the historical ordering records;
and generating historical ordering preference of the target object according to the second dish set and the ordering times.
In a possible implementation manner, the obtaining, according to a preset rule, a dish to be recommended according to the historical ordering preference and the current ordering preference, and displaying the dish to be recommended to the target object includes:
respectively setting the weights of the historical ordering preference and the current ordering preference according to the sales condition of the dishes;
determining the weight of the dish to be recommended according to the historical ordering preference and the weight corresponding to the current ordering preference;
and displaying the dishes to be recommended to the target object according to the sequence of the weights from big to small.
In a possible implementation manner, the performing face attribute recognition on at least a partial image of the face video data that includes a target object to obtain facial expression data and facial pose data of the target object includes:
performing feature extraction on at least partial image including a target object in the face video data through a deep neural network to determine face expression data of the target object;
extracting at least one key point of the target object from each frame of the at least partial image;
and determining the facial pose data of the target object according to the at least one key point.
In one possible implementation, after the acquiring the face video data including at least one frame of image, the method further includes:
carrying out face detection on each frame of image in the at least partial images, and screening out images of which the face definition and the face shielding degree meet the preset quality;
and taking the image which meets the preset quality in the at least partial image as a face image to be subjected to face attribute identification.
In one possible implementation, after the acquiring the face video data including at least one frame of image, the method further includes:
and if the target object is detected to be placed in order, stopping recommending the dishes to be recommended to the target object.
In a second aspect, a dish recommendation system according to an embodiment of the present invention includes:
the system comprises an image acquisition module, a face attribute identification module and a processing module;
the image acquisition module is used for acquiring human face video data comprising at least one frame of image;
the face attribute recognition module is used for carrying out face attribute recognition on at least partial image of the face video data, including a target object, to obtain face expression data and face posture data of the target object, and sending the face expression data and the face posture data to the processing module;
the processing module is used for generating current ordering preference of the target object according to the facial expression data and the facial posture data, generating historical ordering preference of the target object according to the historical ordering preference when determining that historical ordering data of the target object exists, obtaining dishes to be recommended according to the current ordering preference and the historical ordering preference and displaying the dishes to be recommended to the target object according to a preset rule.
In one possible implementation, the processing module is configured to:
determining an expression code for representing the target expression of the target object according to the facial expression data;
determining a target moment corresponding to the expression code;
determining face orientation data corresponding to the target time from the face posture data;
determining a heat power area of interest corresponding to the face orientation data and a first dish set corresponding to the heat power area of interest;
and taking the first dish set as the current ordering preference of the target object.
In one possible implementation, the processing module is configured to:
determining a historical ordering record of the target object according to the historical ordering data;
determining a second dish set and the number of times of ordering each dish from the historical ordering records;
and generating historical ordering preference of the target object according to the second dish set and the ordering times.
In one possible implementation, the processing module is configured to:
respectively setting the weights of the historical ordering preference and the current ordering preference according to the sales condition of the dishes;
determining the weight of the dish to be recommended according to the historical ordering preference and the weight corresponding to the current ordering preference;
the dish recommendation system further comprises a display module, and the display module is used for: and displaying the dishes to be recommended to the target object according to the sequence of the weights from large to small.
The invention has the following beneficial effects:
the embodiment of the invention provides a menu recommending method and a menu recommending system, which comprises the steps of firstly, obtaining face video data comprising at least one frame of image, then, carrying out face attribute recognition on at least partial image comprising a target object in the face video data, for example, when the at least one frame of image comprising the face video data is a plurality of frames, carrying out face attribute recognition on each frame of image comprising the target object in the plurality of frames of image, or carrying out face attribute recognition on partial image comprising the target object in the plurality of frames of image, so as to obtain face expression data and face posture data of the target object, then, generating current click list preference of the target object according to the face expression data and the face posture data, because the face expression data and the face posture data can change in real time, generating actual click list requirement of the target object according to the current click list preference determined according to the face expression data and the face posture data which change in real time of the target object, and further, when determining that the current click list preference of the target object is closer to the historical click list data of the target object, generating historical click list preference of the target object according to the historical click list preference, namely, determining that the historical click list preference of the target object is better according to the historical click list preference of the historical click list data of the target object. Then, according to a preset rule, obtaining the dish to be recommended according to the historical ordering preference and the current ordering preference of the target object, and displaying the dish to be recommended to the target object. In other words, dish recommendation can be performed on the target object by combining the historical ordering preference and the current ordering preference of the target object, so that the dish recommendation not only considers the current ordering preference of the target object, but also considers the historical ordering preference, and the dish recommendation effect is improved.
Drawings
Fig. 1 is a flowchart of a method of ordering a recommendation method according to an embodiment of the present invention;
FIG. 2 is a flowchart of the method of step S103 in FIG. 1;
FIG. 3 is a flowchart of the method of step S104 in FIG. 1;
FIG. 4 is a flowchart of the method of step S105 in FIG. 1;
FIG. 5 is a flowchart of the method of step S102 in FIG. 1;
FIG. 6 is a flowchart of one of the methods after step S101 in FIG. 1;
FIG. 7 is a flowchart illustrating an overall procedure of a method for ordering a list according to an embodiment of the present invention;
fig. 8 is a block diagram of a click list recommendation system according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the drawings of the embodiments of the present invention. It is to be understood that the embodiments described are only a few embodiments of the present invention, and not all embodiments. And the embodiments and features of the embodiments may be combined with each other without conflict. All other embodiments, which can be derived by a person skilled in the art from the described embodiments of the invention without inventive step, are within the scope of protection of the invention.
Unless defined otherwise, technical or scientific terms used herein shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention belongs. The use of "including" or "comprising" and the like in the present invention is intended to mean that the elements or items listed before the word "comprise" or "comprising" and the like, include the elements or items listed after the word and their equivalents, but do not exclude other elements or items.
It should be noted that the sizes and shapes of the figures in the drawings are not to be considered true scale, but are merely intended to schematically illustrate the present invention. And the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout.
In the prior art, dishes are often recommended to a user through historical ordering preferences of the user, real-time feeling and taste change of the user are not considered in the whole recommending process, and dish recommending effect is poor.
In view of this, the embodiment of the invention provides a dish recommendation method and system, which are used for improving the dish recommendation effect.
As shown in fig. 1, a method for recommending dishes according to an embodiment of the present invention includes:
s101: acquiring face video data comprising at least one frame of image;
in a specific implementation process, the face video data may be acquired by an image acquisition unit including a Charge-coupled Device (CCD) or a Complementary Metal Oxide Semiconductor (CMOS), and the face video data may include one frame of image or multiple frames of images, which may be sequentially arranged according to a time sequence, and the images included in the face video data may be set according to an actual use requirement in an actual application, which is not limited herein.
S102: performing face attribute recognition on at least partial image including a target object in the face video data to obtain facial expression data and facial posture data of the target object;
in a specific implementation process, after the face video data is acquired, a machine learning method can be used for carrying out face detection on a first frame image in the face video data, face positions and occupied area in the images are extracted, at least one object is determined according to the face positions and the occupied area of each face position, when a plurality of objects are provided, the object closest to an image acquisition unit for acquiring the face video data in the plurality of objects can be used as a target object, and the object with the largest occupied area of the face positions in the plurality of objects can be used as the target object; when there is at least one object, the object is directly used as a target object, and of course, the target object may also be set according to actual application requirements, which is not limited herein. After the target object is determined, facial attribute recognition may be performed on at least a partial image of the facial video data including the target object, and facial expression data and facial pose data of the target object are determined, wherein the facial expression data is used for representing the expression of the target object, such as happiness, anger and the like, and the facial pose data is used for representing the facial pose of the target object, such as that the face faces the upper left corner of the ticker screen, and further such as that the face faces the lower right corner of the ticker screen and the like.
S103: generating a current point list preference of the target object according to the facial expression data and the facial gesture data;
in a specific implementation process, the current ordering preference of the target object may be generated according to the facial expression data and the facial pose data, the emotion change of the target object in the ordering process may be determined according to the change condition of the facial expression data in the ordering process of the target object, and the change condition of the target object in the ordering process may be determined according to the change condition of the facial pose data in the ordering process of the target object, so that the current ordering preference of the target object generated by the facial expression data and the facial pose data better conforms to the current ordering condition of the target object. In addition, in the target object ordering process, a text position where the target object focuses on the current ordering page and a stay time length of the target object focusing on the dish at the text position can be further determined through the facial pose data, for example, the stay time length is longer, and correspondingly, the weight of the dish at the text position is larger, and the stay time length is closer to the current ordering requirement of the target object. Therefore, the change conditions of the facial expression data and the facial posture data in the target object ordering process can be combined together to determine the current ordering preference of the target object, and therefore the current dish recommending efficiency is improved.
S104: when determining that the historical point list data of the target object exist, generating historical point list preference of the target object according to the historical point list data;
in a specific implementation process, when it is determined that the historical click list data of the target object exists, the historical click list preference of the target object can be generated according to the historical click list data, so that when the target object is an old user or an old customer, the historical click list preference of the target object can be generated according to the historical click list data of the target object, and corresponding guidance opinions can be provided for dish recommendation of the target object according to the historical click list preference.
S105: and according to a preset rule, obtaining a dish to be recommended according to the historical ordering preference and the current ordering preference, and displaying the dish to be recommended to the target object.
In a specific implementation process, the dish to be recommended may be obtained according to the historical ordering preference and the current ordering preference according to a preset rule, or the dish to be recommended may be determined by fusing the historical ordering preference and the current ordering preference together. The preset rule may be a rule preset according to actual application requirements, for example, the weights of the historical point preference and the current point preference are preset, for example, the weight of the current point preference is set to 0.8, and the weight of the historical point preference is set to 0.2, although other rules may also be set, which is not limited herein. The dish recommendation method and the dish recommendation system can perform dish recommendation on the target object by combining the historical ordering preference and the current ordering preference of the target object, so that the dish recommendation not only considers the current ordering preference of the target object, but also considers the historical ordering preference, and the dish recommendation effect is improved. In addition, as the current ordering preference of the target object can be generated according to the facial expression data and the facial posture data, when the historical ordering data of the target object is determined to exist, the historical ordering preference of the target object can be generated according to the historical ordering data, dishes to be recommended are generated according to the current ordering preference and the historical ordering preference and recommended to the target object, in the whole dish recommending process, the dish recommending can be completely carried out under the condition that an ordering user does not sense, the dish recommending effect is guaranteed, and meanwhile the ordering experience of the user is improved.
In the embodiment of the present invention, as shown in fig. 2, step S103: generating a current click list preference of the target object according to the facial expression data and the facial pose data, comprising:
s201: determining an expression code for representing a target expression of the target object according to the facial expression data;
s202: determining target time corresponding to the expression codes;
s203: determining face orientation data corresponding to the target moment from the face posture data;
s204: determining a heat power area of interest corresponding to the face orientation data and a first dish set corresponding to the heat power area of interest;
s205: and taking the first dish set as the current ordering preference of the target object.
In the specific implementation process, the specific implementation process from step S201 to step S205 is as follows:
firstly, face expression detection can be adopted to record real-time face expression data in the process of browsing a single page of the target object, and an expression code for representing the target expression of the target object is determined according to the face expression data, wherein the expression code of each frame of image of the target object in at least partial image can be determined, and then the expression code corresponding to the target expression is screened out. The corresponding relationship between the facial expression and the expression code may be preset, for example, the expression code when the facial expression is angry is preset to be 1, the expression code when the facial expression is normal to be 2, the expression code when the facial expression is happy to be 3, and the expression code when the facial expression is frightened to be 4, and of course, the corresponding relationship between the facial expression and the expression code may also be set according to the actual application situation, which is not limited herein.
After determining an expression code for representing a target expression of the target object according to the facial expression data, determining a target time of the expression code, for example, determining a time a, a time b and a time c when the expression code corresponding to the happy object a is 3 according to the facial expression data of the object a, then determining face orientation data at the target time from the face orientation data by adopting face posture detection, and then determining a concerned thermal area corresponding to the face orientation data and a first dish set corresponding to the concerned thermal area. Still taking the above example as an example, determining that the face of the object a faces the upper left corner of the current click list page p at the time a, and the face of the object a faces the lower right corner of the current click list page p at the time b from the face posture data, correspondingly determining the thermal area of interest i corresponding to the upper left corner of the current click list page p and the dish h1 corresponding to the thermal area of interest i, and the dish h2 corresponding to the thermal area of interest j corresponding to the lower right corner of the current click list page p, and thus obtaining the first dish set corresponding to all the thermal areas of interest including the dish h1 and the dish h2. The dishes of each ordering page in different areas can be preset, and in this way, the dishes corresponding to the thermal areas can be paid attention to on the corresponding ordering pages. Then, the first dish set is used as the current ordering preference of the target object. The current point list preference is generated by analyzing the facial expression data and the facial gesture data of the target object, which change in real time in the point list process, and the generated current point list preference is closer to the actual point list requirement of the target object, so that the point list recommendation efficiency is improved. And the whole current ordering process can be carried out under the condition that the target object is completely unaware, so that the use experience of the ordering user is ensured. In addition, in practical application, the facial expression can be quantized through the expression codes, and the speed of facial expression recognition is improved.
In the embodiment of the present invention, as shown in fig. 3, step S104: when determining that the historical point list data of the target object exist, generating the historical point list preference of the target object according to the historical point list data, wherein the historical point list preference comprises the following steps:
s301: determining a historical ordering record of the target object according to the historical ordering data;
s302: determining a second dish set and the number of times of ordering each dish from the historical ordering records;
s303: and generating historical ordering preference of the target object according to the second dish set and the ordering times.
In the specific implementation process, the specific implementation process from step S301 to step S303 is as follows:
firstly, determining a historical ordering record of the target object according to the historical ordering data, then determining a second dish set and ordering times of various dishes from the historical ordering record, for example, if it is detected that the ordering user B is an old user, the historical ordering data can be obtained, determining a corresponding historical ordering record, and determining that the corresponding dish set comprises a dish s1, a dish s2 and a dish s3 from the historical ordering record, wherein the ordering times of the dish s1 is 1, the ordering times of the dish s2 is 3, and the ordering times of the dish s3 is 5. Then, a historical ordering preference of the target object is generated according to the second menu set and the ordering times, and still taking the above example as an example, the historical ordering preference of the ordering user B, which prefers the menu s3, can be determined by counting the menu sets and the ordering times of the various menus in the historical ordering record of the ordering user B. And generating the historical ordering preference according to the historical ordering data of the target object, so that menu recommendation can be performed on the target object according to the historical ordering preference of the target object, and the ordering recommendation effect is improved.
In the embodiment of the present invention, as shown in fig. 4, step S105: according to a preset rule, obtaining a dish to be recommended according to the historical ordering preference and the current ordering preference, and displaying the dish to be recommended to the target object, wherein the method comprises the following steps:
s401: respectively setting the weight of the historical ordering preference and the weight of the current ordering preference according to the dish sales condition;
s402: determining the weight of the dish to be recommended according to the historical ordering preference and the weight corresponding to the current ordering preference;
s403: and displaying the dishes to be recommended to the target object according to the sequence of the weights from big to small.
In the specific implementation process, the specific implementation process from step S401 to step S403 is as follows:
first, the weights of the historical ordering preference and the current ordering preference are respectively set according to the sales condition of the dishes, for example, according to the statistics of the sales condition of the dishes, a newly-pushed menu is more popular and is better to sell, the weight of the current ordering preference can be set to 1, and the weight of the historical ordering preference can be set to 0.5. Of course, the weight of the historical ordering preference and the weight of the current ordering preference may also be adjusted according to the actual dish sales situation, which is not limited herein. And then, determining the weight of the dish to be recommended according to the historical ordering preference and the weight corresponding to the current ordering preference. Still taking the current order preference with weight of 1 and the historical order preference with weight of 0.5 as an example, the first menu set of the current order preference with weights of c1 and c2 includes dishes, and the second menu set of the historical order preference with weights of c1 and c3 includes dishes, so that the dish c1 with weight of 1.5, the dish c2 with weight of 1, and the dish c3 with weight of 0.5. After the weight of the dish to be recommended is determined, displaying the dish to be recommended to the target object according to the sequence of the weights from large to small. Still taking the above example as an example, c1, c2, and c3 are displayed to the target object in descending order of weight. The dish recommendation method has the advantages that the dish recommendation effect is improved because the dish recommendation method can combine the historical ordering preference and the current ordering preference together to recommend dishes to be recommended according to the dish sales condition.
In the embodiment of the present invention, as shown in fig. 5, step S102: performing face attribute recognition on at least a partial image of the face video data including a target object to obtain face expression data and face posture data of the target object, including:
s501: performing feature extraction on at least partial image including a target object in the face video data through a deep neural network to determine face expression data of the target object;
s502: extracting at least one key point of the target object from each frame of the at least partial image;
s503: and determining the facial pose data of the target object according to the at least one key point.
In the specific implementation process, the specific implementation process of step S501 to step S503 is as follows:
firstly, feature extraction is carried out on at least partial image including a target object in the face video data through a deep neural network, and the facial expression data of the target object are determined. The deep neural network can extract and classify the features of at least partial images, output a feature vector with the size of 4, and output four outputs respectively representing anger, normality, happiness and surprise, so that the recognition of the facial expression data of the target object in each frame of image of at least partial images can be realized, and the real-time facial expression data of the target object can be determined. In addition, face key point extraction may be further used to extract at least one key point of the target object from each frame of the at least partial image, for example, to extract left eye, right eye, nose, left mouth corner, and right mouth corner of the target object from the image p1 in the at least partial image. Then, according to the at least one key point, face pose data of the target object is determined. The method can be characterized in that the face pose is calculated by using matrix operation, the face pose is expressed as deflection angles in the x direction, the y direction and the z direction, the three deflection angles are linearly amplified and converted into point coordinates in a two-dimensional space, and therefore the position of the face of the target object facing a single point screen can be determined, and dishes in corresponding positions can be further determined.
In the embodiment of the present invention, as shown in fig. 6, in step S101: after acquiring the face video data comprising at least one frame of image, the method further comprises:
s601: carrying out face detection on each frame of image in the at least partial images, and screening out images of which the face definition and the face shielding degree meet the preset quality;
s602: and taking the image which meets the preset quality in the at least partial image as an image to be subjected to face attribute identification.
In the specific implementation process, the specific implementation process from step S601 to step S602 is as follows:
firstly, face detection is carried out on each frame of image in at least partial image, images with face definition and face shielding degree meeting preset quality are screened out, face position and size can be extracted from each frame of image by adopting face detection technology, namely, the face part in each frame of image is cut out, then face definition detection and face shielding detection are carried out, images with face definition and face shielding degree meeting preset quality are screened out from the cut faces, blurred and shielded images can be filtered out from the cut images, so that the images with higher face definition and without shielding are screened out and serve as images meeting preset conditions, then the images meeting the preset quality in at least partial image are used as images to be subjected to face attribute identification, and the accuracy of face attribute identification is ensured because face attribute identification is carried out on the basis of the images meeting the preset quality in at least partial image including the target object. For example, the face sharpness is divided into five levels, i.e., 0, 1, 2, 3, and 4, from low to high, and accordingly, the image with the clearest quality is obtained when the face sharpness is 4. The face shielding degree can be divided into two categories of shielding and non-shielding, when the face definition is lower than 4 or the face is shielded in the face attribute identification process, the frame of image is automatically skipped, the face attribute identification is carried out on the next frame, and the face attribute identification is carried out on the frame of image as long as the face definition is 4 and is not shielded, so that the face attribute identification speed is increased, and the face attribute identification precision is ensured.
In the implementation process, in step S602: after the image satisfying the preset quality in the at least partial image is taken as the image to be subjected to face attribute recognition, the image to be subjected to face attribute recognition can be subjected to face identity recognition by machine learning, firstly, a convolutional neural network is used for extracting the face image features of a preset image set comprising a plurality of frames of images, for example, a feature vector with the size of 512 is extracted, and the feature vector is stored in a face library. When an image to be identified by the face attribute is input, feature extraction is carried out on the input image to obtain a corresponding feature vector, and Euclidean distance calculation is carried out on the feature vector and each feature vector in a face library, so that the best matching of the current face to be identified is obtained, and the identity identification of a corresponding object is realized. For example, for an input image to be subjected to face attribute recognition, firstly, feature extraction is performed on the input image by using a convolutional neural network, distance calculation is performed on the extracted features and face features stored in a face library in advance, for example, cosine distance is performed, a set threshold value is 0.95, when the similarity of the face features is higher than the set threshold value, a target object corresponding to the current face features is judged to be an old user, otherwise, the target object is judged to be a new user, when the target object is the old user, historical ordering data of the target object can be searched in an ordering information database so as to determine historical ordering preference of the target object, and when the target object is the new user, current ordering data information of the user can be stored in an ordering information database after a current ordering flow is finished, and the current ordering data information is bound with the face features of the user, so that more accurate dish recommendation can be performed on the user later.
Fig. 7 is a general flowchart of a method for ordering a list according to an embodiment of the present invention, and specific implementation of each step is described in detail in the foregoing, and is not described herein again. Because the face images meeting the preset quality can be screened out in the face detection, the face recognition and the face attribute recognition are carried out, the historical ordering data of the target object is determined when the target object is an old user, and the current ordering preference of the target object is determined in real time according to the face expression data and the face posture data of the target object, so that the dish recommendation is carried out on the target object by combining the historical ordering preference and the current ordering preference of the target object, therefore, the dish recommendation not only considers the current ordering preference of the target object, but also considers the historical ordering preference, and the dish recommendation effect is improved.
In the embodiment of the present invention, in step S101: after acquiring the face video data comprising at least one frame of image, the method further comprises the following steps:
and if the target object is detected to be placed in order, stopping recommending the dishes to be recommended to the target object.
In a specific implementation process, after face video data including at least one frame of image is acquired, if ordering of the target object is detected to be completed, recommending of dishes to be recommended to the target object is stopped, which may be stopping recommending of the dishes to be recommended to the target object if an operation of pressing an order button on an order menu by the target object is detected, or stopping recommending of the dishes to be recommended to the target object if a browsing action of the target object is detected to be completed, for example, if the face image of the target object is not acquired within a preset time period. Therefore, invalid dish recommendation is avoided, and dish recommendation efficiency is improved. Of course, the ordering of the target object may also be determined according to the actual application requirement, which is not limited herein.
In the specific implementation process, the images in the face video data are subjected to algorithm processing such as face detection, face recognition, face attribute recognition and the like, face frames, face key points and output results of the attribute recognition algorithms can be transmitted among the algorithms, and the recording and inputting of the algorithm results can be realized by utilizing a preset face structure, so that the communication among the algorithms is realized, and the dish recommending speed is further improved. In addition, due to the fact that algorithms are communicated with one another in the dish recommending process, the dish recommending method is small in calculation amount and low in parameter amount, the dish recommending method can be applied to edge intelligent equipment, and dish recommending can be performed quickly even for the edge intelligent equipment with low calculation force.
Based on the same inventive concept, as shown in fig. 8, an embodiment of the present invention further provides a ordering recommendation system, including:
the system comprises an image acquisition module 10, a face attribute recognition module 20 and a processing module 30;
the image acquisition module 10 is configured to acquire face video data including at least one frame of image;
the face attribute recognition module 20 is configured to perform face attribute recognition on at least a partial image of the face video data that includes a target object, obtain facial expression data and facial pose data of the target object, and send the facial expression data and the facial pose data to the processing module;
the processing module 30 is configured to generate a current order preference of the target object according to the facial expression data and the facial pose data, generate a historical order preference of the target object according to the historical order data when it is determined that historical order data of the target object exists, obtain a dish to be recommended according to the current order preference and the historical order preference according to a preset rule, and display the dish to be recommended to the target object.
In an embodiment of the present invention, the processing module 30 is configured to:
determining an expression code for representing a target expression of the target object according to the facial expression data;
determining target time corresponding to the expression codes;
determining face orientation data corresponding to the target time from the face posture data;
determining a heat power area of interest corresponding to the face orientation data and a first dish set corresponding to the heat power area of interest;
and taking the first dish set as the current ordering preference of the target object.
In an embodiment of the present invention, the processing module 30 is configured to:
determining a historical ordering record of the target object according to the historical ordering data;
determining a second dish set and the number of times of ordering each dish from the historical ordering records;
and generating historical ordering preference of the target object according to the second dish set and the ordering times.
In an embodiment of the present invention, the processing module 30 is configured to:
respectively setting the weights of the historical ordering preference and the current ordering preference according to the sales condition of the dishes;
determining the weight of the dish to be recommended according to the historical ordering preference and the weight corresponding to the current ordering preference;
the dish recommendation system further comprises a display module, wherein the display module is used for:
and displaying the dishes to be recommended to the target object according to the sequence of the weights from large to small.
In the embodiment of the present invention, the face attribute identification module 20 is configured to:
performing feature extraction on at least part of images including a target object in the face video data through a deep neural network to determine facial expression data of the target object;
extracting at least one key point of the target object from each frame of the at least partial image;
and determining the facial pose data of the target object according to the at least one key point.
In this embodiment of the present invention, after the image obtaining module 10 obtains the face video data including at least one frame of image, the apparatus further includes a face detection module, where the face detection module is configured to:
carrying out face detection on each frame of image in the at least partial images, and screening out images of which the face definition and the face shielding degree meet the preset quality;
and taking the image which meets the preset quality in the at least partial image as an image to be subjected to face attribute identification.
In this embodiment of the present invention, after the image obtaining module 10 obtains the face video data including at least one frame of image, the processing module 30 is further configured to:
and if the target object is detected to be placed in order, stopping recommending the dishes to be recommended to the target object.
In a specific implementation process, the processing module 30 and the image acquisition module 10, the face attribute identification module 20, the face detection module may communicate with each other through a Transmission Control Protocol (TCP), and may encapsulate data into a structural body, and transmit the structural body between a server and a client using a socket tool, thereby ensuring the usability of the ordering recommendation system.
In addition, the principle of solving the problem of the ordering recommendation system is similar to that of the ordering recommendation method, so the implementation of the ordering recommendation system can refer to the implementation of the ordering recommendation method, and repeated parts are not described again.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.
Claims (11)
1. A method for recommending dishes, comprising:
acquiring human face video data comprising at least one frame of image;
performing face attribute recognition on at least partial image including a target object in the face video data to obtain facial expression data and facial posture data of the target object;
generating a current click list preference of the target object according to the facial expression data and the facial pose data;
when determining that the historical point list data of the target object exist, generating historical point list preference of the target object according to the historical point list data;
and according to a preset rule, obtaining a dish to be recommended according to the historical ordering preference and the current ordering preference, and displaying the dish to be recommended to the target object.
2. The method of claim 1, wherein generating the current click-to-order preference for the target object based on the facial expression data and the facial pose data comprises:
determining an expression code for representing the target expression of the target object according to the facial expression data;
determining a target moment corresponding to the expression code;
determining face orientation data corresponding to the target time from the face posture data;
determining a heat area of interest corresponding to the face orientation data and a first dish set corresponding to the heat area of interest;
and taking the first dish set as the current ordering preference of the target object.
3. The method of claim 2, wherein the generating historical point ticket preferences for the target object from the historical point ticket data upon determining that the historical point ticket data for the target object exists comprises:
determining a historical ordering record of the target object according to the historical ordering data;
determining a second dish set and the number of times of ordering each dish from the historical ordering records;
and generating historical ordering preference of the target object according to the second dish set and the ordering times.
4. The method of claim 3, wherein the obtaining a dish to be recommended according to the historical ordering preference and the current ordering preference according to a preset rule and displaying the dish to be recommended to the target object comprises:
respectively setting the weight of the historical ordering preference and the weight of the current ordering preference according to the dish sales condition;
determining the weight of the dish to be recommended according to the historical ordering preference and the weight corresponding to the current ordering preference;
and displaying the dishes to be recommended to the target object according to the sequence of the weights from large to small.
5. The method of claim 1, wherein performing face attribute recognition on at least a portion of the image in the facial video data that includes a target object to obtain facial expression data and facial pose data of the target object comprises:
performing feature extraction on at least part of images including a target object in the face video data through a deep neural network to determine facial expression data of the target object;
extracting at least one key point of the target object from each frame of the at least partial image;
and determining the facial pose data of the target object according to the at least one key point.
6. The method of any one of claims 1-5, wherein after the obtaining video data of a human face comprising at least one frame of image, the method further comprises:
carrying out face detection on each frame of image in the at least partial image, and screening out an image of which the face definition and the face shielding degree meet the preset quality;
and taking the image which meets the preset quality in the at least partial image as an image to be subjected to face attribute identification.
7. The method of any one of claims 1-5, wherein after the obtaining video data of a human face comprising at least one frame of image, the method further comprises:
and if the target object is detected to be placed in order, stopping recommending the dishes to be recommended to the target object.
8. A dish recommendation system, comprising:
the system comprises an image acquisition module, a face attribute identification module and a processing module;
the image acquisition module is used for acquiring face video data comprising at least one frame of image;
the face attribute recognition module is used for carrying out face attribute recognition on at least part of the image of the face video data including the target object to obtain face expression data and face posture data of the target object, and sending the face expression data and the face posture data to the processing module;
the processing module is used for generating current ordering preference of the target object according to the facial expression data and the facial gesture data, generating historical ordering preference of the target object according to the historical ordering preference when the historical ordering preference of the target object is determined to exist, obtaining dishes to be recommended according to the current ordering preference and the historical ordering preference according to a preset rule, and displaying the dishes to be recommended to the target object.
9. The system of claim 8, wherein the processing module is to:
determining an expression code for representing a target expression of the target object according to the facial expression data;
determining target time corresponding to the expression codes;
determining face orientation data corresponding to the target time from the face posture data;
determining a heat area of interest corresponding to the face orientation data and a first dish set corresponding to the heat area of interest;
and taking the first dish set as the current ordering preference of the target object.
10. The system of claim 9, wherein the processing module is to:
determining a historical ordering record of the target object according to the historical ordering data;
determining a second dish set and the ordering times of each dish from the historical ordering records;
and generating historical ordering preference of the target object according to the second dish set and the ordering times.
11. The system of claim 10, wherein the processing module is to:
respectively setting the weights of the historical ordering preference and the current ordering preference according to the sales condition of the dishes;
determining the weight of the dish to be recommended according to the historical ordering preference and the weight corresponding to the current ordering preference;
the dish recommendation system further comprises a display module, and the display module is used for:
and displaying the dishes to be recommended to the target object according to the sequence of the weights from large to small.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110691095.6A CN115587829A (en) | 2021-06-22 | 2021-06-22 | Dish recommendation method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110691095.6A CN115587829A (en) | 2021-06-22 | 2021-06-22 | Dish recommendation method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115587829A true CN115587829A (en) | 2023-01-10 |
Family
ID=84772559
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110691095.6A Pending CN115587829A (en) | 2021-06-22 | 2021-06-22 | Dish recommendation method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115587829A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117541359A (en) * | 2024-01-04 | 2024-02-09 | 江西工业贸易职业技术学院(江西省粮食干部学校、江西省粮食职工中等专业学校) | Dining recommendation method and system based on preference analysis |
-
2021
- 2021-06-22 CN CN202110691095.6A patent/CN115587829A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117541359A (en) * | 2024-01-04 | 2024-02-09 | 江西工业贸易职业技术学院(江西省粮食干部学校、江西省粮食职工中等专业学校) | Dining recommendation method and system based on preference analysis |
| CN117541359B (en) * | 2024-01-04 | 2024-03-29 | 江西工业贸易职业技术学院(江西省粮食干部学校、江西省粮食职工中等专业学校) | Dining recommendation method and system based on preference analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR102293008B1 (en) | Information display method, device, and system | |
| US10979624B2 (en) | Methods and devices for establishing photographing template database and providing photographing recommendation information | |
| CN105426850B (en) | Associated information pushing device and method based on face recognition | |
| CA3111097C (en) | Bounding box doubling as redaction boundary | |
| CN110476182A (en) | Share the automatic suggestion of image | |
| CN110321845B (en) | Method and device for extracting emotion packets from video and electronic equipment | |
| CN110741377A (en) | Face image processing method and device, storage medium and electronic equipment | |
| CN116097320A (en) | System and method for improved facial attribute classification and use thereof | |
| US7003140B2 (en) | System and method of searching for image data in a storage medium | |
| KR102455966B1 (en) | Mediating Apparatus, Method and Computer Readable Recording Medium Thereof | |
| US20190332854A1 (en) | Hybrid deep learning method for recognizing facial expressions | |
| US11710347B2 (en) | Information processing apparatus, information processing method, and program | |
| CN107018330A (en) | A kind of guidance method and device of taking pictures in real time | |
| CN110648170A (en) | Article recommendation method and related device | |
| CN115331150A (en) | Image recognition method, image recognition device, electronic equipment and storage medium | |
| US11308313B2 (en) | Hybrid deep learning method for recognizing facial expressions | |
| CN115587829A (en) | Dish recommendation method and system | |
| CN110543813A (en) | Face image and gaze counting method and system based on scene | |
| CN107845025A (en) | The method and device of article in a kind of recommendation video | |
| KR102577604B1 (en) | Japanese bar menu recommendation system based on artificial intelligence | |
| JP2023026630A (en) | Information processing system, information processing apparatus, information processing method, and program | |
| CN112101479B (en) | Hair style identification method and device | |
| CN112492389B (en) | Video pushing method, video playing method, computer device and storage medium | |
| Anjan et al. | Fashion recommendation system using CNN | |
| JP6210554B2 (en) | Recognition device, recognition program, and recognition method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |