Diet record information prediction method and device based on voice dialogue

Abstract

The invention relates to the technical field of diet record information prediction, and discloses a diet record information prediction method and device based on voice dialogue. The diet record information prediction model is further utilized to accurately predict a target food name, a target food unit and a target food quantizer related to the current diet dialogue of the target user, the target food unit is a metric food unit, different types of user diet units can be unified, and the user diet dialogue accuracy is improved. And subsequent unified comparison of diet record information of different users is facilitated.

Description

Technical Field

[0001] The present invention relates to the technical field of diet record information prediction, and in particular to a diet record information prediction method and device based on voice dialogue. Background Art

[0002] Food dietary record information can help people track daily food intake, nutrients and calories. Therefore, dietary record information is important information for people's daily health management of diet.

[0003] In the related art, after collecting the user's voice conversation information, the food in the conversation information is searched for related diet record information. Since this method is affected by the user's non-standard voice and environmental noise, the relevant food pronunciation is wrong or the voice is interfered, which ultimately leads to inaccurate diet record information recognition results. At the same time, because the user does not understand some localized food names and names them incorrectly during the conversation, it will also lead to errors in recognition results. In addition, due to differences in eating habits among different types of user groups, that is, there will be differences in the daily use of dietary measurement units, therefore, if the user's dietary measurement units are not uniformly expressed, the dietary record information recognition results will eventually be inaccurate. Summary of the invention

[0004] In view of this, the present invention provides a method and device for predicting diet record information based on voice dialogue to solve the problem that traditional methods cannot accurately identify diet record information.

[0005] In a first aspect, an embodiment of the present disclosure provides a method for predicting dietary record information based on voice dialogue, the method comprising:

[0006] Obtain the target user's diet label information and the current diet text corresponding to the current diet conversation;

[0007] Extracting target label features of target food from dietary label information, and extracting target dietary features of target food from current dietary text;

[0008] The target label features and the target diet features are input into the diet record information prediction model for fusion to obtain the target prediction result corresponding to the current diet text. The diet record information prediction model is a pre-trained neural network model.

[0009] The disclosed embodiment obtains the target user's diet label information and the current diet text corresponding to the current diet conversation, and further utilizes the diet record information prediction model to accurately predict the target prediction result related to the target user's current diet conversation.

[0010] In some optional embodiments, in the method for predicting diet record information based on voice conversation in the embodiments of the present disclosure, the target label features include: user gender, user age, user region, user commonly used units, the target diet features include: food name, food unit, food measure words, the target prediction results include: target food name, target food unit, target food measure words, and the target food unit is a metric food unit.

[0011] The disclosed embodiment obtains the target user's dietary label information and the current dietary text corresponding to the current dietary conversation, and further utilizes the dietary record information prediction model to accurately predict the target food name, target food unit, and target food quantifier related to the target user's current dietary conversation, wherein the target food unit is a metric food unit, which can unify different types of user dietary units, and is conducive to the subsequent unified comparison of dietary record information of different users.

[0012] In some optional implementations, the method for predicting dietary record information based on voice conversation in the embodiment of the present disclosure further includes:

[0013] Get the current diet text corresponding to the current diet conversation of the target user;

[0014] Extracting target dietary features of target food from the current dietary text, the target dietary features including: food name, food unit, and food quantifier;

[0015] Input the target diet feature into the diet record information prediction model to obtain the current prediction result corresponding to the current diet text, and the current prediction result includes: the current food name, the current food unit, and the current food quantifier;

[0016] Determine whether the preset food unit corresponding to the current food name is found from the food library;

[0017] If yes, output the current prediction result to the target user;

[0018] If not, proceed to the step of obtaining the target user's diet label information and the current diet text corresponding to the current diet conversation.

[0019] The disclosed embodiment preferentially uses the food library to query the current food name predicted by the diet record information prediction model. When the query fails, the diet record information prediction model is used to learn the target user's dietary habit characteristics through the target user's current dietary conversation and in combination with the target user's dietary label information, and the metric food unit of the target user is obtained to prevent the inability to accurately obtain the target user's relevant dietary information.

[0020] In some optional implementations, obtaining the current diet text corresponding to the current diet conversation of the target user includes:

[0021] Get the current diet conversation of the target user;

[0022] The current diet conversation is converted into the corresponding current diet text through the speech recognition model.

[0023] The disclosed embodiment converts the current diet conversation into the corresponding current diet text through a speech recognition model in order to accurately obtain the diet conversation information of the target user.

[0024] In some optional implementations, before obtaining the current diet text corresponding to the current diet dialogue of the target user, the method further includes:

[0025] Obtain a target training text set for training a diet record information prediction model for target users.

[0026] The disclosed embodiment obtains a target training text set to ensure accurate training of a diet record information prediction model.

[0027] In some optional implementations, obtaining a target training text set for training a diet record information prediction model for a target user includes:

[0028] Perform content correction on each text in the target training text set;

[0029] Extract training text features from each text after error correction, the training text features including: food units, food names, and food quantifiers;

[0030] The diet record information prediction model is trained using the error-corrected training text features, and a training prediction result corresponding to a target training text set is generated, as well as a loss function between the training prediction result corresponding to the target training text set and the training text features. The loss function is used to update the parameter values ​​of the parameters of the diet record information prediction model.

[0031] The disclosed embodiments correct text content through a diet record information prediction model and train the diet record information prediction model in order to improve the accuracy of the diet record information prediction model.

[0032] In some optional implementations, the method for predicting dietary record information in voice conversation in the embodiments of the present disclosure further includes: displaying a target prediction result or a current prediction result.

[0033] In a second aspect, a device for predicting dietary record information based on voice dialogue in an embodiment of the present disclosure includes:

[0034] The first acquisition module is used to acquire the target user's diet label information and the current diet text corresponding to the current diet conversation;

[0035] A first extraction module is used to extract target label features of target food from diet label information, and to extract target diet features of target food from current diet text;

[0036] The first prediction module is used to input the target label features and the target diet features into the diet record information prediction model for fusion to obtain the target prediction result corresponding to the current diet text. The diet record information prediction model is a pre-trained neural network model.

[0037] In a third aspect, an embodiment of the present disclosure provides a computer device, including:

[0038] The memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method for predicting dietary record information based on voice dialogue according to the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0039] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the method for predicting dietary record information based on voice dialogue according to the first aspect or any corresponding embodiment thereof.

[0040] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, including computer instructions, which are used to enable a computer to execute the method for predicting dietary record information based on voice dialogue according to the above-mentioned first aspect or any corresponding embodiment. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0042] Figure 1 is a flow chart of a method for predicting dietary record information based on voice dialogue according to an embodiment of the present invention;

[0043] Figure 2 is a flow chart of another method for predicting dietary record information based on voice dialogue according to an embodiment of the present invention;

[0044] Figure 3 is a simplified flowchart of diet record information prediction based on voice dialogue according to an embodiment of the present invention;

[0045] Figure 4is a flow chart of a device for predicting dietary record information based on voice dialogue according to an embodiment of the present invention;

[0046] Figure 5 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0048] According to an embodiment of the present invention, an embodiment of a method for predicting dietary record information based on voice conversation is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in an order different from that shown here.

[0049] In this embodiment, a method for predicting dietary record information based on voice dialogue is provided, which can be used in mobile terminals such as mobile phones, tablet computers, etc. Figure 1 is a flow chart of a method for predicting dietary record information based on voice dialogue according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0050] Step S101, obtaining the target user's diet label information and the current diet text corresponding to the current diet conversation.

[0051] Specifically, the target user can represent any user, and the user can be a user of different professions or identities, for example, a student or a classroom or an astronaut or an athlete or some people who are losing weight. The current diet conversation represents the relevant content of the food-related conversation, for example, I had a piece of cake and some fruit loops. The diet label information is personalized information of the target user, indicating some dietary habit characteristics of the target user's daily diet.

[0052] Step S102, extracting target label features of the target food from the diet label information, and extracting target diet features of the target food from the current diet text.

[0053] In some optional implementations, the target label features include: user gender, user age, user region, user commonly used units, and the target diet features include: food name, food unit, and food quantifier.

[0054] Specifically, for example, the current diet text is I had a sandwich and a cup of juice for lunch, the first food name extracted from the current diet text is sandwich, the corresponding extracted food unit is serving, and the corresponding extracted food quantifier is 1; the second food name extracted from the current diet text is juice, the corresponding extracted food unit is cup, and the corresponding extracted food quantifier is 1. For example, the diet label information is a default user portrait (gender: female, age: 25, weight 58kg, region Beijing, common unit is g), when the diet label information is another default user portrait (gender: male, age: 35, weight 75kg, region Hebei, common unit is g).

[0055] Step S103, inputting the target label features and the target diet features into the diet record information prediction model for fusion, and obtaining the target prediction result corresponding to the current diet text. The diet record information prediction model is a pre-trained neural network model.

[0056] In some optional implementations, the target prediction result includes: target food name, target food unit, target food quantifier, and the target food unit is a metric food unit.

[0057] Specifically, the diet record information prediction model can be a third-party automatic speech recognition model, for example, it can be a GPT4o model or a Claude-Haiku model. For example: If the diet record information prediction model is regarded as a nutritionist and diet expert, planning a diet for a target user who is losing fat, the diet label information provided by the target user is: gender female, age 26, region Beijing, common unit g, and the current diet dialogue is chicken burgers and two corns for lunch (lunch is a chicken burger and 2 corns). The diet record information prediction model integrates the target label features and target diet features used by the target, learns the daily eating habits of the target user, and then obtains the target prediction result corresponding to the current diet text (target food name, target food unit, target food quantifier, and the target food unit is a metric food unit). The diet record information prediction model in the embodiment of the present disclosure can be based on the default portrait (diet label information) of the target user from different regions, combined with the current diet dialogue, and then obtains the relevant target food name, target food unit, and target food quantifier. The embodiment of the present disclosure uses metric food units to ensure that different types of users use unified units, which is conducive to accurately expressing diet record information. For example, if a 40-year-old man who is very heavy and a 20-year-old woman who is very light both say that they ate a bowl of rice, it is very likely that the rice weight of the "40-year-old man who is very heavy" is greater than that of the "20-year-old woman who is very light". Therefore, using metric food units can help to personalize and predict more appropriate metric quantities for different groups of people.

[0058] Therefore, the disclosed embodiment obtains the target user's dietary label information and the current dietary text corresponding to the current dietary conversation, and further utilizes the dietary record information prediction model to accurately predict the target food name, target food unit, and target food quantifier related to the target user's current dietary conversation, wherein the target food unit is a metric food unit, which can unify different types of user dietary units, and is conducive to the subsequent unified comparison of different users' dietary record information.

[0059] The method for predicting dietary record information based on voice dialogue provided in this embodiment can be used in mobile terminals, such as mobile phones, tablet computers, etc. Figure 2 is a flow chart of a method for predicting dietary record information based on voice dialogue according to an embodiment of the present invention. Figure 2 As shown, the process also includes the following steps:

[0060] Step S201, obtaining a current diet text corresponding to a current diet conversation of a target user.

[0061] Specifically, the target user can represent any user, and the user can be a user of different professions or identities, for example, a student or a classroom or an astronaut or an athlete or some people who are losing weight. The current diet dialogue represents the relevant content containing food-related dialogues, for example, I ate an apple this morning (I ate an apple for breakfast).

[0062] Step S202, extracting target dietary features of target food from the current dietary text, where the target dietary features include: food name, food unit, and food quantifier.

[0063] Specifically, for example, the food name extracted from the current diet text I ate an apple this morning is apple, the corresponding extracted food unit is serving, and the corresponding extracted food quantifier is 1.

[0064] Step S203, inputting the target diet feature into the diet record information prediction model for prediction, and obtaining the current prediction result corresponding to the current diet text, the current prediction result including: the current food name, the current food unit, and the current food quantifier.

[0065] For example, the current diet text is I ate an apple this morning (I ate an apple for breakfast), the food name extracted from the current diet text is apple, the corresponding extracted food unit is serving, and the corresponding extracted food quantifier is 1.

[0066] Step S204, determining whether a preset food unit corresponding to the current food name is found from the food library.

[0067] Step S205: If yes, output the current prediction result to the target user.

[0068] Specifically, by passing the current food name into the food library, the information of apple in the food library is searched, and it is found that apple supports the unit of serving. Therefore, the unit predicted by the LLM module in the diet record information prediction model is valid and can be output to the target user.

[0069] Step S206: If not, proceed to step S101 of obtaining the target user's diet label information and the current diet text corresponding to the current diet conversation.

[0070] Specifically, taking the above current diet text "I ate an apple this morning" as an example, the current food unit extracted by the LLM module (language recognition module of the speech recognition model, full name Large Language Model) is "serving". If the unit "serving" is not supported in the food library, the data will be lost and cannot be returned to the user. Therefore, the step S101 of obtaining the diet label information of the target user and the current diet text corresponding to the current diet dialogue is entered, and the target prediction result corresponding to the current diet text is re-predicted, and then the target food unit (metric food unit) corresponding to the current food name is obtained. On the basis of the above example, the metric food unit g / ml of apple is re-obtained. Generally, non-metric food units are preferred. For example, most people are accustomed to recording with one apple or two apples, and are not accustomed to recording with 50g apples or 100g apples. The prediction of metric food units and quantities complements food units and quantities. "Common measurement units used by users to add food records" is useful in some special cases. For example, if a user says "I ate apple", this sentence does not know the unit and portion of the food. Therefore, it is necessary to return a unit and an estimated portion to the user. If the unit is the unit he is accustomed to using, then he only needs to adjust the portion in the result page, which is more convenient to use. )

[0071] In some optional implementations, obtaining the current diet text corresponding to the current diet conversation of the target user in the above step includes:

[0072] Step a1, obtaining the current eating conversation of the target user.

[0073] Step a2: convert the current diet conversation into the corresponding current diet text through the speech recognition model.

[0074] Specifically, the speech recognition model may be a third-party speech recognition model, for example, a GPT4o model and a Claude-Haiku model.

[0075] In some optional implementations, before obtaining the current diet text corresponding to the current diet dialogue of the target user, the method further includes:

[0076] Obtain a target training text set for training a diet record information prediction model for target users.

[0077] Specifically, the target training text set is converted from the target training dialogue.

[0078] In some optional implementations, obtaining a target training text set for training a diet record information prediction model for a target user includes:

[0079] Step b1: Perform content correction on each text in the target training text set.

[0080] Specifically, in the embodiments of the present disclosure, a speech recognition model can be used to perform content correction on each text in the target training text set. The speech recognition model itself has an error correction function. The dietary record information prediction model is a speech recognition model, which has a keen insight into food and can detect errors in speech. It corrects the speech-transcribed text, mainly targeting incorrectly pronounced food names and portions.

[0081] For example, for text content correction, the text content input into the diet record information prediction model is I had five tumbled eggs (I ate 5 scrambled eggs), and the text tumbled has a spelling error. After the text is corrected, it becomes I had five scrambled eggs.

[0082] For example, to correct the food name in the text content, the text content input into the diet record information prediction model is Ihad a peace of cake and some froot loops, and the food name froot is misspelled. After the food name is corrected, it becomes Ihad a piece of cake and some fruit loops (I had a sandwich and a glass of juice for lunch).

[0083] For example, to correct the food quantifier in the text content, the text content input into the diet record information prediction model is To milk today, and the food quantifier To is misspelled. After the food quantifier is corrected, it becomes Two milk today (two servings of milk today).

[0084] Step b2: extracting target dietary features of target food from the current dietary text, wherein the target dietary features include: food name, food unit, and food quantifier.

[0085] Specifically, for example, the food name extracted from the corrected current diet text I ate an apple this morning (I ate an apple for breakfast) is apple, the corresponding extracted food unit is serving, and the corresponding extracted food quantifier is 1.

[0086] Step b3, using the corrected training text features to train the diet record information prediction model, and generate the training prediction results corresponding to the target training text set, as well as the loss function between the training prediction results corresponding to the target training text set and the training text features, the loss function is used to update the parameter values ​​of the parameters of the diet record information prediction model.

[0087] Specifically, a diet record information prediction model is trained and updated through a large number of training text features to ensure the accuracy of the diet record information prediction model, which is conducive to accurately predicting the diet record information of the target user.

[0088] In some optional implementations, the method for predicting diet record information based on voice dialogue in the embodiment of the present disclosure further includes: displaying a target prediction result or a current prediction result.

[0089] Specifically, the target prediction results (target food name, target food unit, target food quantifier) ​​predicted by the diet record information prediction model, or the target prediction results (current food name, current food unit, current food quantifier) ​​are displayed to prompt users to immediately know their diet situation.

[0090] like Figure 3 FIG. 1 is a simplified flowchart of the prediction of dietary record information based on voice dialogue in the embodiment of the present disclosure. Figure 3 In the example, the speech resource is the target user's dietary conversation content. The speech resource is passed to the speech recognition model (diet record information prediction model) through the automatic speech recognition interface (ASR interface, full name Automatic Speech Recognition). The speech recognition model corrects the dietary conversation content through the LLM module, extracts the food name, food unit and food quantifier after error correction, and then uses the food library to query the preset unit of the food library corresponding to the food name in the above. If it is not found, the speech recognition model (diet record information prediction model) is combined with the target user's personalized information (diet label information) to predict the food metric unit of the food, and finally outputs the target prediction result (target food name, target food unit, target food quantifier), and returns this information as a food list.

[0091] In this embodiment, a device for predicting dietary record information based on voice dialogue is also provided, which is used to implement the above-mentioned embodiments and preferred embodiments, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware for a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0092] This embodiment provides a device for predicting dietary record information based on voice dialogue, such as Figure 4 As shown, including:

[0093] The first acquisition module 401 is used to acquire the target user's diet label information and the current diet text corresponding to the current diet conversation;

[0094] A first extraction module 402 is used to extract target label features of target food from diet label information, and to extract target diet features of target food from current diet text;

[0095] The first prediction module 403 is used to input the target label features and the target diet features into the diet record information prediction model for fusion to obtain the target prediction result corresponding to the current diet text. The diet record information prediction model is a pre-trained neural network model.

[0096] In some optional embodiments, the target label features include: user gender, user age, user region, and user commonly used units; the target diet features include: food name, food unit, and food quantity words; the target prediction results include: target food name, target food unit, and target food quantity words; the target food unit is a metric food unit.

[0097] In some optional implementations, the device for predicting dietary record information based on voice dialogue in the embodiment of the present disclosure further includes:

[0098] The second acquisition module is used to acquire the current diet text corresponding to the current diet conversation of the target user;

[0099] The second extraction module extracts the target dietary features of the target food from the current dietary text, and the target dietary features include: food name, food unit, and food quantifier;

[0100] The second prediction module is used to input the target diet feature into the diet record information prediction model for prediction, and obtain the current prediction result corresponding to the current diet text, and the current prediction result includes: the current food name, the current food unit, and the current food quantifier;

[0101] An information judgment module is used to judge whether a preset food unit corresponding to the current food name is found from the food library;

[0102] A first determination module, configured to output the current prediction result to the target user if yes;

[0103] The second determination module is used to enter a module for obtaining the target user's diet label information and the current diet text corresponding to the current diet dialogue if no.

[0104] In some optional implementations, the second acquisition module includes:

[0105] The acquisition submodule is used to obtain the current diet conversation of the target user;

[0106] The conversion submodule is used to convert the current diet dialogue into the corresponding current diet text through the speech recognition model.

[0107] In some optional implementations, before the first acquisition module 401, the device further includes:

[0108] The third acquisition module is used to acquire a target training text set for training a diet record information prediction model for a target user.

[0109] In some optional implementations, the third acquisition module includes:

[0110] The error correction submodule is used to perform content error correction on each text in the target training text set;

[0111] The extraction submodule is used to extract training text features from each text after error correction, and the training text features include: food units, food names, and food quantifiers;

[0112] The training submodule is used to train the diet record information prediction model using the error-corrected training text features, and generate the training prediction results corresponding to the target training text set, as well as the loss function between the training prediction results corresponding to the target training text set and the training text features. The loss function is used to update the parameter values ​​of the parameters of the diet record information prediction model.

[0113] In some optional implementations, the method for predicting dietary record information in voice conversation in the embodiments of the present disclosure further includes: displaying a target prediction result or a current prediction result.

[0114] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0115] The diet record information prediction device based on voice conversation in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0116] An embodiment of the present invention further provides a computer device having the above-mentioned diet record information prediction device based on voice dialogue.

[0117] See also Figure 5 , Figure 5 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Figure 5As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 5 A processor 10 is taken as an example.

[0118] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0119] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0120] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0121] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0122] The computer device also includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means. Figure 5 The example of connecting through bus is taken in the following.

[0123] The input device 30 can receive input digital or character information, and generate key signal input related to the user settings and function control of the computer device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator bar, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display and a plasma display. In some optional embodiments, the display device can be a touch screen.

[0124] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0125] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.

[0126] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for predicting dietary record information based on voice dialogue, characterized in that: The method comprises: Obtain the target user's diet label information and the current diet text corresponding to the current diet conversation; Extracting target label features of target food from the diet label information, and extracting target diet features of the target food from the current diet text; The target label features and target diet features are input into a diet record information prediction model for fusion to obtain a target prediction result corresponding to the current diet text. The diet record information prediction model is a pre-trained neural network model.

2. The method according to claim 1, characterized in that The target label features include: user gender, user age, user region, and user commonly used units; the target diet features include: food name, food unit, and food quantifier; the target prediction result includes: target food name, target food unit, and target food quantifier; and the target food unit is a metric food unit.

3. The method according to claim 1 or 2, characterized in that: Also includes: Get the current diet text corresponding to the current diet conversation of the target user; Extracting target dietary features of the target food from the current dietary text, wherein the target dietary features include: food name, food unit, and food quantifier; Input the target diet feature into the diet record information prediction model for prediction, and obtain a current prediction result corresponding to the current diet text, wherein the current prediction result includes: current food name, current food unit, and current food quantifier; Determine whether a preset food unit corresponding to the current food name is found in the food library; If yes, output the current prediction result to the target user; If not, proceed to the step of obtaining the target user's diet label information and the current diet text corresponding to the current diet conversation.

4. The method according to claim 3, characterized in that The method of obtaining the current diet text corresponding to the current diet dialogue of the target user includes: Obtaining the current diet conversation of the target user; The current diet dialogue is converted into the corresponding current diet text through a speech recognition model.

5. The method according to claim 3, characterized in that: Before obtaining the current diet text corresponding to the current diet dialogue of the target user, the method further includes: Obtaining a target training text set for the target user to train the diet record information prediction model; the step of obtaining a target training text set for the target user to train the diet record information prediction model comprises: Performing content error correction on each text in the target training text set; Extracting training text features from each text after error correction, wherein the training text features include: food units, food names, and food quantifiers; The diet record information prediction model is trained using the error-corrected training text features, and a training prediction result corresponding to the target training text set and a loss function between the training prediction result corresponding to the target training text set and the training text features are generated. The loss function is used to update the parameter values ​​of the parameters of the diet record information prediction model.

6. The method according to claim 3, characterized in that Also includes: Display the target prediction result or the current prediction result.

7. A dietary record information prediction device based on voice dialogue, characterized in that: The device comprises: The first acquisition module is used to acquire the target user's diet label information and the current diet text corresponding to the current diet conversation; A first extraction module, configured to extract a target label feature of a target food from the diet label information, and to extract a target diet feature of the target food from the current diet text; The first prediction module is used to input the target label features and target diet features into a diet record information prediction model for fusion to obtain a target prediction result corresponding to the current diet text. The diet record information prediction model is a pre-trained neural network model.

8. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method for predicting dietary record information based on voice dialogue as described in any one of claims 1 to 6 by executing the computer instructions.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the method for predicting dietary record information based on voice dialogue according to any one of claims 1 to 6.

10. A computer program product, characterized in that The method comprises computer instructions for causing a computer to execute the method for predicting dietary record information based on voice dialogue according to any one of claims 1 to 6.

Images