Interactive guest broadcasting interaction method based on intelligent glasses

Abstract

The invention provides an interactive guest playing interaction method based on intelligent glasses, and belongs to the technical field of interaction between intelligent equipment and guest playing, and the method comprises the steps: responding to the received first interaction information, generating first text information based on the first interaction information, and sending the first text information to second equipment, the first text information is used for indicating the second equipment to generate first broadcast text information matched with the first text information based on the target AI large model; and in response to the received first broadcaster text information sent by the second equipment, converting the first broadcaster text information into first broadcaster voice information, and playing the first broadcaster voice information. According to the interactive guest broadcasting interaction method based on the intelligent glasses, the interactivity and individuation can be improved, and the requirement of a user for obtaining specific information in real time is met.

Description

Technical Field

[0001] The present disclosure belongs to the technical field of interaction between smart devices and podcast playback, and more specifically, to an interactive podcast interaction method based on smart glasses. Background Art

[0002] Traditional podcasts are usually pre-recorded audio content, and users need to select the content they are interested in on their mobile phones before listening. Therefore, existing interactive methods lack interactivity and personalization, and cannot meet users' needs for obtaining specific information in real time. Summary of the invention

[0003] The purpose of the present disclosure is to provide an interactive podcast interaction method based on smart glasses to improve interactivity and personalization and meet the user's needs for real-time acquisition of specific information.

[0004] According to a first aspect of an embodiment of the present disclosure, there is provided an interactive podcast interaction method based on smart glasses, which is applied to a first device and includes: In response to receiving the first interaction information, generating first text information based on the first interaction information, and sending the first text information to the second device, wherein the first text information is used to instruct the second device to generate first podcast text information matching the first text information based on the target AI big model; In response to receiving the first podcast text information sent by the second device, converting the first podcast text information into first podcast voice information, and playing the first podcast voice information; Among them, the second device is a device that has established a connection relationship with the first device, the first interaction information includes voice information of the first user, and the first user is a user wearing the first device.

[0005] According to a second aspect of the embodiments of the present disclosure, there is provided an interactive podcast interaction device based on smart glasses, which is applied to a first device and includes: A first podcast text information module, configured to generate first text information based on the first interaction information in response to receiving the first interaction information, and send the first text information to the second device, wherein the first text information is used to instruct the second device to generate first podcast text information matching the first text information based on the target AI large model; The first podcast voice information module is used to convert the first podcast text information into first podcast voice information and play the first podcast voice information in response to receiving the first podcast text information sent by the second device; wherein the second device is a device that has established a connection relationship with the first device, the first interaction information includes the voice information of the first user, and the first user is a user wearing the first device.

[0006] A third aspect of an embodiment of the present disclosure provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the above-mentioned interactive method of an interactive podcast based on smart glasses when executing the computer program.

[0007] According to a fourth aspect of the embodiments of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned interactive method of an interactive podcast based on smart glasses are implemented.

[0008] The beneficial effects of an interactive podcast interaction method based on smart glasses provided by the embodiments of the present disclosure are: The disclosed embodiments greatly enhance the personalization and immediacy of podcast content. Users can trigger content generation through voice, so that podcast content can be dynamically adjusted according to users' immediate needs or interests, enhancing the flexibility and interactivity of user experience.

[0009] On the one hand, using smart glasses as an interactive medium breaks the one-way reception mode of traditional podcasts, realizes interactive content consumption anytime and anywhere, and adds convenience and fun to fast-paced life.

[0010] On the other hand, with the help of the powerful capabilities of the target AI big model, the generated podcast content is high-quality and highly relevant, which can meet the diverse information needs of users, while also promoting the innovative application and popularization of AI technology in daily life.

[0011] Therefore, this embodiment improves interactivity and personalization, and meets the user's needs for obtaining specific information in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0013] Figure 1 A schematic flow chart of an interactive podcasting method based on smart glasses provided in one embodiment of the present disclosure; Figure 2 A schematic diagram of a signaling interaction process between devices provided in an embodiment of the present disclosure; Figure 3 A structural block diagram of an interactive device for interactive podcasts based on smart glasses provided in one embodiment of the present disclosure; Figure 4A schematic block diagram of an electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0014] In the following description, specific details such as specific system structures and technologies are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present disclosure. However, it should be clear to those skilled in the art that the present disclosure may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obstructing the description of the present disclosure with unnecessary details.

[0015] In order to make the purpose, technical solutions and advantages of the present disclosure more clear, specific embodiments will be described below in conjunction with the accompanying drawings.

[0016] Please refer to Figure 1 , Figure 1 A flowchart of an interactive podcast interaction method based on smart glasses is provided for an embodiment of the present disclosure. The method is applied to a first device and may include S101 and S102.

[0017] S101: In response to receiving first interaction information, generate first text information based on the first interaction information, and send the first text information to a second device, wherein the first text information is used to instruct the second device to generate first podcast text information matching the first text information based on a target AI big model.

[0018] In this embodiment, the first device may be a wearable device that can realize instant and direct information interaction with the user, and the first user is a user wearing the first device. For example, the first device may be smart glasses that have the functions of collecting and playing voice and communicating and interacting with other devices.

[0019] The first interaction information includes voice information of the first user, for example, the first user speaks out the podcast topic of interest through voice.

[0020] Exemplarily, the first user speaks out the podcast topic of interest through voice, such as "explain Journey to the West", and the smart glasses can collect the first user's voice information through the built-in microphone, that is, the first device receives the first interaction information.

[0021] The first device includes a microphone, a speaker, a processor, a wireless communication module (such as BLE or WIFI), a speech recognition (Automatic Speech Recognition, ASR) module, a text to speech (Text To Speech, TTS) module, a background noise detection and voice activity detection module (Voice activity detection, VAD).

[0022] After the first device responds to the received first interaction information, the ASR module of the first device converts the first interaction information into first text information, and the first text information can be sent to the second device through a wireless communication module (such as BLE or WIFI, etc.).

[0023] The second device is a device that has established a connection relationship with the first device, and is responsible for receiving the first text message of the first device and performing subsequent processing. For example, the second device can be a cloud-based AI large model server, which has the functions of receiving text messages, understanding and expanding text messages, and generating podcast information.

[0024] In this embodiment, the first interaction information may be converted into the first text information by using an ASR module, wherein the ASR module may have a built-in natural language processing technology.

[0025] Specifically, the step of converting the first interaction information into the first text information may be as follows: First, speech signal acquisition and preprocessing.

[0026] Voice signal collection: The voice signal of the first user is collected through the built-in microphone of the smart glasses as the first interaction information.

[0027] Preprocessing: Since the voice signal is a continuous analog signal, it contains the voice characteristics of the first user and various environmental noises, etc. Therefore, it is necessary to preprocess the collected voice signal.

[0028] For example, the speech signal may be subjected to noise reduction (eg, filtering) to remove or reduce the interference of environmental noise on the speech signal and improve the clarity of the speech.

[0029] Alternatively, pre-emphasis can be performed to enhance the high-frequency part of the speech signal and compensate for the high-frequency attenuation of the speech signal during transmission, making the spectrum of the speech signal flatter, which is beneficial for subsequent feature extraction.

[0030] Second, feature extraction is performed on the speech signal, that is, parameters that can represent the speech features are extracted from each frame of the speech signal.

[0031] For example, the Mel-Frequency Cepstral Coefficients (MFCCs) of the speech signal can be extracted. The Mel-Frequency Cepstral Coefficients simulate the human ear's perception characteristics of sounds of different frequencies and can effectively represent the spectral characteristics of speech.

[0032] In this embodiment, the frequency spectrum can be obtained by performing fast Fourier transform on the speech signal, and then the MFCCs feature can be obtained by performing processing such as Mel filter and logarithmic operation.

[0033] Third, the processed first interaction information may be sequentially input into a pre-built acoustic model and a language model to decode the first interaction information to obtain the first text information.

[0034] The acoustic model can be used to calculate the probability relationship between speech features and phonemes (or other speech units). For example, the acoustic model can be a Hidden Markov Model (HMM). The HMM model regards the speech signal as being generated by a series of hidden states, each state corresponding to a phoneme or subphoneme. In this embodiment, the parameters of the HMM can be determined in advance by training a large amount of speech data, thereby obtaining a trained HMM model.

[0035] The language model is used to calculate the probability of a word sequence appearing. The language model takes into account the grammatical, semantic, and contextual information of the language, and can further constrain and optimize the phoneme sequence output by the acoustic model to obtain more accurate text results. For example, the language model uses deep neural networks to learn the relationship between words, which can better capture the long-range dependencies and semantic information of the language.

[0036] Exemplarily, after the smart glasses receive the first interactive information (such as "Explain Journey to the West"), the ASR module can convert the first user's voice information into the first text information (such as "Explain Journey to the West"), that is, the first interactive information generates the first text information. Then, the first text information (such as "Explain Journey to the West") is sent to the cloud AI big model server via WIFI, that is, the first text information is sent to the second device.

[0037] The first text information is used to instruct the second device to generate first podcast text information that matches the first text information based on the target AI big model. The second device generates the first podcast text information (detailed explanation of Journey to the West) that matches the first text information based on the target AI big model. For example, the target AI big model can be a pre-trained Generative Adversarial Network (GAN), which understands and expands the content of the first text information, thereby generating the first podcast text information that matches it.

[0038] Specifically, the steps of converting the first interactive information into the first text information using the pre-trained GAN model may be as follows: First, data preparation.

[0039] First, we need to collect a large amount of speech data and corresponding text data pairs as the basic data set for training the GAN model. It should cover a variety of speakers, accents, speaking speeds, environmental noise, etc. to ensure that the model can learn a wide range of speech features and text correspondences.

[0040] Second, GAN model architecture design.

[0041] The generator converts the input first interactive information (speech signal) into the first text information. It is usually composed of neural network layers, such as the Convolutional Neural Networks (CNN) layer for extracting the features of the speech signal and the Recurrent Neural Network (RNN) layer for processing sequence data and generating text sequences. After receiving the speech signal, the generator converts it into an intermediate representation by performing spectral analysis, feature extraction and other operations on the speech signal. Then, the RNN is used to gradually generate text characters, one character at a time, until the complete first text information is generated.

[0042] The discriminator determines whether the generated text is real. It receives the text generated by the generator and the real text data as input, and outputs a probability value indicating the possibility that the input text is real text.

[0043] The discriminator is also composed of a neural network, usually including a multilayer perceptron (MLP), etc. The authenticity of the text is judged by analyzing the grammar, semantics, and statistical features of the text.

[0044] Third, model training.

[0045] Alternately train the generator and discriminator: First, the parameters of the discriminator are fixed and the generator is trained. The goal of the generator is to generate text that is as realistic as possible, making it difficult for the discriminator to distinguish between the generated text and the real text. The parameters of the generator are updated by minimizing the difference between the generated text and the real text, for example, using the cross entropy loss function.

[0046] Then, the parameters of the generator are fixed and the discriminator is trained. The goal of the discriminator is to accurately distinguish between real text and text generated by the generator. The parameters of the discriminator are updated by maximizing the probability of correct discrimination, that is, maximizing the probability that the real text is judged to be real and the probability that the generated text is judged to be fake.

[0047] This process is repeated many times until the generator and the discriminator reach a balance, where the generator is able to generate sufficiently realistic text, while the discriminator has difficulty distinguishing between the generated text and the real text.

[0048] Introducing adversarial loss: In the training process, in addition to using traditional loss functions (such as cross entropy loss), adversarial loss is also introduced. Adversarial loss measures the adversarial relationship between the generator and the discriminator, prompting the generator to generate more realistic text to deceive the discriminator, while also prompting the discriminator to judge the authenticity of the text more accurately.

[0049] Fourth, text generation.

[0050] After training, when the first device receives the first interaction information (voice signal), it is input into the generator. The generator generates the first text information based on the learned correspondence between voice and text. The generated text can undergo post-processing steps, such as removing unnecessary punctuation marks, correcting spelling errors, etc., to improve the readability and accuracy of the text.

[0051] Exemplarily, the first interactive information may be a question or the starting point of a topic discussion. The generated first text information may be a further elaboration of the question or a guiding description of the topic. After the second device receives the first text information, the target AI big model may generate a detailed podcast script, including the host's opening remarks, the guest's point of view, case analysis, etc., to meet the podcast needs indicated by the first text information.

[0052] S102: In response to receiving the first podcast text information sent by the second device, converting the first podcast text information into first podcast voice information, and playing the first podcast voice information; Among them, the second device is a device that has established a connection relationship with the first device, the first interaction information includes voice information of the first user, and the first user is a user wearing the first device.

[0053] In this embodiment, the first device receives the first podcast text information (detailed explanation content about Journey to the West) sent by the second device, and then the first device converts the first podcast text information into first podcast voice information, which can use the TTS module.

[0054] Exemplarily, the text content is converted into a playable voice signal by calling the TTS module. In this process, the voice parameters can be adjusted according to the needs and preferences of the first user, such as selecting different voice timbres, speech speeds, pitches, etc., to provide a more personalized voice playback experience.

[0055] Then, the first device plays the first podcast voice information through a speaker of the first device or a connected wireless headset or other device, so that the first user wearing the first device can listen to the podcast content about the topic of interest to him in a timely manner.

[0056] For example, if the first interaction information includes the voice information of the first user, it means that the first user expresses his needs through voice, triggering the entire podcast generation and playback process. As the subject wearing the first device, the first user can enjoy a convenient and personalized podcast service, without manual operation, and can obtain the required information and entertainment content only through voice interaction.

[0057] For example, when the first user says the voice message "Explain Journey to the West", the first device converts it into a first text message and sends it to the second device, and the second device generates a first podcast text message about the explanation of Journey to the West and sends it back to the first device. The first device then converts the first text message into a first podcast voice message and plays it to the first user, so that the first user can listen to the explanation of Journey to the West anytime and anywhere.

[0058] It can be concluded from the above that this embodiment greatly improves the personalization and immediacy of podcast content. Users can trigger content generation through voice, so that podcast content can be dynamically adjusted according to users' immediate needs or interests, enhancing the flexibility and interactivity of user experience.

[0059] On the one hand, using smart glasses as an interactive medium breaks the one-way reception mode of traditional podcasts, realizes interactive content consumption anytime and anywhere, and adds convenience and fun to fast-paced life.

[0060] On the other hand, with the help of the powerful capabilities of the target AI big model, the generated podcast content is high-quality and highly relevant, which can meet the diverse information needs of users, while also promoting the innovative application and popularization of AI technology in daily life.

[0061] Therefore, this embodiment improves interactivity and personalization, and meets the user's needs for obtaining specific information in real time.

[0062] In one embodiment of the present disclosure, after playing the first podcast voice information, the method further includes: In response to receiving the second interaction information, generating second text information based on the second interaction information, and sending the second text information to the second device, wherein the second text information is used to instruct the second device to expand the first podcast text information based on the target AI big model and the second text information to obtain the second podcast text information; In response to receiving the second podcast text information sent by the second device, converting the second podcast text information into second podcast voice information, and playing the second podcast voice information; The second interaction information includes voice information of the first user.

[0063] In this embodiment, the second device can extract multiple first keywords from the second text information, and extract multiple second keywords from the first podcast text information. Then, the relevance of the multiple first keywords and the multiple second keywords is calculated respectively, and multiple second keywords related to each first keyword (with a relevance greater than a preset relevance) are selected from the multiple second keywords as multiple target keywords.

[0064] First, define the first keyword set as , the second keyword set is .

[0065] Define the correlation function , indicating the first keyword With the second keyword The default correlation is .

[0066] First, calculate the importance weight of the keywords.

[0067] For keywords in the first keyword set , its importance weight It can be calculated by the following formula:

[0068] in, Indicates keywords The frequency of the word in the second text information, Indicates keywords the inverse document frequency in the entire corpus, and is an adjustable weight coefficient, and .

[0069] For keywords in the second keyword set , its importance weight It can be calculated by the following formula:

[0070] in, Indicates keywords The frequency of the word in the second text message, Indicates keywords the inverse document frequency in the entire corpus, and is an adjustable weight coefficient, and .

[0071] Second, calculate the correlation function.

[0072] The relevance is calculated by combining the vector space model with semantic similarity.

[0073] Suppose the word vector of the keyword is represented as , the semantic similarity function is , a pre-trained language model can be used to calculate the semantic similarity of two keywords.

[0074] The correlation function is defined as:

[0075] in, and is an adjustable weight coefficient, and .

[0076] Third, determine the target keyword set.

[0077] The calculation formula for the target keyword set is as follows:

[0078] This embodiment comprehensively considers multiple factors such as keyword frequency, inverse document frequency, word vector similarity, and semantic similarity, and can more accurately determine the target keyword. In actual applications, the weight coefficients and preset relevance can be adjusted according to the specific situation. , in order to achieve the best results.

[0079] Multiple first keywords and multiple target keywords are input into the target AI big model to expand the first podcast text information to obtain second podcast voice information that matches the second interaction information.

[0080] After obtaining the second podcast voice information, the second device may send the second podcast voice information to the first device.

[0081] The second interaction information includes the voice information of the first user. At this time, the voice information of the first user is a further question to the first interaction information, such as "help me explain Sun Wukong in depth", which is based on the first interaction information (explain Journey to the West).

[0082] While playing the first podcast voice message, the first device (smart glasses) will dynamically respond and update based on further interaction of the first user.

[0083] When the first device is playing the first podcast voice information, if the first user provides the second interaction information, the second interaction information is also received by the first device in the form of voice. The first device will process the second interaction information in the same way as the first interaction information. First, the second text information is generated based on the second interaction information. This process can still use ASR technology, including voice signal acquisition and preprocessing, feature extraction, pre-building of acoustic models and language models, etc., to convert the first user's voice information into accurate second text content.

[0084] Then, the first device sends the second text information to the second device. After receiving the second text information, the second device will expand the first podcast text information based on the target AI big model (such as a pre-trained GAN model) and the second text information. The target AI big model will further mine relevant content based on the new second text information and the previously generated first podcast text information, conduct more in-depth analysis and expansion, and thus obtain the second podcast text information. In this process, the target AI big model will fully consider the context of the first podcast text information and the new needs and directions provided by the second text information to generate a newer second podcast text information.

[0085] Next, when the first device receives the second podcast text message sent by the second device, it converts it into the second podcast voice message. This step is similar to converting the first podcast text message into the first podcast voice message, and the new second podcast text message is converted into the playable second podcast voice message through the TTS module.

[0086] Finally, the first device plays the second podcast voice information so that the first user can listen to the expanded podcast content.

[0087] For example, when the first user is listening to the first podcast voice information about "Journey to the West", he says "Help me explain Sun Wukong in depth" as the second interactive information. The first device converts it into the second text information and sends it to the second device. The second device uses the target AI big model to expand the description of Sun Wukong's personality characteristics based on the first podcast text information about "Journey to the West", generates the second podcast text information and sends it back to the first device. The first device then converts it into the second podcast voice information and plays it to the first user.

[0088] From the above, it can be concluded that this interactive podcast interaction method further enhances the coherence and depth of the first user experience. The first user can immediately provide feedback and guide the expansion of the podcast content, achieving a natural transition from a single podcast to in-depth conversational content. The dynamic content expansion not only satisfies the first user's curiosity to explore the unknown, but also promotes a more personalized and immersive information acquisition process, making the podcast experience richer and more interesting.

[0089] In one embodiment of the present disclosure, before playing the second podcast voice information, the method further includes: Stop playing the first podcast voice message.

[0090] In this embodiment, before playing the second podcast voice information, stopping playing the first podcast voice information can avoid confusion and interference caused by the simultaneous playing of voices of different contents. If the first podcast voice information and the second podcast voice information are played at the same time, the first user will find it difficult to hear any one of the contents, which greatly affects the efficiency and experience of the first user in obtaining information. By stopping playing the first podcast voice information, the first user can focus on the new extended content, that is, the second podcast voice information, ensuring that the first user can clearly receive and understand the current podcast content.

[0091] When the second interaction information triggers the expansion of the first podcast text information and generates the second podcast text information, the first device needs to make it clear that the current playback focus should be shifted to the new content. Stopping the playback of the first podcast voice information is a clear signal switching method, indicating that the first device is responding to the new first user demand and is ready to play the second podcast voice information.

[0092] For example, when the first user is listening to the first podcast voice message about "Journey to the West", the second interactive message is sent to learn more about Sun Wukong. At this time, stopping the first podcast voice message can divert the first user's attention from the previous explanation content and better prepare to receive the second podcast voice message about Sun Wukong's explanation, thereby achieving a more smooth and coherent interactive experience.

[0093] It can be concluded from the above that this embodiment ensures the smoothness of content switching, avoids information overlap, allows the first user to focus on the current podcast content, and improves the coherence and clarity of the overall listening experience.

[0094] In one embodiment of the present disclosure, after playing the second podcast voice information, the method further includes: In response to completing the playing of the second podcast voice information, outputting a first voice prompt, the first voice prompt being used to ask the user whether to continue playing the first podcast voice information; In response to receiving the third interaction information, generating third text information based on the third interaction information, and sending the third text information to the second device, where the third text information is used to instruct the second device to expand the second podcast text information based on the target AI big model and the third text information to obtain the third podcast text information; In response to receiving the third podcast text information sent by the second device, converting the third podcast text information into third podcast voice information, and playing the third podcast voice information; In response to receiving the fourth interaction information, continue playing the first podcast voice information.

[0095] In this embodiment, when the second podcast voice information is played, the first device outputs a first voice prompt, the purpose of which is to ask the first user whether to continue playing the first podcast voice information. For example, the voice prompt may be "Do you want to continue playing the previous explanation content about Journey to the West?".

[0096] Then, if the first user responds to the voice prompt, that is, "Talk about the personality characteristics of Sun Wukong", the third interactive information is generated. The first device generates a third text message based on the third interactive information and sends it to the second device. The second device expands the second podcast text information according to the target AI big model and the third text information to generate the third podcast text information. This process is similar to the expansion of the first podcast text information. The target AI big model will conduct in-depth analysis and expansion based on the new third text information and the existing second podcast text information to meet the further needs of the first user.

[0097] Then, after receiving the third podcast text information, the first device converts it into the third podcast voice information and plays it, providing the user with new extended content, allowing the user to continue to obtain richer information.

[0098] On the other hand, if the user sends a fourth interactive message to continue playing the first podcast voice message after hearing the voice prompt, the first device will respond to the fourth interactive message and continue playing the first podcast voice message. At this time, the first user can return to the previous content at any time to review or further think about it, which improves the flexibility and adaptability of the first device.

[0099] For example, after playing the second podcast voice information about Sun Wukong's explanation, the first device sends a voice prompt to ask the first user whether to continue playing the explanation content about "Journey to the West". If the first user answers "yes", the first device continues to play the first podcast voice information; if the first user raises a new question or demand as the third interactive information, the first device will expand and play the third podcast voice information accordingly.

[0100] From the above, it can be concluded that after playing the second podcast voice information, the first user is interacted with through voice prompts, which improves the flexibility of the first user's participation. Allowing users to choose whether to continue the first podcast or explore the third podcast content extended by the target AI large model not only meets personalized needs, but also enriches the listening experience and increases the diversity of podcasts.

[0101] In one embodiment of the present disclosure, when playing the first podcast voice information, it also includes: In response to receiving the target voice information, determining whether the target voice information matches the first user; If the target voice information does not match the first user, controlling the first device to be in a conversation mode; Wherein, when the first device is in the conversation mode, the first device stops playing the podcast voice information.

[0102] In this embodiment, when the first device receives the target voice information during the process of playing the first podcast voice information, the first device will determine whether the target voice information matches the first user. The matching process can be performed in a variety of ways, such as analyzing the voice's timbre, pitch, speaking speed and other characteristics, or identifying the speaker's identity information through specific ASR technology. If the judgment result shows that the target voice information does not match the first user, it means that someone else is talking to the first user at this time, and the first device automatically enters the conversation mode.

[0103] When the first device is in conversation mode, the first device stops playing the podcast voice information, preventing the podcast voice from interfering with the current conversation and ensuring that the conversation can proceed clearly. Turning on the conversation mode allows the first device to better adapt to different usage scenarios, improving the flexibility and practicality of the first device.

[0104] For example, when the first user is playing the first podcast voice information about Journey to the West through smart glasses, the person next to him suddenly speaks. After receiving the target voice information, the first device makes a judgment and finds that it does not match the first user, so it immediately enters the conversation mode and stops playing the podcast voice information so that the first user can focus on communicating with the person next to him. After the conversation is over, the first user can restart the podcast as needed.

[0105] It can be concluded from the above that when playing the first podcast voice information, by identifying the matching degree between the target voice information and the user, it intelligently switches to the conversation mode, effectively avoiding the interference of the voice information and ensuring smooth communication between users. This embodiment not only improves the user experience, but also demonstrates the first device's keen perception and flexible response capabilities to the first user's situation, creating a more humane and convenient use environment for the first user. At the same time, it also enhances the applicability of the first device in a variety of usage scenarios.

[0106] In one embodiment of the present disclosure, a method for interactive podcast based on smart glasses further includes: In response to the first device being in the conversation mode, if no voice information is detected within a target duration, the first device is controlled to exit the conversation mode and continue to play the podcast voice information.

[0107] In this embodiment, when the first device is in conversation mode, it means that other people other than the first user are currently interacting with the environment where the first device is located, and the playback of the podcast voice information is paused to avoid interfering with the conversation. At this time, if no voice information is detected within the target duration, it means that the surrounding conversation has ended or entered a relatively quiet stage. In order to ensure that the first user can continue to enjoy the podcast content, the first device can exit the conversation mode and continue to play the podcast voice information.

[0108] The target duration setting can be adjusted according to the actual usage scenario and the needs of the first user. For example, it can be set to tens of seconds or minutes. If the target duration is set to be shorter, it can respond more quickly to the end of the conversation, allowing the first user to return to the podcast listening state as soon as possible; if the target duration is set to be longer, it can avoid frequent switching between the conversation mode and the podcast playback mode due to short silences, thereby improving the stability of the first device.

[0109] For example, the first user is using smart glasses to listen to a podcast, and suddenly someone talks to the first user, and the smart glasses enter the conversation mode and stop playing the podcast voice information. If no voice is detected within the next minute, the smart glasses will automatically exit the conversation mode and continue to play the previously paused podcast voice information, allowing the first user to continue to enjoy the podcast content without manual operation.

[0110] From the above, we can conclude that the introduction of the function of automatically exiting the conversation mode significantly improves the consistency of the first user experience. When there is no voice activity in the conversation mode within the target duration, it automatically ends and resumes the podcast playback, which not only avoids unnecessary waiting, but also ensures that the first user can seamlessly connect to the original podcast content. Not only does it reduce the manual operation of the first user, but it also demonstrates the first device's deep understanding and accurate response to the first user's behavior pattern, further enhancing the smoothness and convenience of the podcast experience.

[0111] In one embodiment of the present disclosure, determining whether the target voice information matches the first user includes: Extracting multiple speech features of target speech information; Fusing multiple speech features to obtain target fusion features; Calculate the matching degree between the target fusion feature and the standard fusion feature; Based on the matching degree, it is determined whether the target voice information matches the first user.

[0112] In this embodiment, the target voice information includes multiple segments of voice information, and multiple voice features of the target voice information are extracted, including: Extract multiple segments of speech features of target segment speech information, where the target segment speech information is a segment of speech information among multiple segments of speech information.

[0113] Specifically, the target voice information can be divided into multiple voice information segments according to the preset duration. For each voice information segment, the information amount of the voice information segment is calculated. If the information amount of the voice information segment is less than the preset information amount, the voice information segment is marked as invalid voice information. If the information amount of the voice information segment is greater than or equal to the preset information amount, the voice information segment is marked as valid voice information.

[0114] Exemplarily, the information amount of each segment of speech information is calculated based on speech feature statistics, and the steps are as follows: First, spectrum feature analysis.

[0115] Perform fast Fourier transform (FFT) on each piece of speech information to obtain its spectrum.

[0116] The entropy value of the spectrum is calculated as a measure of the amount of information. The spectrum entropy reflects the uniformity of the signal distribution at different frequencies. The more uniform the distribution, the greater the entropy value, and the smaller the amount of information; the more concentrated the distribution, the smaller the entropy value, and the larger the amount of information.

[0117] For example, for a relatively single-frequency speech signal, its spectrum will be concentrated near a specific frequency, and the entropy value will be small, indicating that the amount of information is large, indicating that specific speech content is conveyed, such as a clear word or phrase.

[0118] Second, time domain feature analysis.

[0119] Calculate the zero-crossing rate of the speech signal. The zero-crossing rate refers to the number of times the signal passes through the zero value per unit time. Speech signals with a higher zero-crossing rate usually contain more high-frequency components and may have a relatively large amount of information.

[0120] By analyzing the energy distribution of speech signals, we can calculate the energy of speech signals in different time periods and observe their changes. The parts with larger energy changes contain more information.

[0121] In this embodiment, the multiple voice features of the target voice information are extracted as valid voice information. The target voice information can be a voice information among multiple valid voice information, and the preset information amount can be set according to the actual situation. For example, according to the information amount of the voice information in the historical conversation scene of the first user, multiple valid voice information can be calculated and then averaged.

[0122] From the above, we can conclude that by carefully dividing the speech information and evaluating its information content, invalid speech segments are effectively eliminated, and valid speech segments rich in information are retained, which significantly improves the accuracy and efficiency of speech processing.

[0123] Accordingly, multiple speech features are fused to obtain target fusion features, including: Fusing the multiple speech features corresponding to the target segment speech information to obtain the target segment fusion feature; The target fusion feature is determined based on all target segment fusion features.

[0124] In this embodiment, multiple segment speech features corresponding to the target segment speech information are fused to obtain the target segment fusion feature: Determine the feature weights of each segment of speech features.

[0125] The speech features of each segment are weighted and calculated based on their feature weights to obtain the fusion features of the target segment.

[0126] Multiple segments of speech features may include: timbre, intonation, and speech speed. First, the dimensions of each speech feature are unified. For example, standardization is performed as follows: (1) For numerical features (such as speech rate), calculate the mean and standard deviation of the feature in all speech data. For each speech rate value, standardize it using the formula "(original value - mean) / standard deviation" so that the value of the speech rate feature is distributed in a specific range, usually an interval with a mean of 0 and a standard deviation of 1. Therefore, speech rate values ​​in different ranges can be unified to the same scale.

[0127] (2) For the quantification of non-numerical features (such as timbre), timbre is a relatively complex feature that can be converted into numerical representation through some feature extraction methods. For example, the Mel-frequency cepstral coefficients (MFCCs) method can be used to extract the numerical vector of timbre features.

[0128] Then, the extracted timbre feature vectors are standardized in a similar way to the numerical features. The mean and standard deviation are calculated and standardized so that different timbre feature vectors are also within the same scale range.

[0129] (3) For the encoding of categorical features (such as intonation), if intonation can be divided into several different types, one-hot encoding can be used to convert it into a numerical vector. For example, if there are three types of intonation, they are represented by vectors [1, 0, 0], [0, 1, 0], and [0, 0, 1] respectively. Therefore, different intonation types are converted into numerical vectors of uniform dimension, which can be further processed with other features.

[0130] Optionally, the first formula may be used to calculate the target segment fusion feature.

[0131] The first formula is:

[0132] in, is the target segment fusion feature, For the The weight of the speech feature of each segment, For the Segment speech features, is the number of speech features in the segment.

[0133] This embodiment can unify the vector dimension of each target segment fusion feature. For example, the features of timbre, intonation, and speech speed are represented as vectors , , , which can be combined into a new vector ,in , , The dimensions are , , 1. Among them, it is necessary to ensure that different features have reasonable positions and orders in the combined vector for subsequent processing and analysis.

[0134] In this embodiment, after all target fusion features are obtained, the target fusion features can be calculated using the second formula.

[0135] The second formula is:

[0136] in, is the target fusion feature, For the The target segment fusion features, For the The weight of the fusion feature of each target segment, is the number of target segment features. It can be determined based on the information amount of the target segment speech corresponding to the target segment speech feature.

[0137]

[0138] Among them, a <b<c, For the The information content of the speech features of the target segment, is the first information quantity, is the second amount of information.

[0139] For example, a, b, and c are 0.2, 0.6, and 1.2 respectively, and the greater the amount of information, the greater the weight.

[0140] In this embodiment, the standard fusion feature can be calculated based on the voice information of the first user within a certain period of time, and the calculation method adopted is the same as the above-mentioned calculation method. For example, by adopting the cosine similarity calculation method, the similarity between the target segment fusion feature and the standard fusion feature is calculated as the matching degree between the two.

[0141] The specific calculation method is as follows: First, calculate the dot product. Calculate the dot product of the target segment fusion feature vector and the standard fusion feature vector. The dot product reflects the sum of the products of two vectors in each dimension and can measure the common features of two vectors in the same dimension. If two vectors have similar values ​​in some dimensions, their dot product will be large, indicating a high correlation in these dimensions.

[0142] Second, calculate the modulus. Calculate the modulus of the target segment fusion feature vector and the standard fusion feature vector. The modulus represents the length of the vector, which reflects the comprehensive size of the vector in each dimension. The modulus can be calculated by summing the squares of the values ​​of each dimension of the vector and then taking the square root.

[0143] Third, calculate the similarity. Divide the dot product by the product of the modulus lengths of the two vectors to get the cosine similarity value. This value ranges from -1 to 1, where 1 means the two vectors are exactly the same, 0 means the two vectors are completely unrelated, and -1 means the two vectors are completely opposite.

[0144] In the application of speech feature matching, a higher cosine similarity value indicates a higher match between the target segment fusion features and the standard fusion features, that is, the two speech segments have a higher similarity in timbre, intonation, speaking speed, etc.

[0145] In this embodiment, judging whether the target voice information matches the first user based on the matching degree may include: If the matching degree is greater than or equal to the first matching degree, it is determined that the target voice information matches the first user.

[0146] If the matching degree is less than the first matching degree, it is determined that the target voice information does not match the first user.

[0147] It can be concluded from the above that this embodiment significantly improves the accuracy and personalization of speech recognition. Through multi-dimensional feature fusion, such as timbre, intonation and speaking speed, the comprehensiveness and representativeness of speech features are ensured. By using feature weight calculation, the influence of important features is enhanced and the effectiveness of feature fusion is improved. At the same time, the matching method based on cosine similarity can quickly and accurately evaluate the similarity between speech segments and effectively distinguish the speech features of different users. This embodiment not only enhances the reliability of user identity recognition, but also provides strong support for personalized voice services.

[0148] refer to Figure 2 , Figure 2 A schematic diagram of a signaling interaction process between devices provided in an embodiment of the present disclosure. The signaling interaction process between the first device and the second device is as follows: A101: Receive first interaction information.

[0149] The first device receives first interaction information.

[0150] A102: Generate first text information.

[0151] The first device generates first text information based on the first interaction information.

[0152] A103: Send the first text message.

[0153] The first device sends the first text message to the second device.

[0154] A104: Generate first podcast text information.

[0155] The first text information is used to instruct the second device to generate first podcast text information that matches the first text information based on the target AI big model.

[0156] A105: Send the first podcast text message.

[0157] The second device sends the first podcast text information to the first device.

[0158] A106: Generate the first podcast voice information.

[0159] The first device converts the first podcast text information into first podcast voice information.

[0160] A107: Play the first podcast voice message.

[0161] The first device plays the first podcast voice information.

[0162] A108: Receive second interaction information.

[0163] The first device receives second interaction information.

[0164] A109: Generate second text information.

[0165] The first device generates second text information based on the second interaction information.

[0166] A110: Send the second text message.

[0167] The first device sends the second text message to the second device.

[0168] A111: Generate second podcast text information.

[0169] The second text information is used to instruct the second device to expand the first podcast text information based on the target AI large model and the second text information to obtain the second podcast text information. A112: Send the second podcast text message.

[0170] The first device receives the second podcast text information sent by the second device.

[0171] A113: Generate second podcast voice information.

[0172] The first device converts the second podcast text information into second podcast voice information.

[0173] A114: Stop playing the first podcast voice message.

[0174] Before playing the second podcast voice information, the first device stops playing the first podcast voice information.

[0175] A115: Play the second podcast voice message.

[0176] The first device plays the second podcast voice information.

[0177] A116: The second podcast voice message has been played.

[0178] The first device finishes playing the second podcast voice message.

[0179] A117: Output the first voice prompt.

[0180] The first device outputs a first voice prompt, where the first voice prompt is used to ask the user whether to continue playing the first podcast voice information.

[0181] A118: Receive third interaction information.

[0182] The first device receives third interaction information.

[0183] A119: Generate third text information.

[0184] The first device generates third text information based on the third interaction information.

[0185] A120: Send the third text message.

[0186] The first device sends the third text information to the second device.

[0187] A121: Generate third podcast text information.

[0188] The third text information is used to instruct the second device to expand the second podcast text information based on the target AI large model and the third text information to obtain the third podcast text information. A122: Send a third podcast text message.

[0189] The first device receives the third podcast text information sent by the second device.

[0190] A123: Generate third podcast voice information.

[0191] The first device converts the third podcast text information into third podcast voice information.

[0192] A124: Play the third podcast voice message.

[0193] The first device plays the third podcast voice information.

[0194] A125: Receive fourth interaction information.

[0195] The first device receives fourth interaction information.

[0196] A126: Continue playing the first podcast voice message.

[0197] The first device continues to play the first podcast voice information.

[0198] Corresponding to the above embodiment, an interactive podcast interaction method based on smart glasses, Figure 3 This is a block diagram of a structure of an interactive device for interactive podcasts based on smart glasses provided by an embodiment of the present disclosure. For ease of description, only the parts related to the embodiment of the present disclosure are shown. Figure 3 The interactive device 20 of the interactive podcast based on smart glasses is applied to the first device, including: a first podcast text information module 201 and a first podcast voice information module 202.

[0199] The first podcast text information module 201 is used to generate first text information based on the first interaction information in response to receiving the first interaction information, and send the first text information to the second device, where the first text information is used to instruct the second device to generate first podcast text information matching the first text information based on the target AI big model; The first podcast voice information module 202 is used to convert the first podcast text information into first podcast voice information and play the first podcast voice information in response to receiving the first podcast text information sent by the second device; wherein the second device is a device that has established a connection relationship with the first device, the first interaction information includes the voice information of the first user, and the first user is a user wearing the first device.

[0200] In one embodiment of the present disclosure, an interactive podcasting device 20 based on smart glasses further includes: a second podcasting text information module and a second podcasting voice information module; A second podcast text information module, configured to generate second text information based on the second interaction information in response to receiving the second interaction information, and send the second text information to the second device, wherein the second text information is used to instruct the second device to expand the first podcast text information based on the target AI big model and the second text information to obtain the second podcast text information; The second podcast voice information module is used to convert the second podcast text information into second podcast voice information and play the second podcast voice information in response to receiving the second podcast text information sent by the second device; the second interactive information includes the voice information of the first user.

[0201] In one embodiment of the present disclosure, an interactive podcasting device 20 based on smart glasses further includes: a stop module; The stop module is used to stop playing the first podcast voice information.

[0202] In one embodiment of the present disclosure, an interactive podcasting device 20 based on smart glasses further includes: a first voice prompt module, a third podcast text information module, a third podcast voice information module, and a continue playing module; A first voice prompt module, configured to output a first voice prompt in response to completion of playing the second podcast voice information, wherein the first voice prompt is used to ask the user whether to continue playing the first podcast voice information; A third podcast text information module, configured to generate third text information based on the third interaction information in response to receiving the third interaction information, and send the third text information to the second device, wherein the third text information is used to instruct the second device to expand the second podcast text information based on the target AI big model and the third text information to obtain the third podcast text information; A third podcast voice information module, configured to, in response to receiving the third podcast text information sent by the second device, convert the third podcast text information into third podcast voice information and play the third podcast voice information; The continue playing module is used to continue playing the first podcast voice information in response to receiving the fourth interaction information.

[0203] In one embodiment of the present disclosure, the first podcast voice information module 202 is specifically used to: In response to receiving the target voice information, determining whether the target voice information matches the first user; If the target voice information does not match the first user, controlling the first device to be in a conversation mode; Wherein, when the first device is in the conversation mode, the first device stops playing the podcast voice information.

[0204] In one embodiment of the present disclosure, the first podcast voice information module 202 is further used to: In response to the first device being in the conversation mode, if no voice information is detected within a target duration, the first device is controlled to exit the conversation mode and continue to play the podcast voice information.

[0205] In one embodiment of the present disclosure, the first podcast voice information module 202 is further configured to: Extracting multiple speech features of target speech information; Fusing multiple speech features to obtain target fusion features; Calculate the matching degree between the target fusion feature and the standard fusion feature; Based on the matching degree, it is determined whether the target voice information matches the first user.

[0206] See also Figure 4 , Figure 4 A schematic block diagram of an electronic device provided by an embodiment of the present disclosure. Figure 4 The electronic device 300 in the embodiment shown may include: one or more processors 301, one or more input devices 302, one or more output devices 303 and one or more memories 304. The processors 301, input devices 302, output devices 303 and memories 304 communicate with each other via a communication bus 305. The memory 304 is used to store computer programs, which include program instructions. The processor 301 is used to execute the program instructions stored in the memory 304. The processor 301 is configured to call the program instructions to execute the functions of each module / unit in the above-mentioned device embodiments, such as Figure 3 The functions of modules 21 to 22 are shown.

[0207] It should be understood that in the embodiment of the present disclosure, the processor 301 may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0208] The input device 302 may include a touch panel, a fingerprint collection sensor (for collecting the user's fingerprint information and fingerprint direction information), a microphone, etc., and the output device 303 may include a display (LCD, etc.), a speaker, etc.

[0209] The memory 304 may include a read-only memory and a random access memory, and provide instructions and data to the processor 301. A portion of the memory 304 may also include a non-volatile random access memory. For example, the memory 304 may also store information about the device type.

[0210] In a specific implementation, the processor 301, input device 302, and output device 303 described in the embodiments of the present disclosure can execute the implementation methods described in the first and second embodiments of an interactive podcast interaction method based on smart glasses provided in the embodiments of the present disclosure, and can also execute the implementation methods of the electronic device described in the embodiments of the present disclosure, which will not be repeated here.

[0211] In another embodiment of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by the processor, all or part of the processes in the above-mentioned embodiment method are implemented, and the computer program can also be completed by instructing the relevant hardware through the computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by the processor, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0212] The computer-readable storage medium may be an internal storage unit of the electronic device of any of the aforementioned embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium may also be an external storage device of the electronic device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Furthermore, the computer-readable storage medium may also include both an internal storage unit of the electronic device and an external storage device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium may also be used to temporarily store data that has been output or is to be output.

[0213] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this disclosure.

[0214] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the electronic devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0215] In the several embodiments provided in the present application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces or units, or it can be an electrical, mechanical or other form of connection.

[0216] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present disclosure.

[0217] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0218] The above are only specific embodiments of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present disclosure, and these modifications or replacements should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.

Claims

1. An interactive podcast interaction method based on smart glasses, characterized in that: Applied to a first device, comprising: In response to receiving the first interaction information, generating first text information based on the first interaction information, and sending the first text information to the second device, wherein the first text information is used to instruct the second device to generate first podcast text information matching the first text information based on the target AI big model; In response to receiving the first podcast text information sent by the second device, converting the first podcast text information into first podcast voice information, and playing the first podcast voice information; Among them, the second device is a device that has established a connection relationship with the first device, the first interaction information includes voice information of the first user, and the first user is a user wearing the first device.

2. The interactive podcast interaction method based on smart glasses as claimed in claim 1, characterized in that: After playing the first podcast voice information, the method further includes: In response to receiving the second interaction information, generating second text information based on the second interaction information, and sending the second text information to the second device, wherein the second text information is used to instruct the second device to expand the first podcast text information based on the target AI big model and the second text information to obtain the second podcast text information; In response to receiving the second podcast text information sent by the second device, converting the second podcast text information into second podcast voice information, and playing the second podcast voice information; The second interaction information includes voice information of the first user.

3. The interactive podcast interaction method based on smart glasses as claimed in claim 2, characterized in that: Before playing the second podcast voice information, the method further includes: Stop playing the first podcast voice message.

4. The interactive podcast interaction method based on smart glasses as claimed in claim 3, characterized in that: After playing the second podcast voice information, the method further includes: In response to completion of playing the second podcast voice information, outputting a first voice prompt, wherein the first voice prompt is used to ask the user whether to continue playing the first podcast voice information; In response to receiving the third interaction information, generating third text information based on the third interaction information, and sending the third text information to the second device, wherein the third text information is used to instruct the second device to expand the second podcast text information based on the target AI big model and the third text information to obtain third podcast text information; In response to receiving the third podcast text information sent by the second device, converting the third podcast text information into third podcast voice information, and playing the third podcast voice information; In response to receiving the fourth interaction information, continue playing the first podcast voice information.

5. The interactive podcast interaction method based on smart glasses as claimed in claim 1, characterized in that: When playing the first podcast voice message, it also includes: In response to receiving the target voice information, determining whether the target voice information matches the first user; If the target voice information does not match the first user, controlling the first device to be in a conversation mode; Wherein, when the first device is in conversation mode, the first device stops playing the podcast voice information.

6. The interactive podcast interaction method based on smart glasses as claimed in claim 5, characterized in that: Also includes: In response to the first device being in the conversation mode, if no voice information is detected within a target duration, the first device is controlled to exit the conversation mode and continue to play the podcast voice information.

7. The interactive podcast interaction method based on smart glasses as claimed in claim 5, characterized in that: The determining whether the target voice information matches the first user includes: Extracting multiple speech features of the target speech information; Performing feature fusion on the multiple speech features to obtain a target fusion feature; Calculating the matching degree between the target fusion feature and the standard fusion feature; Based on the matching degree, it is determined whether the target voice information matches the first user.

8. An interactive podcast device based on smart glasses, characterized in that: Applied to a first device, comprising: A first podcast text information module, configured to generate first text information based on the first interaction information in response to receiving the first interaction information, and send the first text information to a second device, wherein the first text information is used to instruct the second device to generate first podcast text information matching the first text information based on a target AI big model; The first podcast voice information module is used to convert the first podcast text information into first podcast voice information and play the first podcast voice information in response to receiving the first podcast text information sent by the second device; wherein the second device is a device that has established a connection relationship with the first device, the first interaction information includes the voice information of the first user, and the first user is a user wearing the first device.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images