79results about How to "Increase wake rate" patented technology

The invention discloses a voice wake-up method and apparatus, a terminal, and a processing method thereof. The voice wake-up method comprises: a wake-up word voice signal sent out by a terminal user is collected and a wake-up word identification result and a vocal print feature of the wake-up word voice signal are obtained and stored; when a terminal is in a sleep state, a voice signal inputted externally is detected; a voice identification result and a vocal print feature of the voice signal are obtained; the voice identification result and the stored wake-up word identification result as well as a matching result of the vocal print feature and the stored vocal print feature are obtained; and according to the matching result, whether the terminal needs to be woken up is determined; and if so, the terminal is woken up. With the method, technical problems of low user wake-up rate, high wrong wake-up rate, and poor user experience according to the existing voice wake-up technology are solved.

The invention discloses a wake-up word preset confidence threshold adjustment method and system. The method comprises steps that S11, daily use behavior data of a user is obtained; S12, the wake-up word preset confidence threshold is adjusted according to the daily use behavior data. The method is advantaged in that the wake-up word preset confidence threshold is dynamically set and adjusted through acquiring the daily usage behavior data of the user, personal use characteristics of the user are fully considered, firstly, whether the preset confidence threshold is lowered overall is determinedthrough the daily behavior confidence average value of the user, the preset confidence threshold is further adjusted based on the use number for each time period, the device wake-up rate is effectively improved, and the false wake-up rate is reduced.

The invention relates to a differential-training-based voice wake-up customization optimization method and system. After acoustic data with marks are given, neural network forward propagation is carried out by combining an initial neural-network-based acoustic model; a modeling probability for wake-up word marking is obtained by combining a marking sequence and a frame-by-frame acoustic probability; statistics is carried out on lots of texts to obtain a phoneme-level language model and a corresponding searching network is constructed; wake-up word competition item modeling is carried out based on the searching space of the phoneme-level language model and a frame-by-frame acoustic model, thereby obtaining a posterior probability; differential training of the acoustic model is carried out by combining the wake-up word competition item modeling and the wake-up word marking modeling; and the acoustic model obtained by the differential training determines whether to carry out a wake-up operation by comparison of a wake-up word confidence coefficient with an automatic estimation threshold value.

The invention provides a training method and using method of a voice wake-up hybrid model and related equipment; during model training, a voice wake-up hybrid model is obtained through hybrid learningtraining of a voice separation network, a feature transformation network and a wake-up word detection network. When the model is used, an audio to be recognized is input into the voice wake-up hybridmodel, and the wake-up probability is directly obtained. When the wake-up probability is greater than the threshold, the wake-up word identified from a to-be-identified audio is judged. In a model training process, a first loss function obtained through the voice separation network and a second loss function obtained through the wake-up word detection network are weighted to obtain a comprehensive loss function, and weight parameters of the voice separation network, the feature transformation network and the wake-up word detection network are obtained through back propagation and learning according to the comprehensive loss function. Meanwhile, the networks are unified into one framework, and a joint optimization method is used, so that the model can learn optimal separation and wake-up network parameters at the same time, and the wake-up rate is effectively improved.

The invention provides a far-field voice wake-up system, belonging to the field of the voice wake-up of a smart device. The problem that the wake-up identification of the smart device is easily disturbed and the recognition is inaccurate under different noise environments is solved. The technical solution is that the system includes a microphone module, a voice processing module and a wake-up engine module, the voice processing module includes multiple parallel processing modules which process voice recognition information by using different algorithms and transmit processing results to the wake-up engine module, the wake-up engine module includes multiple wake-up recognition modules corresponding to the multiple parallel processing modules, the multiple wake-up recognition modules are used for performing recognition processing of a voice wake-up command on the corresponding processing results through the wake-up recognition modules after receiving the processing results, and a smart device where the system is located can be woken up by recognized voice recognition information if at least one processing result is successfully identified.

The embodiment of the invention discloses an application awakening method and device, a storage medium and electronic equipment. The electronic equipment comprises two microphones and can collect twochannels of audio data and obtain data of background audio played during audio collecting through the two microphones; then the two channels of audio data are subjected to echo cancellation processingaccording to the background audio data so as to cancel self-noise; then the two channels of audio data subjected to echo cancellation are subjected to beam forming processing so as to cancel externalnoise, and the enhanced audio data are obtained; and then text features and voiceprint features of the enhanced audio data are subjected to two-stage verification, when two-stage verification is passed, voice interaction application is awakened, and thus voice interaction between the electronic equipment and a user is achieved. Thus, interference of the self-noise and the external noise can be removed, verification accuracy is guaranteed through two-stage verification, and the purpose of increasing the awakening rate of the voice interaction application is achieved.

The invention provides voice equipment and a control method, device and equipment thereof, and a storage medium. The method comprises the following steps of acquiring wake-up failure frequency of thevoice equipment in a preset time period, if the wake-up failure frequency is greater than or equal to a first preset frequency, reducing a wake-up threshold corresponding to wake-up information of thevoice equipment, if the wake-up failure frequency is smaller than or equal to a second preset frequency, increasing a wake-up threshold corresponding to wake-up information of the voice equipment, and if the wake-up failure frequency is less than the first preset frequency and greater than the second preset frequency, maintaining a wake-up threshold corresponding to wake-up information of the voice equipment, thereby realizing automatic adjustment of the wake-up threshold of the voice equipment in any usage scenario of the user and in the speaking volume of the user habit. By adopting the technical scheme of the invention, the wake-up rate of the voice equipment can be improved at a relatively low false wake-up rate.

The invention relates to a differential-training-based voice wake-up customization optimization method and system. After acoustic data with marks are given, neural network forward propagation is carried out by combining an initial neural-network-based acoustic model; a modeling probability for wake-up word marking is obtained by combining a marking sequence and a frame-by-frame acoustic probability; statistics is carried out on lots of texts to obtain a phoneme-level language model and a corresponding searching network is constructed; wake-up word competition item modeling is carried out based on the searching space of the phoneme-level language model and a frame-by-frame acoustic model, thereby obtaining a posterior probability; differential training of the acoustic model is carried out by combining the wake-up word competition item modeling and the wake-up word marking modeling; and the acoustic model obtained by the differential training determines whether to carry out a wake-up operation by comparison of a wake-up word confidence coefficient with an automatic estimation threshold value.

The invention relates to the field of smart home, in particular to an intelligent awakening method and system. The method comprises the steps of: acquiring illumination intensity information in a first preset time period; judging whether the illumination intensity is greater than a preset illumination threshold value or not; if so, sending an opening instruction of an intelligent lamp and an intelligent curtain; acquiring human body sleep state information in a second preset time period, with the second preset time period being after the first preset time period; judging whether the human bodyis in a light sleep state; and if so, sending an intelligent alarm clock ringing instruction. According to the scheme, before the intelligent alarm clock rings, natural light, lamplight and an intelligent mattress are used for ensuring that the human body enters the light sleep state, and the probability that the human body is awakened in the light sleep state is increased according to the alarmringing time period set by the user.

The invention relates to the technical field of intelligent loudspeakers, in particular to a computer readable storage medium and a voice interaction loudspeaker box utilizing the readable storage medium. The medium stores a computer program, and the computer program can be executed by a processor of the voice interaction loudspeaker box. According to the voice interaction loudspeaker box, reference parameters are set for correcting echo errors of the loudspeaker box at different sound volumes, the obtained corresponding relationship between an actual audio signal directly output by a main controller and the external echo collected by a microphone array is more accurate, therefore human voices can be more accurately obtained, the loudspeaker box can recognize a waking command according tothe more accurate human voices received by the loudspeaker box, and the waking rate of the loudspeaker box is increased.

The invention discloses an intelligent equipment wake-up method, system and device, electronic equipment and a storage medium. The method comprises the following steps: when a wake-up signal is received, obtaining the time when intelligent equipment enters a non-working mode; determining the wake-up priority of the intelligent equipment according to the time; taking the intelligent equipment withthe highest priority as target intelligent equipment; and sending a wake-up instruction to the target intelligent equipment to wake up the target intelligent equipment. According to the invention, when a plurality of wake-up equipment exist, the equipment with the highest priority is selected to be waked up, and only one equipment with the highest wake-up probability can be waked up at the same time.

The invention discloses a voice interaction and voice wake-up detection method and device, equipment and a storage medium. The method comprises the steps of analyzing the voice input of a user; and adjusting parameters related to the voice interaction based on an analysis result. The method can be used for voice interaction after the equipment is awakened, for example, voice characteristics such as voice speed, volume, tone and timbre of voice input of a user can be analyzed, and the emotional state of the voice input of the user can also be analyzed. Based on the analysis result, the relatedparameters of the voice output fed back to the user can be correspondingly adjusted to improve the interaction experience of the user. The method can also be used for voice wake-up detection, under the condition that a wake-up detection result is lower than a current wake-up threshold value, the parameter of the wake-up threshold value can be lowered, for example, the wake-up threshold value can be lowered in the next period of time, and therefore the wake-up success rate in the next period of time can be increased, and the wake-up experience of a user can be improved.

Embodiments of the invention disclose a voiceprint awakening method. The voiceprint awakening method comprises the following steps of: determining a preset awakening parameter matched with a current scene according to a preset condition, wherein the preset awakening parameter comprises a preset awakening threshold and a preset voiceprint threshold; receiving an input voice signal; determining theawakening value and voiceprint feature of the voice signal; and under the condition that the awakening value is greater than the preset awakening threshold and the voiceprint feature is greater than the preset voiceprint threshold, awakening a to-be-awakened application. The embodiments of the invention further provide a voiceprint awakening device, voiceprint awakening equipment and a storage medium.

The invention provides a wake-up model updating method. The method comprises the following steps of: separating a command word and a wake-up word in original voice data; determining the probability ofcommand words according to the language model, and deleting the command words with the probability lower than a first threshold value in the command words; and inputting the wake-up word into a wake-up model, determining a score output by the wake-up model, and deleting the wake-up word with the score lower than a second threshold value in the wake-up word. Therefore, the built wake-up model withhigher false wake-up probability is updated, command words and wake-up words are screened respectively, and the wake-up words are closer to preset wake-up words by eliminating false wake-up data or distortion data. Meanwhile, the command words serve as training data, the number of the training data is guaranteed, and the false awakening probability of the updated awakening model is lower.

The invention provides a voice noise reduction method and device, equipment and a medium, and belongs to the technical field of voice. According to the invention, the gain modulation information of the to-be-denoised voice is obtained according to the probability distribution of the to-be-denoised voice and the modulation gain function corresponding to each noise type, and then the to-be-denoisedvoice is denoised according to the gain modulation information. Because noise time domain characteristics, frequency domain characteristics and the like of different noise types are different, different noise reduction processes are needed; compared with the prior art in which the same noise reduction algorithm is used to reduce the noise of the to-be-denoised voice, the situation that the noise types of the noise contained in the to be denoised voice are different is considered; different gain modulation information is acquired according to different noise types of noise contained in the to-be-denoised voice, so that the denoising processes of different to-be-denoised voices are distinguished, more targeted denoising is realized, the denoising effect is improved, and the wake-up rate andthe recognition rate of the denoised voice are improved.

The embodiment of the invention discloses an audio data storage method and device, a terminal and a storage medium, and belongs to the technical field of terminals. The method comprises the followingsteps: acquiring audio data acquired by a microphone; recognizing the audio data through a k-level voice wake-up recognition model, wherein the voice wake-up recognition models of different levels correspond to different recognition dimensions, the recognition results are used for representing recognition passing conditions of the audio data on the corresponding recognition dimensions, and k is aninteger larger than or equal to 2; and storing the audio data to a first storage area according to the recognition result, wherein the first storage area is used for storing the audio data passing through the at least one stage of voice wake-up recognition model. According to the method and the device, the reason of wakeup failure, namely which level of recognition dimension fails, can be accurately positioned, so that the reason of wakeup failure can be accurately analyzed and optimized, and the wakeup rate in a voice wakeup scene is further improved.

The invention discloses a self-defined awakening method and device applied to intelligent voice equipment. The method comprises the steps of receiving a registration text set by a user and collectingregistration voice of the user; computing a registration awakening threshold based on the registration text; computing a registration awakening score of the registration voice based on a universal awakening model, and judging that whether the registration awakening score is more than the registration awakening threshold or not; if the registration awakening score is more than or equal to the registration awakening threshold, aligning the registration voice with the registration text, and extracting a voice segment corresponding to each character in the registration text; and extracting a Gaussian posterior feature of the voice segment corresponding to each character, and generating a feature template of the whole registration voice based on the Gaussian posterior feature of each voice segment. According to the scheme provided by the method and the device, the quality of input voice can be optimized in a registration stage, so that the system can be much finer in a score computing module, and finally, a mistakenly awakening rate of similar words also can be reduced while an awakening rate is improved.

The invention discloses a multi-carrier signal-based low-operand underwater acoustic wake-up method, and the method comprises the following steps: constructing a corresponding multi-carrier signal for a communication node, performing absolute value summation on signals in a window at a receiving end to perform signal arrival detection, performing multiple times of frequency point detection on the signals by using real number fast Fourier transform, and realizing the wake-up detection. The method is suitable for accurate awakening of any distance within the maximum communication distance of the two underwater acoustic nodes, is small in operand, and is suitable for a low-power-consumption single chip microcomputer. The communication machine can be in a low-power-consumption dormant state underwater for a long time.

The invention discloses an awakening performance optimization method and system based on mode switching. The method comprises the steps: acquiring awakening words; setting awakening modes, including asecondary awakening mode, a deep sleep mode and a common awakening mod, of to-be-awakened equipment according to an application scene; determining a secondary threshold value corresponding to the secondary awakening mode, a sleep threshold value corresponding to the deep sleep mode and a standard threshold value corresponding to a common awakening mode; and controlling the to-be-awakened equipment to switch among the three modes according to the awakening word, the secondary threshold, the sleep threshold and the standard threshold. Compared with a single common awakening mode, the awakeningperformance optimization method and system disclosed by the invention have the advantages that the secondary awakening mode is added, so that the awakening rate is improved; the deep sleep mode is added, so that the false awakening rate is reduced, and increasingly stringent performance indexes of a user can be met.

The invention discloses an audio and video wake-up method, system and device, and a storage medium, which introduces a video mode to improve and enhance the performance of a wake-up system, can adapt to a wake-up task in a real complex scene, improves the wake-up rate, and improves the interaction experience. moreover, aiming at the characteristic that the audio and video multi-mode wake-up data volume is relatively small, the invention provides effective information which is obtained by using a cross-mode-based teacher-student model and migrating and utilizing abundant large-data-volume single-mode acoustic data training, so system performance loss caused by relatively small multi-mode audio and video wake-up training data volume is ameliorated, and the wake-up rate is improved.

The invention provides a voice wake-up method utilizing electronic equipment in a noisy environment. The voice wake-up method utilizes the electronic equipment with n groups of voice input components,and comprises a sound recording step, a recorded sound processing step, a wake-up judging step and a voice identifying step; the wake-up judging step particularly comprises: calculating wake-up parameters according to a second signal by a set method and judging whether the wake-up parameters belong to a set wake-up set; if the wake-up parameters belong to the set wake-up set, proceeding to the voice identifying step; and if the wake-up parameters do not belong to the set wake-up set, returning to the sound recording step, wherein the second signal is a preprocessed sound signal obtained in the recorded sound processing step, n is an integer which is not smaller than 2. The voice wake-up method has the advantages of simple and convenient steps, small calculation amount and high wake-up rate, and can solve a problem of excessively low voice wake-up rate in the noisy environment or when multiple people talk and a problem of insufficient accuracy of a positioning angle of a double-microphone sound source.

The embodiment of the invention provides a voice processing method and device, user equipment and an intelligent loudspeaker box. The method comprises the steps: acquiring, by the user equipment, voice information input by a user; and according to the voice information, sending control information to a paired intelligent sound box so that the intelligent loudspeaker box outputs corresponding response information according to the control information. According to the voice processing method and device, the user equipment and the intelligent loudspeaker box, voice information input by a user canbe acquired by the user equipment and according to the voice information, the control information corresponding to the paired intelligent loudspeaker box, therefore, the intelligent loudspeaker box can output corresponding response information according to the control information. When the intelligent loudspeaker box cannot acquire clear voice information, the user equipment can be used as a relay to control the intelligent loudspeaker box according to the acquired voice information, so that the awakening rate and the voice interaction performance of the intelligent loudspeaker box are improved, and convenience is provided for a user.

The invention provides an automobile three-screen intelligent cabin host system. The system comprises a full-lamination instrument display screen, a full-lamination central control touch screen and a full-lamination co-driver touch screen which are sequentially arranged on an automobile instrument desk side by side, and further comprises a connecting wire harness and an intelligent cabin host, the connecting wire harness comprises a host end connector and a screen end connector, the full-lamination instrument display screen, the full-lamination central control touch screen and the full-lamination co-driver touch screen are connected with the screen end connector, and the intelligent cabin host is connected with the host end connector. According to the invention, a multi-screen interaction mode with a large screen, high resolution, rich display content and finer display is used, and more intelligent and more comfortable experience can be given to a user in a cross-screen multi-screen interaction mode.

The invention provides an awakening method and device for intelligent vehicle-mounted equipment. The awakening method comprises the steps that voice and light ray change of the intelligent vehicle-mounted equipment in a preset range are collected at the same time; if the voice is collected, the collected voice frequency is converted into characters through a semantic parsing engine, the charactersare matched with preset awakening words, and when the characters are the awakening words through matching, the intelligent vehicle-mounted equipment is awakened; if light ray change of the intelligent vehicle-mounted equipment in the preset range is collected, the intelligent vehicle-mounted equipment is awakened. According to the awakening method, on the basis of voice awakening, the light sensation awakening technology is added, the awakening action of the device is simple, the awakening success rate is increased, and the use experience of a user is improved.

The invention relates to a voice wake-up method and a device thereof, a storage medium and electronic equipment, and aims to design a relatively high threshold value for common wake-up words so as to reduce false wake-up and set a relatively low threshold value for rare wake-up words so as to improve the wake-up rate. The method comprises the following steps: acquiring a voice wake-up word of a to-be-set threshold; determining a word score of the voice wake-up word through a language model, wherein the word score is used for representing the occurrence probability of the voice wake-up word; according to the word score of the voice wake-up word and a preset corresponding relation between the wake-up threshold and the word score, determining a target wake-up threshold corresponding to the voice wake-up word, wherein the wake-up threshold in the preset corresponding relation is in positive correlation with the word score, and the target wake-up threshold value is used for comparing with a collected target voice wake-up word in a voice wake-up process so as to determine a voice wake-up result corresponding to the target voice wake-up word.

A voice wake-up method is applied to the field of voice interaction, and comprises the following steps: after a voice signal is received, extracting acoustic features of the voice signal, predicting posterior probabilities of the acoustic features on acoustic modeling units, and inputting the posterior probabilities into a preset evaluation model to obtain a first evaluation value, and inputting the posterior probability into a trained convolutional neural network model to obtain a second evaluation value, and judging whether to wake up or not according to the first evaluation value and the second evaluation value. The invention further discloses a voice wake-up device, electronic equipment and a storage medium. Whether wake-up words are contained in voice signals or not is judged by combining a two-stage wake-up mechanism, the influence on a result by only adopting a single model is prevented, and the wake-up rate is increased and the false wake-up rate is reduced under the conditionof ensuring low power consumption.

The embodiment of the invention discloses a voice wake-up method and device, an apparatus and a storage medium. The method comprises the following steps of acquiring the voice information in real time, and identifying an environment type corresponding to the voice information; determining a target wake-up model from a preset wake-up model group according to the environment type corresponding to the voice information, the wake-up model group comprising at least two wake-up models, and different wake-up models corresponding to different environment types; and inputting the voice information intothe target wake-up model to detect a wake-up word, and executing a device wake-up operation when it is determined that the wake-up word is detected. By means of the technical scheme, the higher wake-up rate and the lower false awakening rate can be considered at the same time in different scenes.

The invention discloses an awakening method and device of intelligent equipment, equipment and a medium, which are used for solving the problem that the awakening rate of the intelligent equipment islow because the intelligent equipment is influenced by the change of voiceprint characteristics of a target user. According to the embodiment of the invention, the method comprises the steps: obtaining a first voiceprint feature of to-be-recognized first voice information; under the condition that the first voiceprint feature is not matched with a second voiceprint feature pre-stored in a voiceprint library, determining that the voiceprint feature is not matched with the voiceprint feature, outputting default validation information, according to the method, receiving the reply information input for the verification information, and determining whether the intelligent equipment is awakened according to whether the reply information is matched with the target reply information of the presetverification information, so that the awakening process of the intelligent equipment is not influenced by voiceprint feature change of the target user, and the awakening rate of the intelligent equipment is improved.

The invention provides an interactive awakening method, device, apparatus and system for risk transaction, and the method comprises the steps: actively carrying out interaction with a transaction userof a to-be-processed risk transaction, carrying out intention recognition and dialogue recommendation, splicing determined interaction question answering information and recommended risk interactioninquiry information, and transmitting the spliced information to the transaction user; and guiding the transaction user to perform multiple rounds of interactive awakening conversations. Based on multiple rounds of conversations with transaction users, risk identification confirmation of risk transactions is realized, the risk identification accuracy of the risk transactions is improved, fraud means of the risk transactions can be mined, and a data foundation is laid for subsequent risk transaction identification. Meanwhile, the interactive question answering information and the recommended risk interactive inquiry information are spliced, so that questions of the transaction user are answered, the transaction user can be actively guided to confirm risk transaction, the transaction user isawakened in time, and the safety of network transaction and the awakening rate of risk transaction are improved.