8results about How to "Improve voice interaction experience" patented technology

The invention relates to a dialogue round end judgment method and device, electronic equipment, a medium and a vehicle. The method comprises the steps that a first effective voice segment of a to-be-recognized voice frame is acquired, and first text content corresponding to the first effective voice segment is recognized; inputting the first effective voice segment and the first text content into a preset voice recognition model, and obtaining a first confidence score output by the preset voice recognition model; the first confidence score is used for representing the probability that the user does not speak a sentence according to the first effective voice segment and the first text content; according to the first confidence score, the preset mute duration is adjusted to be the target mute duration; the target mute duration is in positive correlation with the first confidence score; and determining whether the dialogue of the to-be-recognized voice frame is ended based on the target mute duration. By adopting the method, whether the current conversation of the user is finished can be more flexibly judged, and the voice interaction experience of the user is improved.

The invention discloses a voice interaction method and device, relates to the technical field of voice interaction, and mainly aims at giving a response matched with the emotion of a user in a voice interaction process to realize emotion interaction with the user so as to improve the voice interaction experience of the user. According to the main technical scheme, the method comprises the steps of obtaining voice data of a user in a current dialogue process, and generating a reply text for feeding back the voice data; emotion recognition is carried out on the voice data, and acoustic emotion features and semantic emotion features of the user are determined; determining a real emotion feature of the user by combining the acoustic emotion feature and the semantic emotion feature; determining a target emotion feature and a target timbre feature adapted to feed back the real emotion feature; and playing the reply text by adopting the target emotion feature and the target tone feature.

The invention provides a door lock voice interaction method and device and a related product, and relates to the technical field of intelligent interaction. When the method is executed, a PIR passive infrared sensor is triggered, whether a voice instruction of a user is received or not is judged, the voice instruction is a preset voice instruction, and the voice instruction of the user is used for indicating that a first use scene is met, that is, when the first use scene is met, an interaction strategy corresponding to the first use scene is executed, and the interaction strategy corresponding to the first use scene is executed. Obtaining a door lock voice interaction result, and finally outputting the door lock voice interaction result. Thus, in response to the fact that the PIR passive infrared sensor is triggered and the use scene is met, the interaction strategy corresponding to the met use scene is executed, so that a user interacts with the door lock, and the voice interaction experience of the user and the door lock can be improved.

The invention mainly relates to the technical field of natural language processing. The short text anaphora resolution method comprises the following steps: firstly, constructing an anaphora recognition rule, and carrying out anaphora part-of-speech and position recognition on a currently input statement; then, building a large model cue word based on the recognized part-of-speech of the anonyms, and extracting entity contents conforming to the part-of-speech of the anonyms in a statement input in the previous round; and finally, replacing the anonyms based on the identified anonym positions and the extracted entity content, and rewriting the currently input statement, thereby improving the anonym resolution accuracy and widening the application range.

The invention discloses a commercial vehicle central control screen man-machine interaction system, which relates to the technical field of man-machine interaction, and comprises an information acquisition module, a keyword extraction module, a disassembly analysis module, a voice screening module, a space analysis module and a vehicle machine execution module, the system has the advantages that through cooperation of the keyword extraction module and the disassembly analysis module, deep understanding of semantics of voice instructions is facilitated, the accuracy of the instructions is ensured through multi-dimensional analysis, the system can process the voice instructions more comprehensively and reliably, the accuracy and stability of man-machine interaction are improved, and the system is suitable for popularization and application. The central control screen man-machine starting and dormancy are more flexible, the spatial analysis module is beneficial to understanding the source direction and intensity of a voice instruction, effective voice can be better collected in a multi-sound-source environment, interference of environmental noise can be effectively reduced through the voice screening module, and it is ensured that collected voice information is clear and effective.

The invention discloses a speech recognition system based on a convolutional enhanced self-attention model of an improved scoring matrix. In the invention, a new relation calculation module (RCM) is introduced, so that the attention recovery decoding method is optimized. The module aims to mitigate error accumulation in an attention recovery decoding method, thereby reducing the influence of CTC decoding errors on subsequent attention recovery. In addition, the system also introduces a jump fusion module (SFM) to integrate shallow and deep features. The limitation that a Conformer encoder lacks inter-layer skipping connection is solved, and the capability of effectively capturing and utilizing contextual information of the model is enhanced. Comparison experiments show that a graph-based method improves the recognition rate, especially on a relatively small Uygur language data set. Compared with a baseline model, the method has the advantages that the character error rates are respectively reduced by 0.03%, 0.35% and 0.44%. On a GeneralSpeech16.1 Uygur language data set, the character error rate of the method is 4.19%, and is reduced by 1.56% compared with that of a baseline model.

The utility model relates to a voice-controlled tea bar machine, which comprises a cabinet body containing a water storage container and a back plate arranged at the top of the cabinet body, and is characterized in that a voice module is arranged on the back plate, the voice module comprises a voice output module used for voice broadcast, a mounting structure used for fixing the voice output module is arranged on the back plate, and the voice output module is arranged on the mounting structure. And when the sound output module is fixed on the mounting structure, the mounting structure, the inner wall of the back plate and the back side of the sound output module are matched to form a sound cavity structure. According to the voice-controlled tea bar machine provided by the utility model, the sound cavity structure is formed in the mounting area of the sound output module, and the sound cavity structure can gather the sound of the sound output module at first and then discharge the sound through the sound cavity hole, so that the sound quality and the volume of the sound output module are ensured, and the condition that the sound is not clear is prevented; meanwhile, the sound cavity structure can provide certain protection for the sound output module, and the risk of water inflow is reduced.

The present disclosure relates to a voice response method, device and storage medium, the method comprising: receiving a first voice signal; performing gain compensation processing on the first voice signal according to a correspondence relationship between frequency points of a voice signal and gain compensation values obtained in advance to obtain a second voice signal; wherein the correspondence relationship is determined according to acoustic performance parameters corresponding to the same voice signal received by a first device in a current sound field environment and a preset sound field environment; determining first energy information of the second voice signal; wherein the first energy information is used to determine a target device responding to the first voice signal.