Speech Interaction System Using Wake-Up Detection and Semantic Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current speech interaction technologies require users to wake up devices before each request, leading to inconvenience and potential incorrect responses when no interaction follows, resulting in a poor user experience due to difficulties in accurately distinguishing human-machine from non-human-machine interactions.
Innovation Solution
Implementing a semantic understanding technology using an improved deep neural network model and online real-time intervention mechanism to derive accurate semantic results from speech recognition, determining whether the interaction is human-machine or non-human-machine, and controlling speech interaction accordingly, thereby reducing the need for repeated wake-ups and improving user experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If speech interaction technology is implemented without wake-up requirements, then interaction convenience is improved, but false activation and incorrect responses increase
Solution Approach 1:
The patent introduces an intermediary mechanism (wake-up word detection module) that mediates between the speech recognition system and the user's speech. This intermediary filters out ambient speech and only activates the full speech recognition system when the specific wake-up word is detected, thereby maintaining high interaction convenience while preventing false activations and improving response accuracy.
2Reliability
If wake-up word detection is implemented, then false activation is reduced, but interaction threshold increases
Solution Approach 1:
The patent applies preliminary action by having the wake-up word detection module continuously monitor speech signals in advance before activating the full speech recognition system. This preliminary detection ensures that the system is ready to respond accurately when needed while remaining dormant during normal ambient speech, thus improving response accuracy without significantly increasing the interaction threshold for users.
3Measurement precision
If semantic understanding technology is used, then speech interaction control accuracy is improved, but system complexity increases
Solution Approach 1:
The patent segments the speech processing system into distinct functional modules: wake-up word detection module, speech recognition module, and semantic understanding module. Each module handles a specific aspect of speech processing independently, which improves control accuracy through specialized processing while managing system complexity through modular design that allows independent optimization and maintenance of each component.
Data Source
AI summary
Embodiments of the present disclosure relate to a method and an apparatus for speech interaction, and a computer readable storage medium. The method may include determining text information corresponding to a received speech signal. The method also includes obtaining label information of the text information by labeling elements in the text information. In addition, the method further includes determining first intention information of the text information based on the label information. The method further includes determining a semantic of the text information based on the first intention information and the label information.