Display Device Voice Wake-Up Verification Using Feature Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Speech recognition through TV devices often results in misrecognition due to environmental noise or other conversations, leading to unintended operations, causing user discomfort.
Innovation Solution
A display device performs verification of the voice wake-up word by dividing the recognition process into two steps: determining successful recognition and verifying if the voice wake-up word falls within a normal range using feature information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If speech recognition is implemented through TV using environmental microphones, then user interaction convenience is improved, but recognition accuracy deteriorates due to environmental noise and other conversations
Solution Approach 1:
The speech recognition process is divided into two distinct stages: wake-up word detection stage and command recognition stage. Each stage uses different recognition thresholds and processing methods, allowing the system to optimize for each specific function rather than using a single general-purpose recognition mode.
Solution Approach 2:
The system dynamically adjusts its recognition sensitivity and processing mode based on the current operational state. When in wake-up mode, it uses high sensitivity to detect potential wake-up words. After successful wake-up, it transitions to command recognition mode with appropriate threshold adjustments to filter out background noise and other conversations.
2Reliability
If the recognition threshold is lowered to detect more wake-up words, then detection sensitivity is improved, but false recognition increases due to similar spoken words and noise
Solution Approach 1:
The system performs preliminary wake-up word detection before entering full command recognition mode. This preliminary action uses a dedicated wake-up word detector with pre-configured parameters optimized specifically for wake-up word patterns, creating a filtering layer before the main recognition system processes commands.
Solution Approach 2:
Different recognition parameters and thresholds are applied depending on the recognition stage. The wake-up word detection uses parameters optimized for high sensitivity to specific wake-up phrases, while command recognition uses different parameters tuned for accurate interpretation of user commands, effectively managing false recognition through parameter adaptation.
Data Source
AI summary
According to an embodiment of the present disclosure, a display device includes a microphone, a wake-up word recognition unit configured to determine whether a voice wake-up word received through the microphone is recognized, and a main processor configured to, when recognition of the voice wake-up word is successful in the wake-up word recognition unit, determine whether the voice wake-up word is a normal wake-up word based on feature information of the voice wake-up word, and, when it is determined that the voice wake-up word is the normal wake-up word, perform a function corresponding to a received speech command as the speech command is received through the microphone.