Hearing Device Semantic Conversation Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Hearing devices face challenges in optimizing directivity in situations with multiple voice sources, making it difficult to distinguish and amplify the desired voice signals while suppressing unwanted ones.
Innovation Solution
The method involves using spatial sound filters and machine learning-based natural language understanding to extract directional sound signals, determine semantic representations, and cluster conversations, allowing the hearing device to amplify specific conversations and suppress others based on semantic content and position.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If hearing devices steer directivity to the front or dominant sound sources in situations with multiple voice sources, then the device can amplify some sound sources, but it cannot distinguish which specific conversations the user wants to hear, leading to inability to optimize directivity effectively
Solution Approach 1:
The system performs preliminary speech recognition and semantic analysis on sound sources before directing audio processing resources. By pre-identifying conversations through semantic content analysis, the hearing device can prioritize processing for relevant conversations and suppress irrelevant ones, solving the problem of not knowing which conversations to amplify
Solution Approach 2:
The patent replaces traditional mechanical/spatial directivity control alone with a hybrid system that incorporates semantic analysis and machine learning. Instead of relying solely on physical microphone array steering, the system uses digital signal processing combined with speech recognition to identify and prioritize conversations based on their semantic content, enabling intelligent selection of which conversations to amplify
2Quantity of substance
If hearing devices amplify multiple sound sources simultaneously, then more voices are captured, but sound clarity decreases due to inability to distinguish desired voices from unwanted ones
Solution Approach 1:
The system extracts and identifies individual conversations from the mixed audio environment using speech recognition and semantic analysis. By separating conversations based on their semantic content and speaker identity, the device can selectively amplify desired voices while suppressing others, maintaining voice distinction accuracy even when multiple sound sources are present
Solution Approach 2:
The patent introduces speech recognition and semantic analysis as intermediary processing layers between the microphones and the audio output. This intermediary system analyzes the semantic content of each sound source and provides control signals to the directivity optimization module, enabling intelligent selection and separation of desired conversations from the mixed audio environment
3Adaptability or versatility
If hearing devices use traditional directivity steering methods, then the device structure remains simple, but the device cannot adapt to complex acoustic environments with multiple conversations
Solution Approach 1:
The hearing device integrates multiple functions into a unified system: speech recognition, semantic analysis, speaker identification, and directivity optimization all work together as interconnected modules. This multi-functional approach allows the device to adapt to complex acoustic environments by combining these capabilities, while sharing common hardware resources like microphones and processors across all functions
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach effectively steers the directivity of the hearing device to focus on conversations the user is participating in, enhancing sound clarity by amplifying relevant voices and suppressing irrelevant ones, thereby improving hearing performance in complex acoustic environments.
Implementation Method 1
receiving the sound signal from a microphone of the hearing device
Implementation Method 2
extracting directional sound signals and optionally a user voice signal from the sound signal. Such an extraction may be performed with spatial sound filters of the hearing device
Data Source
Figure 1~2
Figure 3
AI summary
A method for directionally amplifying a sound signal (36) of a hearing device (12) comprises: receiving the sound signal (36) from a microphone (20) of the hearing device (12); extracting a user voice signal (62) and directional sound signals (40, 46) from the sound signal (36); determining a word sequence (54) from the user voice signal (62) and each directional sound signal (40, 46); determining a semantic representation (58) from each word sequence (54); identifying conversations (34) from the semantic representations (58), wherein each conversion (34) is associated with one or more directional sound signals (40, 46) and wherein each conversation (34) is identified by clustering semantic representations (58); and processing the sound signal (36), such that directional sound signals (40, 46) associated with one of the conversations (34) are amplified.